Sheep at the New Jersey State Fair 1995.  Flemington, New Jersey.

The book behind the book behind the book...

("the book" being Database Backed Web Sites by Philip Greenspun)

Pelican.  Audubon Zoo.  New Orleans, Louisiana.
"For a desperate disease a desperate cure."
-- Michael de Montaigne
This document exists to

In the beginning...

Lower Klamath National Wildlife Refuge Well, not the beginning but in 1993 anyway. I wrote Travels with Samantha, a book about North America, about North Americans, and about traveling alone. It won Best of the Web '94 and became one of the 10 most heavily trafficed sites on the Internet. More than a thousand people each day read at least part of the book. A lot of them sent email asking to buy a hardcopy after they'd read the entire book, with its 250 photos, on-line.

My phone started ringing off the hook. Dead trees publishers wanted to do a coffee-table edition of Travels with Samantha. There was a big pent-up demand for the book already, it would establish them as a leader on the Internet, and the spillover traffic from a 100,000 hit/day site onto their Web pages would be worth a lot. ... Ooops. Sorry for that excursion into La La Land.

Meanwhile, back on Earth: Dead trees publishers (1) don't surf the Internet, (2) don't read books, (3) don't understand the value of a high Internet profile. Ergo, they'd never heard of Travels with Samantha. If it were in front of them, they wouldn't read it (they'd read a summary saying "North American travelogue" and pass on the project; travelogues are known to be unprofitable). If you take a look at most publishers' Web sites even as I write this in mid-1997, it is obvious that they haven't seriously thought about using the Internet to communicate with readers.

In the beginning... (Part Deux)

Cedars Market.  Adin, California Publishers might not read or love books but they do love good food. In fact, as far as I can tell, the principal skill and activity of an acquisition editor is choosing restaurants. Simon Hayes of Ziff-Davis Press (Macmillan) came to MIT in September 1996 and asked some friends of mine to round up a dinner table worth of "Web geeks who might be able to write a book. Failing that, just geeks." My name apparently appeared in the second category because nobody called me until the day before. I didn't want to write a dead trees book but I figured that ordering lobster was my only chance of getting anything out of Ziff-Davis, which had been copying my images off my Web site and running them without credit or payment in its magazines (my emailed complaints about this practice were ignored). So we all met at Salamander, an expensive restaurant that somehow manages to survive the periodic bankruptcies of the tech start-ups over its head in an East Cambridge office block.

Things started to go awry in the opening minutes. Salamander had run out of the lobster special; Simon & friends were from "the other Ziff-Davis." It turned out that the book and magazine divisions had parted ways a few years back. So the dinner wasn't being paid for by the people who'd stolen my images (Ziff-Davis Inc.).

A few days later, Simon called to say "I've looked at your [Web Tools Review] URL and I think it is great. This is exactly the kind of stuff we need." Web Tools Review was one of the sorriest corners of my Web server. I started it basically as an FAQ for the hundreds of people who'd emailed me questions about Web publishing. But I hadn't really bothered to maintain it because thousands of people were asking me questions about photography (leading to the massive and IMHO-well-maintained

Simon said "Turn it into a book. I know you can do it. We want your voice."

I told him to forget it. If I wanted to update Web Tools Review, I could do it at my own pace and stick it on my own server and reach plenty of people. If I wanted to feel warm and fuzzy, I'd be better off finishing my Ph.D. research. If I wanted to make money, I'd be better off consulting. Sorry but no thanks.

"You can make $60,000."

"Hmmmm... maybe I'll write an outline," I responded.

After I spent a couple of days working on an outline, I realized that I had much more to say about Web publishing than I'd expected. Three years and 50 sites had taught me a lot of subtle lessons that I thought I could communicate in a coherent manner. Finally, after reading the outline I decided that I had no choice but to write the book. The on-line content at Web Tools Review was an embarrassment. So I called Simon back and said I'd write the book if I could keep the Web rights and use the material to refresh Web Tools Review.

The Contract

Brooks Falls, Katmai National Park Macmillan and I wasted three months negotiating the contract. Typically writers are desperate to get money and fame and therefore the publisher is able to ram almost anything down their throats. I was busy and indifferent. My personal Web site got 500,000 hits/day. I didn't need Macmillan to reach readers. And if I wanted money, I could just invoice some of the big rich companies for whom I'd done consulting. Furthermore, though I'd written many magazine articles, this was my first dead trees book. I had no idea what constituted a fair deal.

My first step was to sign up to the Studio B mailing list. Studio B is a group of agents who represent computer book authors. I would send out Macmillan's proposed contract terms to this list and ask people to comment. My second step was to peruse the O'Reilly Web site. They have done the industry a good service by making their standard contract public, with annotations. The bottom line at O'Reilly is that the author gets 10% of what O'Reilly gets.

Ten percent?

At this point a lot of authors think that there is something wrong. Given that books are typically sold to bookstores for about 50% of the cover price, 10% of net means the author gets only 5% of the cover price. The publishing industry soaks up 95% of the booty. How can that be fair?

Sawmill.  Northeast California. Five percent of retail is fair if you abandon one erroneous assumption: that the publishing industry exists to compensate authors. Magazine publishers exist to sell advertising. The writing and photography are filler. Book publishers exist to pay the salaries of people who work in book publishing. The people who work in book publishing are well aware of this. That's why hardly anybody ever quits a job in book publishing to become an author. Consider the two situations. Author: sits alone at home editing manuscripts and praying that proposals will be accepted. Employee of large publisher: hangs out in comfortable office surrounded by fun people, makes twice the salary, assumes no risk (i.e., gets paid whether or not particular book proposals are accepted).

Most publishers' standard book contracts include a whole page of conditions under which the author gets half royalties or no royalties at all. Special sales, discount sales, book club sales, sales outside the US, direct sales by the publisher, copies sold by an on-line service, etc., etc., etc. ... The message is basically that the author should be grateful to be published at all. In fact, the author is lucky that the publisher isn't charging him a fee per copy printed and sold!

The most upsetting contract provision is indemnification. It is axiomatic that if you publish the truth about anything that matters, you will eventually get sued. That's why newspapers have legal departments. It is an insurable predictable business risk in the publishing world. People will be unhappy with what you write and they will sue you for libel. Sometimes they might even win though they should not have.

Book publishers have decided that they don't want to insure against the risk of litigation. Why should they bother when they are multi-billion dollar companies negotiating with individual authors? So they force authors to sign an indemnification clause. If the publisher gets sued because someone doesn't like what you wrote, you have to pay the publisher's legal bills and, if it comes to that, any legal judgment.

What is an author supposed to do about this? You could go out and buy a business liability policy but that isn't very efficient if you're only publishing one book. What most authors do is write very bland prose that isn't going to upset anyone. Here's how the contract affects texts:

Before IndemnificationAfter Indemnification
The Evil Monopolistic Microshaft hired a bunch of incompetent 22-year-old C hackers to work on Word and that's why it crashes trying to save a document as Text with Line Breaks. Pretty sad considering that Word is just a marginally extended copy of 1970s word processors. Microsoft, the wonderful and generous company that invented the PC and then the WYSIWYG word processor, occasionally makes minor mistakes. Some people say that they've had trouble saving documents as Text with Line Breaks. What you should do is go out and buy an upgrade to the latest version of Word and see if it works for you.
Sybase's basic pricing strategy is to hang the user up by his heels, see how much money falls out, take it all and then ask for another $50,000 for "support". Large RDBMS vendors have different prices for users with different requirements to ensure that each user gets the support and system configuration that he needs.
Here the book publishing industry may have outsmarted itself. An author earning $12,000 from a book is not going to expose himself to liability by giving bottom-line recommendations. So people will turn to the Web and USENET when they need advice, rather than shell out $30 for a book of warmed-over homilies.

Agent or no Agent?

Point Lobos, California Folks on the Studio B mailing list were generally in favor of hiring an agent, delegating to him or her the contract negotiation, and giving the agent 15 percent. If I'd had a proposal and no publisher, I would have hired an agent, mostly to save myself the degrading process of selling the proposal/manuscript. However, I did not hire an agent for the following reasons:

The Decision

I signed the contract on December 19, 1996, about three months after the dinner at Salamander. Macmillan had agreed to my most important condition: that I be able to develop a free on-line edition of the book. I agreed to most of their conditions. The advance that Macmillan offered made me think that the sun would fall out of the sky before I collected the $60,000 figure we'd kicked around. However, by then I'd fallen in love with the outline and resolved to do the project with or without them.


If you glance around the computer section of a bookstore, it is pretty clear that publishers contribute nothing in terms of editorial quality to computer books. If you glance at the server log of a popular Web site, it is pretty clear that publishers don't contribute much in terms of distribution either. So what function does a dead trees book publisher serve these days? Nagging.

Computer books are produced in a pipelined fashion. The author is writing Chapter 9 while the development and technical editors are marking up Chapter 6 while the copy editor is correcting spelling errors in Chapter 3 while the production people are typesetting Chapter 1. This is the only way to get Netscape 4.0 Unleashed into bookstores before Netscape 6.0 hits

A gondolier in a quiet moment If the author decides he isn't in the mood to write one week, the entire pipeline comes to a halt. So the publisher puts a $1000/week penalty clause in the contract and calls the author every few days to make sure that things aren't slipping. I signed a contract promising to deliver Chapters 1-4 on January 9, 1997, Chapter 11 on February 13 and the final Chapter 15 on March 6. I submitted my final chapter on March 26, 1997. Apparently being 20 days late isn't so bad because Macmillan did not enforce the penalty clause against me.

If not for Macmillan's nagging, I would have written a 150-page book in three months and then finished the rest over a couple of years. Though it didn't make financial sense for me to focus on the book so much, I felt that I should because so many people were waiting for my efforts. It seems very inefficient that readers have to pay a bookstore $30 so that a computer book author can earn $1.50 in royalties. But without institutional nagging such as is provided by the publishers, I doubt that many book-length documents would be completed.

Writing is depressing. Either you don't have good ideas, you have some good ideas but can't figure out how to express them, or you can figure out how to express your good ideas but not to the target audience. My worst moment during this entire project was when I read Lives of the Monster Dogs by Kirsten Bakis. Lives was such a beautiful realization of a simple idea. Life is short. How could I justify working on a tech book when I could be aspiring to if not actually creating art?

After submitting the last chapter I felt drained. I wanted to sleep for two weeks and felt that no amount of money could possibly be worth getting out of bed for. Unfortunately, I'd put a host of Web projects on hold during the three months that I was writing the book. So readers, users, contributors, and consulting clients were hammering me with reminder messages and phone calls. I read a New Yorker article about Robert Hughes, TIME magazine's art critic who only managed to finish his American Visions : The Epic History of Art in America with massive doses of Prozac and psychotherapy. If it hadn't been for Alex, I'm sure that I would have become just as depressed.


Charter fishing captain cleaning a tourist's catch. On the wharf in Menemsha, Martha's Vineyard, Massachusetts The editing of my book was interleaved with the writing but for the purposes of this "book behind the book" story, I'm going to pretend that it happened serially.

Up at the top I said that book publishers don't read books. That's true of the people involved in buying manuscripts from authors and marketing finished products to bookstores. However, there are staffers called "development editors" and contractors called "copy editors" and "tech reviewers" who do read the prose submitted by an author.

At least at Macmillan, everyone collaborates using Microsoft Word. I'd wanted to write my book in HTML using Emacs, the text editor I've been using since 1978. That way I wouldn't have to do any extra work to produce the on-line edition and I wouldn't be slowed down by leaving Emacs (the world's most productive text editor, though a bit daunting for first-time users and useless for the kind of fancy formatting that one can do with Frame, Pagemaker, or Word). Macmillan said that the contract provision to use Word was non-negotiable and now I understand why.

Microsoft Word incorporates a fairly impressive revision control system. With revision control turned on, you can see what you originally wrote with a big line through it. If you put the mouse over the crossed-out text, Word tells you that "Angela Allen at Ziff Davis Press crossed this out on March 1, 1997 at 2:30 pm." Similarly, new text shows up in a different color and Word remembers who added it. Finally, it is possible to define special styles for, say, Tech Reviewer Comments. These show up in a different color and won't print in the final manuscript.

[Note: Paul Takemura has written to point out that the one thing I like about Word was not in fact developed by Microsoft. He notes that in "About Microsoft Word", credit is given to "Compare Versions" by Advanced Software, Inc.]

Magnolia biting Alex.  Massachusetts Institute of Technology. If, thanks to the genius of Bill Gates, the mechanics of collaboration were smooth, the process of collaboration was problematic. My goal was to write a book that would inspire and inform intelligent people with enough technical meat to educate an MIT-trained computer scientist but explained clearly enough to be understandable to someone with no programming or CS background. Macmillan's goal was to have a book that would sell well.

It is an article of faith in the computer publishing that bigger books sell better. They take up more space on the shelf and readers with a tech problem find a bigger book comforting. Readers aren't really sure what is wrong with their computer and they only have a few minutes in the bookstore so they figure the thickest book is the most likely to contain the solution. Another important albeit implicit tenet is that readers are incredibly stupid. Combine that tenet with the observation that most computer book authors are incompetent and you get the "book with user interface."

The first component of the tech book user interface can be found at the front of most books: 10 pages explaining which chapters are relevant and for whom (note that this would not be necessary if the authors were capable of writing an adequate table of contents). These 10 pages also usually introduce a whole raft of typographic conventions and icons. The second component of the user interface is icons strewn throughout a book. Instead of a blank line and "Note:" in front of a little aside, you have the big notepad icon (explained in the first 10 pages). Anything that is supposedly technical is preceded by a bolt-with-attached-nut icon.

Another way to bulk up the book is with screen shots. Here's a fragment from my first chapter:

"General Electric ( is a rare good example of a corporate site that takes the Web as Distribution Medium seriously.. They have brochures, yes, but also owner's manuals and installation guides for every GE product. If you moved into an apartment with a GE appliance but the previous tenants did not leave you the instructions, you can grab a PDF file from the GE site and print it."
My development editor for this chapter (Angie) wanted a screen shot of the GE site. I pointed out that the user could just type "ge" into Netscape and surf around. She responded that perhaps someone would be reading the book on a bus. At some level Angie's point was valid but really there was no possible way that one screen shot of the GE Web site could convey what I'd said in that above paragraph. I argued that the the screen shot would break the flow of my writing and distract the reader. The cost of this distraction was far greater than any illustrative value it might have. Mine was apparently a novel argument. Screen shots and figures, no matter how tenuously related to the text, are never gratuitous in the eyes of computer book publishers.

I successfully fought off almost all suggestions to bulk up my book, mostly because the Ziff-Davis Press team was already in accord with me and at odds with the rest of the industry. I insisted that Macmillan leave out the user interface and let the book live or die on its organization and clarity of writing. I put in only those screen shots that I thought were really helpful and mostly towards the end of the book when I'd already earned the reader's attention or lost him completely.

So we'd all agreed to do an uncommercially thin book. But we started fighting again over whether or not it was safe to assume that the average tech book buyer is an illiterate moron. Publishers are terrified that someone might buy a book and be confused. Again from the first chapter:

Some people like a one-truth world. If you have a huge advertising and PR budget then you can control your public image very effectively in a literate world. Ford Motor Company has enough money to remind you 2,000 times a year that "Quality is Job One;" unless you lost a friend in a Pinto gas tank explosion, you probably will eventually come to agree. Microsoft via the genius of Bill Gates invented the mouse-windows user interface, reliable operating systems, affordable computing, and the Internet; if you don't think all that is true, ask someone who has never used a computer and whose only exposure to the industry is through mass media.
Even if he hadn't read the back of the book jacket and learned that I was the author of The Bill Gates Personal Wealth Clock, a native English speaker would normally have no trouble figuring out that I don't personally believe that Bill Gates invented the Internet (or anything else for that matter). Yet my development editor wanted me to stick in a sentence or two disclaiming such a belief. To couch this in commercial terms, avoiding one confused reader was worth, say, $100 to Macmillan. But to have a graceful paragraph instead of an awkward one wasn't worth a penny more. The boredom of a literate reader doesn't carry a cost. That's why computer books read like they were written for retarded 10-year-olds. Don't blame the author. The publishers make them do it.

Macmillan also surprised me by their prudishness. First they rejected my title: "How to be a (small type) WEB WHORE (big type) just like me (small type)" [see below]. Then they took exception to a part of Chapter 5:

Sometimes the referer URL will contain the query string. The very first time I ran a referer report on a server log was on a commercial site. I was all set to e-mail it to "the suits upstairs" when I looked a little more closely at one line of the report. We were giving away "Cosmo Hunk calendars" where each month there was a picture of Fabio or something. A WebCrawler user had grabbed this page and the referer header gave us some real insight into his interests
I decided not to use this particular report to demonstrate my powerful new logging system.
This was way over the top as far as Angela was concerned. God forbid that someone should walk into the computer book section expecting to find yet another paean to the flawless products of the American software industry and come out with a book containing the phrase "hunks with big dicks." They couldn't even believe that I'd written it. Why had I included something so shocking?

"It is true. And it is funny," was my reply.

But I could have written something "just as true" and "just as funny". She typed in the following replacement as a placeholder:

We were giving away sports car calendars where each month there was a picture of a flashy car. A WebCrawler user had grabbed this page and the referer header showed what he was looking for.
Just as true? I'm not exactly a hostage to traditional journalistic ethics but I didn't understand how something made up could be "just as true." Nor did I see how the above example was anywhere close to as funny as the executives of a $3 billion publisher seeing that "hunks with big dicks" query string right at the top of their brand new HTTP referer report.

I tried to persuade them with the example of my cousin who once exiled himself to Muncie, Indiana for two years where he worked for Jim Davis, the man behind the Garfield industry.

It turns out that Jim Davis in person is fond of telling jokes about blacks, Jews, etc. But he knows that this kind of humor isn't commercial so he is extremely careful to avoid offending anyone in the actual Garfield strips. If he must make fun of a group, he'll choose fat people.

My cousin explained that "Garfield isn't funny because it doesn't offend anyone. To be funny, most of the time, you have to run the risk of offending at least a few people. Jim isn't willing to take that risk."

Magnolia on top of Alex.  Killian Court, Massachusetts Institute of Technology. I argued back and forth with various Macmillan folks over "hunks with big dicks" for awhile and I thought that I'd won. At any rate, the pablum suggested by Macmillan wasn't in the "final" draft that I submitted to them. After the book came out, I started getting email from readers. They were confused by a seemingly disconnected section in Chapter 5. I opened up my copy of the book and discovered a whole half-page that I hadn't ever seen. It started off "Someone I know runs a computer answer service, and regularly scans the logs for the 'gooseggs' -- queries that didn't find any matches in the database." This was followed by a bunch of server log entries from this person's site showing users hitting a local CGI script with a variety of search strings.

The first problem with this sentence is that it isn't true. I don't know anyone who runs a computer answer service. In fact, I don't even know what such a service would do. Second, I've never heard of the term "gooseggs" and certainly am in no position to introduce it to anyone else. But these paled compared to the real problem with this half-page: it has nothing to do with the topic of the chapter, i.e., how people get from public search engines such as AltaVista to one's site. I'd written about site-local full-text search engines on page 4 of the book (in Chapter 1) and had no intention of coming back to the subject. If I had wanted to return to it, I would have done so elsewhere. Naive readers were confused. Readers with a bit of Web experience thought I was a loser who couldn't figure out the difference between a Web-crawling search engine like AltaVista and his own CGI scripts talking to local site indexers such as PLS, Verity, or Excite for Web Servers.

Lesson: delivering a text that won't shock anyone is more important than being correct or clear or truthful.

Ziff Davis Press chose to split development editing responsibility for my book. Angie got the earlier chapters and Paula Hardin, thanks to her enthusiasm for database management systems, picked up the later chapters. Unfortunately, by the time I got Paula's suggestions the project was overdue and I was in no mood to make major changes.

--- beginning of flaming digression ----

Flaming Summary: Why Computer Books Suck

Before lighting the candle, let me say that I think I was working with some of the best people in the industry. Many of them agreed with my biggest gripes against computer books, which is (1) why they let me get away with writing what I did, (2) why my book doesn't have a CD-ROM in back, (3) why my book won't destabilize your desk, (4) why the interior of my book isn't filled with user interface widgets, and (5) why the cover of my book doesn't scream "this is for idiots". Do not read this as a complaint against Macmillan or Ziff-Davis Press (now Que). But working with the best people enables one to more clearly see problems that are endemic to the industry. Here then is my explanation for why the interiors of computer books are so bad.

First, since publishers don't pay real money for computer books, the only people who are attracted to work as authors are the clueless and unemployed. If I actually know something about Web publishing, why should I write a book instead of consulting for $1,000/day? But if I've never typed a line of SQL in my life, that makes me the perfect candidate to write a book about databases. Yes the publisher is only going to pay me $10,000 but it works out because I get an excuse to learn a bunch of new things. Maybe I can get a job as a junior database programmer when I'm done.

[Note: there are actually people who choose to make a living writing computer books. They do this by cranking out as many as 8 or 10 books each calendar year. Publishers cherish these people who become, if not famous for technical competence or quality of writing, beloved for meeting their deadlines.]

Second, we must come back to the question of hugeness. If I stole a copy of PhotoShop, which lacks on-line help, and need to know what every command does, then I want the biggest possible book with as many screen shots as possible. But in other situations, size can be intimidating. My principal home computer is a Windows NT 4.0 box that I don't understand. I once tried to buy a book on NT that explained something about the philosophical underpinnings so that I'd be better prepared to use the on-line help. But all the NT books at my local Micro Center were 1200 pages long.

I don't have time to read a 1200 page book. I am afraid to even let one in my house.

Then I looked a little more carefully at the NT books. They are full of screen shots and user interface icons. There isn't really all that much text in them. But the text is scattered among so many "features" that it is impossible to follow. It seemed like it would be too much work to plow through one of these so I bought nothing.

Flaming Example: by idiots for idiots

Shortly after signing my contract with Macmillan, I came upon Creating Cool Web Databases (Sinclair & McCullough; IDG Books). I kept it around on my coffee table so that I could show it to friends and tell them that I'd finished my book. The electric blue cover, garish red "NEW", text-filled packaging, and CD-ROM all instantly screamed "by idiots for idiots". We all got a kick out of watching new people try to say something polite about the book because they thought I'd actually written it.

The interior of Creating Cool Web Databases is about what you'd expect from the publishers of the "for Dummies" books and I thought that surely no worse Web/Database book would ever be brought to market. Then one day I got an Airborne package from Macmillan. Enclosed was a book that they thought exemplified the correct approach to explaining Web/DB technology: Web Database Construction Kit (Khurana & Khurana; Waite Group Press). Here was my email reply:

Here's what I don't like about _Web Database Construction Kit_:

1) it is entirely specific to Microsoft Access and WebSite 1.1.  Reading
this would get me no closer to building a site with any other tool. In
662 pages, I don't even learn SQL syntax (only how to drive Access
through some proprietary forms).  After reading each of 662 pages, I
could not sit down at an Oracle, Informix, Sybase, DB2, or SQL Server
and even see what was in a table.  [Most of what is in this book
wouldn't even translate very well to using another Web server program
with the same operating system and Access, e.g., IIS.]  On page 396,
they do say "One main hurdle of using an SQL statement is that you have
to construct the statement through code, which requires a good working
knowledge of the SQL syntax.  Fortunately, there is an easy way to
bypass this hurdle: Let Microsoft Access tackle the major portion of SQL
composition."  How I translate this: the authors tried and failed to
learn SQL; the authors think they are intelligent; an intelligent person
is able to learn things quickly; ergo, SQL must be incredibly hard to

2) the authors don't understand what a database is for (or they don't
say it).  They talk about structuring data in columns (something you can
do very nicely in Excel or flat files) but not about transactions.

3) there are no IDEAS in this book, nothing that could inspire a
publisher to make a better site.  I wouldn't tell anyone to read this
book unless he wanted to sit down and follow 662 pages of instruction to
build a simple Access/WebSite site.  A manager at Hearst could not read
this book.  The same person might get a bit lost in a few of my chapters
but basically he or she could read my book and get a lot of useful
(IMHO) things out of it.  A young person with a tech degree, e.g., the
typical programmer hired by Hearst New Media, could not read this book
because it sounds like it was written by idiots for idiots (also because
so little information is conveyed in so many pages).

Oh no, I guess it sounds like I'm trashing this book.  I don't really
mean to.  However, I want to be sure that we're on the same
wavelength.  I don't want to write like this.  I want to convey
information and skills that will be useful for many years with many
tools.  I don't want to go step by step through something with Illustra
because the reader may be using Sybase.

Now I'm looking at the message from Mitchell Waite in the front. I guess
I SHOULD trash this book.  He says he is "Creating the Highest Quality
Computer Books in the Industry".  He says I should send him e-mail if I
want to comment (but he doesn't give his e-mail address).  Oh my God,
now I'm surfing his Web site.  It has frames, white text on a black
background (bleah).  Custom link colors. I clicked on News from the
menubar at the bottom of the cover page and got an image map file (the
actual imagemap defining rectangles that the server was supposed to use
to redirect me).  The site has no provision for incorporating reader
comments or doing anything with the Web that couldn't be done on paper.
Book promo pages aren't linked to sources of software in electronic
form.  The site seems out of date.  I couldn't find Web Database
Construction Kit, not even after using the search engine.

In short, the Khurana book embodies everything that I've been saying is
bad about computer books.  The Waite group Web site is an excellent
example of all of the practices that I deplore in the early chapters of

not working for Waite and glad about it
I went and revisited the Khurana book just now, feeling that perhaps I'd dissed it too hard. Couldn't I find something in its 662 pages to like? I opened it to page 16 and looked at Figure 2-3: "A map of the United States showing the state boundaries." Above the caption, surprisingly enough, a stock disk grayscale map of the US took up a third of the page. This was somehow supposed to illustrate the idea of an HTML imagemap. Below it was a list of bulleted items. Instead of a regular bullet, each item was preceded by a drawing of a satellite dish. This instantly reminded me of the icons my friend developed for a bandwidth-stealing site.

Just Plain Flame: Cheerfulness Considered Harmful?

As long as I'm setting down all of the things that I hate about computer books, let me add cheerfulness. Any complex software system has portions that are well-designed and do the job and portions with holes big enough to fall through. There is some kind of conspiracy among computer book authors to cover both kinds of subsystems in the same uniformly cheerful tone. This masks important distinctions among sections of a program and can be incredibly annoying.

Suppose that you are up all night tearing your hair out because something has gone wrong with your RDBMS. You turn to your technical bookshelf and thumb through all the dbadmin guides. Perhaps you do find some useful information but you become enraged by the cheerful tone of the book. You are in this mess because the RDBMS vendor skimped on the design and implementation of a critical system component. This skimping may well have been documented somewhere ("the difference between a bug and a feature is documentation") but you didn't see the relevant caveats before the skimping brought down your service. Partly this is because tech books don't have sections like "design idiocies that are likely to fuck you over." When you finally do find the relevant passage, it is phrased as though the design shortcoming were perfectly reasonable. How else would you want the system to work?

So you feel utterly alone. As far as you can tell from reading the vendor's official documentation and the combined products of the computer book industry, nobody else has ever had a problem with this RDBMS. You are the stupidest, unluckiest, and most incompetent person ever to walk the face of the Earth.

Why do editors push authors into writing this way? Because there is a belief that computers are intimidating. They aren't friendly enough to the users so we'll make tech books cloyingly friendly to compensate. IMHO, this idea sucks.

Something Nice About Editors

Much as I love to vent my spleen, fairness demands that I acknowledge the substantial contribution my development editors made. Sometimes they took stuff that was good and said "this sucks and/or is confusing". In those cases, I mostly ignored them. But they often found stuff that was bad and said "this really sucks and/or is really confusing." On about 100 pages out of 350, there is at least a sentence or two that I owe to their hypersensitive noses. My nerd friends sometimes found minor technical errors or stuff where I'd oversimplified a bit. But they were all too nerdly to effectively simulate a non-technical reader. They filled in details without ever being aware that they were doing it. My development editors kept me from being sloppy and lazy.

--- end of flaming digression ----

We continue with the rest of the story of the production of my book....


I got a C in handwriting in Third Grade and that was more or less the high point of my artistic career. One of the best things that Macmillan contributed to this project was a cleverly designed and executed set of napkin drawings, Paula Hardin's idea executed by Mina Reimer. I roughed these out and FEDEXed them to Macmillan and they had Mina spend hours making drawings that looked like I roughed them out. At least half of the people who read the dead trees version comment on these drawings.

Illustrators are cheap ($150-250/drawing?) but very few Web publishers ever hire them. They somehow think that they can do it themselves in Freehand, Illustrator or Canvas. Most of us can't, though.


As stupid as I found the idea of printing a book about Web publishing, the idea of stuffing a CD-ROM in the back seemed to belong to a whole new category of stupidity. Macmillan initially wanted a CD-ROM, on the grounds that readers think such books have more value. I said that if we couldn't get some complete RDBMS packages for the CD-ROM then there was no point in having it (and in fact as my book came out Oracle decided to make all of its software available for download on the Web so there would not have been any point even if we could have gotten a full Oracle for the disk).

I asked Macmillan to put in the standard CD-ROM pocket but fill it with a black cardboard disk, said disk to be printed with the URL for the book's virtual CD-ROM ( Macmillan said that would be more expensive than a real disk so we ended up printing the inside back cover with a nice "no CD" symbol underneath which ran my text:

Would you really want to take Web publishing advice from someone who had to burn a CD-ROM to distribute his software? Come to for electronic versions of the source code examples in this book, for live demos of the software in use, and for the packaged source code to larger systems. IMHO, this URL is better than a CD-ROM. You can't lose it. You can't scratch it. You can't leave it in your office when you need it at home. You can give it to your friends and still keep it for yourself.
People laugh when they read this so I think it worked.


Adin California. In the beginning, computer books were thin hardbound volumes in muted colors with minimal text on the cover. Then a marketing genius produced an incredibly thick book, in an incredibly bright color, with an enormous amount of text on the cover. Readers wandered in the bookstore knowing that they had a problem with, say, Lotus 1-2-3 macros. Attracted by the thick brightly-colored spine, the reader would pull the book from the shelf and then scan the front and back cover. Because there were so many words and software product names festooning the book, there was a good chance that the reader would spot "Lotus 1-2-3" and "macros". Thus did the book get bought without the reader carefully examining its interior and/or competing titles.

Like stockpiling atomic bombs, this approach works best if you're the only one doing it. When you walk into the computer book section of a store today, you're seeing the result of a decade-long arms race. Every book is fat. Every book is bold. Every book has enough keywords on the cover that you almost need an index for the cover itself.

I said to Macmillan that I spent all of Chapter 3 arguing against the Web equivalents of this approach. Could we not try to set my book apart by being tasteful? Why not make a computer book that looked like a novel? We'd do it hardcover with just the title and author name on the front. On the back, we'd have blurbs from famous Web nerds hyping the book. On the inside flaps of the dustjacket, we'd have a description of the content and a brief author bio.

They chuckled at my naivete: "You might know how to power cycle a dead Web server, Greenspun, but you've got a lot to learn about marketing computer books."

Macmillan initially wanted to title the book "Greenspun's Guide to Web Publishing". This is apparently prestigious for the author, i.e., to have one's name in the title. Still, I thought the title was lame and wanted the book to look like this:

How to be a

just like me

by Philip Greenspun

They responded: "Bookstores won't carry it; it is offensive to women." Macmillan persuaded me finally with the argument that people who shop in the computer book section are problem-oriented. Books that are specifically about Product X sell better than books that aren't associated with a particular product. Books that are just good general background reading don't sell at all. My own experience bears this out. I very seldom go looking for a book on "database fundamentals" or "programming language theory" but I might well walk into a bookstore looking for a book about Oracle or a Perl manual.

So my book would go into the bookstores around June 1, 1997 looking just like every other computer book ever printed: a soft cover covered in text and a problem-oriented title.

The Quest for Blurbs

If I'd been smart, I would have gone on a blurb quest the moment I started writing. But instead I did it in a frantic rush after I'd finished the draft manuscript and could email people a full rough copy. For "real books", the publisher and publicist go out and collect blurbs. But in the computer book biz, apparently the author is on his own.

My first discovery was that if you've offended everyone you've ever talked to or written about, it is tough to get blurbs. My friend Neil says that this is why stock analysts never say anything truly negative about a company. If they are known for trashing companies then executives will no longer talk to them. Without access to information, the stock analyst is then out of business.

The ultimate blurb would have been from Bill Gates. So I sent off some email to and got back

April 10, 1997

Philip Greenspun
MIT Dept. of Electrical Engineering and Computer Science

Dear Philip: 

Thank you for your email inviting Bill Gates to review your new book
on Database-backed Web publishing and to provide a quote for the
book's publicity. 

Bill appreciates the invitation and your interest in his insights and
opinions, but regretfully must decline.  As you can probably imagine,
Bill receives an extremely large number of requests for his time
throughout the year.  Unfortunately, due to his heavy work and travel
schedule he is limited as to the number he can accommodate.  

Thanks again for extending the invitation and for your interest
towards Bill.  We wish you all the best with your new book. 


Karen Tuazon for
Mich Mathews
Director, Corporate Public Relations 

Perhaps Chapter 9 would have been just as good without the phrase "Bill Gates gave bloated monopoly a name, a face, and a smell". Or perhaps I should not have written The Bill Gates Personal Wealth Clock back in 1995.

Here's a table of people I asked for blurbs and my various sins against them over the years:

Name/CompanyMy SinsResponse
Larry Ellison/Oracle Various Web and book references to his babes and sexual harassment lawsuits did not answer email
Steve Jobs/Apple/NeXT/Whatever dismissal of Mac OS as useless for Web server did not answer email
Scott McNealy/Sun said Sun hardware reliability and hardware/software support was inferior to HP did not answer email
Tim Berners-Lee/Web Consortium harassed him repeatedly in corridors of MIT CS lab so that he would support semantic tags in Web documents and gamma information on IMGs; sent him email complaining about all the dead links and lack of searchability at did not answer email
Nicholas Negroponte/MIT Media Lab linked the phrase "Media Lab" all over my site to the savagely funny Voodoo Magazine parody of the Media Lab; noted that the Media Lab was just about the last department at MIT to put up a Web site said he wanted to read the book before blurbing and didn't have time
I got at least 100 readers/day coming to Web Tools Review for its brutally frank tales of my bloody confrontations with operating systems, Web servers, and RDBMS. But these people never forked over any money. Much better to suck up to the rich and powerful and eat crumbs from their plates.

The very first person I contacted via phone for a blurb was Dave Siegel. He's a friend of a friend of mine so I figured I had a chance. I desperately wanted his name because he's the author of a bestselling Web site design book (Creating Killer Web Sites). This was an incredibly dishonest act on my part. I could have walked one floor downstairs at MIT and asked Dave Clark, the author of the TCP/IP spec, for a blurb. Clark is the father of the modern Internet and it is because of his engineering ability that I'm able to type this document right now (I'm running Emacs on Unix Box A, editing a file that is NFS-mounted from Unix Box B, all from the X server on my home computer). But hardly anybody has ever heard of Dave Clark so I didn't bother to ask him for a blurb. Instead, I asked Dave Siegel, whom my New York graphics nerd friends refer to as the "Barry Manilow of graphic design" and whose book struck me as completely misguided if you intended to build a site with more than 10 .html pages and/or if you were going to build a site that performed a service rather than just static .html.

Worse yet, six months earlier his on-line advertisement for a Swiss babe had caught my eye and I'd stuck him into my article Using the Internet to Pick up Babes and/or Hunks ("because sometimes getting f**ked by Unix isn't all that satisfying").

So I got my friend to call Siegel and I walked him through a draft of the book. Though Chapter 3 is in many ways one of the least interesting, he zeroed in on it immediately and on this unfortunate paragraph:

Graphic designers get it so wrong because they never figured out that they aren't building CD-ROMs. With a CD-ROM, you can control the user's access to the content. Borrow a copy of David Siegel's Creating Killer Web Sites and note that he urges you to have an "entry tunnel" of 3 pages with useless slow-to-load GIFs on them. Then there should be an "exit tunnel" with 3 more full-page GIFs. In between, there are a handful of "content" pages that constitute the site per se. Suppose that Siegel's implicit assumptions are correct, i.e., (1) there are no users with text-only browsers; (2) users want to wait 45 seconds before getting to the content of a site; (3) there are no users who've turned off auto image loading; (4) there is some obvious place to put these tunnels on a site like mine with 1000 .html pages. Even if all of those things are true, if the internal pages indeed contain any content, Altavista will roar through and wreck everything. People aren't going to find the site by typing in "" so that they can be led around by the nose by you. They will find the site by typing a query string that is of interest to them into a search engine. The search engine will cough up a list of URLs that it thinks are of interest to them. Altavista does not think a Dave Siegel "entry tunnel" is cool. In fact, it might not even bother to index a page that is just one GIF.
Siegel was not amused. He apparently does not subscribe to the multi-truth vision I painted in Chapter 1 of my book. He asked "What's in it for me?"

[Note: I eventually did get a few blurbs for the dead trees edition, though really none as effective as the reader reviews at]

The First Copy

Towards the end of May, 1997, I got the first printed copy of the book. Being a glass-is-half-empty kind of person, I was disappointed in the color insert. It looked like it had been printed out by a 10-year-old on his $200 Epson inkjet. I'm used to having my photographs printed in travel magazines and books from publishers who regularly do 4-color printing. A friend looked at the vaguely marbled greenish cover and said "it looks like someone hurled on a white book and then wiped most of it off." My friends at MIT Press (who'd wanted to publish the book but I'd dissed them because Macmillan said they would sell and promote it 100 times better) said "oh they really messed up the cover."

I stuffed the book in my backpack.

The First Reader

Flying out to San Francisco, I happened to sit next to a Harvard MBA. I showed him the book. He read the entire book, every page, during the 6-hour flight. When he finished he said "I learned from every page. I don't have any technical background but I found all of the explanations very clear. The book was funny and easy to read all the way through." I was adjusting my position so that I could pat myself on the back but then he continued: "I never would have bought this book if I'd seen it in a bookstore."

The title and cover text made him think it was a book for hardcore nerds.


Sell Your Bad Memories Here.  Everett, Massachusetts
Without publicity a terrible thing happens: nothing.
-- P.T. Barnum
The only computer book authors that I personally know are Nicholas Negroponte (Being Digital) and Michael Dertouzos (What Will Be). When these books were new, they were in a big stack at the front of every bookstore in the United States. The authors were interviewed on the radio, on TV, in magazines and in newspapers. Any American who might conceivably have wanted to read either book would have been forced to at least look at the cover.

So I waited by the phone. Nobody called.

I went into Wordsworth, the bookstore in Harvard Square with the biggest nerd book collection. They carry every "how to program HTML" book ever written. They carry a bunch of Web/database titles by authors who don't know SQL. My friends would go into the store and say "I've heard great things about this Greenspun Database Backed Web Site book. Do you have it?" No. "Is it on order?" No. "Can you order it?" Yes, with payment in advance.

My friends at MIT Press had ordered 50 copies for their little bookstore and wanted me to come in and sign them. A signing! Just like real authors. So I went down there every day. No books. MIT Press had ordered their books a month before publication date to be sure that they had it as early as possible. Two months later: no books. The store manager had sent Macmillan a couple of FAXes and made three phone calls. Finally, Macmillan answered "Oh yes, we decided that you're too small for us to deal with. You'll have to get the books from a distributor."

My worst humiliation was in New York City (whose isn't?). Here's how I described it to my brother:

I was walking down Columbus Avenue with my friend Bobby.  His literary
agent also represents Saul Bellow.  Despite his credentials as a serious
novelist, he loved the copy of my book that I'd given him and, as we
walked by what is allegedly the largest Barnes & Noble ever (across from
Lincoln Center in Manhattan), Bobby said "Philip, let's see how B&N is
presenting your book.  I'm sure it is going to be huge, not like
literary fiction."

[Background:  last time Bobby had a book at B&N it was on their 10 best
sellers list and copies filled an entire front window.]

In the computer section, they had a table out front with a bunch of
Internet books for people who couldn't figure out what those nine
buttons at the top of the Netscape browser do.  They had a massive "for
Dummies" section behind the table.  Behind the "for Dummies" section was
an even larger "how to program HTML" section.

They had a big Web site design area with Dave Siegel's book.  Right
underneath was a two-shelf Web/Database design area.  They had a big stack of
IDG's "Creating Cool Web Databases" (the book I kept on my coffee table
as a joke), presented cover out.  They had some better books.  They did
not have my book.

Nearby, Barnes & Noble also had a big database section.  It did not
contain my book.  

We asked the clerk.  He looked up the book in the computer.  "Ah yes, we
have one copy.  It is in the network section."  So we walked across an
aisle to a completely non-Web non-Internet non-database area of the
bookstore.  There, among books describing the TCP/IP protocol (something
you don't need to know about unless you are writing your own copy of
Unix or Windows NT from scratch), was one copy of my book.  Spined.

If a person walked into that store knowing that he wanted to build a
database-backed Web site, he would have never come within 20 feet of my
book.  I don't know if Barnes & Noble uses a central computer system
but, if they do, I imagine that all 700 copies they ordered are in the
wrong place.

My brother promptly wrote me back that he'd gone into a Barnes & Noble in Washington, D.C. and found my book in the network section there as well. I stopped going into bookstores.

When the dead trees world lets you down, you can always turn to the Web, no? So I visited the The Macmillan Computer Publishing Web Site and clicked on "what's hot". My book apparently wasn't. But they did have a nice banner ad for Scientology: "with Scientology technology you could understand the purpose of life and through understanding, achieve the goals you set out." Using the search engine, I managed to find a generic page for my book. There was no link to my Web site, no sample chapter, no scan of the cover, no blurbs from the inside cover, no author biography (just to list a few of the things that were actually in Macmillan's possession). This was on July 4th, almost three months after my book was completed.

I'd talked to Macmillan earlier about them running some advertisements for the book. They said "typically, if a book is selling well after a few months then we think about running ads." This plays into my theory that publishers don't read books. It is easier to hire 10 losers to write 10 books on the same topic and pay them each $10,000. Then they dump all of these books into the bookstores and let the public sort it out. After a few months, their inventory computer shows them which of the 10 books is selling best and they get behind that one. Wouldn't we all be better off if the publisher paid a skilled writer $50,000 to do a good Perl/CGI book and then actively pushed it?

Anyway, Macmillan apparently puts almost all of its efforts into schmoozing bookstore buyers and cooperative advertising with retailers. So the chance that an author would see an ad for his own book is minimal.

One thing that worked great in the case of my book was sending out promotional copies. We shipped out 100 three weeks after the book was printed. Almost everybody read it. Almost everybody who read it loved it. Almost everybody who loved it recommended it to several friends. Almost everybody who recommended it to friends also wrote a reader review at (more about that below).

[Advice to authors: insist on a clause in your contract that says the publisher will ship out at least 200 promo copies (to addresses of your choice) within one month of the book coming out of the bindery. Fill this list with either people who are really well-connected and/or people who respect you (preferably both). These people will drive initial sales and reviews.]

Why Arnold Schwarzenegger is depressed

As with most of my Web content, you've probably noticed the overwhelmingly negative tone of this story. However, there has been an upside to the entire experience: I've figured out how rich successful popular movie stars can be depressed.

I should have been happier with the book done. I wanted to write a computer book that an Ivy League graduate with no technical background could learn from, enjoy the experience, and enjoy the writing. By all reports, I had done it. Some of my computer science professor friends were saying that the book was "a major contribution." My brother loved it. My writer friends who'd never "jacked into this World Wide Cybernet thing" loved it.

I should also have been reasonably happy with Macmillan's efforts. They'd unloaded 2,500 copies onto Barnes & Noble and Borders in the first month. Compared to the 5,000 readers/day who visited my Web site, this wasn't significant but it was still a potential 2,500 extra readers. That's better than 0.

So why was I depressed?

Because my expectations had been raised by the effort to the point that I couldn't be satisfied. Why couldn't the entire staff of Macmillan stop what it was doing and spend full time promoting my book? Why didn't Barnes & Noble clear out all those Teach Yourself to be an Unleashed Dummy in 21 Days (TM) books and replace them with Database Backed Web Sites?

It has got to be a lot worse if you're Arnold. Despite my overinflated ego, I'd be perfectly happy to make the front page of the entertainment section of the newspaper and to be called "action hero Philip Greenspun". But Arnold Schwarzenegger's ego is inflated at least another 50 psi and he would be enraged and depressed by the same billing. Why wasn't it on the front page of the newspaper? How could that scrawny no-talent loser from Hong Kong (Jackie Chan) be mentioned in an earlier paragraph of the same article? Fire the publicist! Send a nastygram to the studio.

My solution: stick to the Web. I don't have time to surf so I have no idea whether my Web fame is growing or not. I don't have time to pore through my 50 MB of daily server logs so I don't really even know what's happening with my traffic. The HP Unix box looks physically more or less the same whether it is serving 500,000 hits/day or 50. Nobody is supposed to be promoting my Web site so I don't have any reason to be upset if it isn't featured in magazines, newspapers, or bookstores.

The future of dead trees book marketing:

Rhya Fisher at Harvard Bookstore.  Cambridge, MA 1998. Looking at the way my book was marketed made me realize that is going to rule the world. A traditional bookstore is useful as an entertainment venue. You can arrange to meet someone there. You can kill 20 minutes browsing. But if you're picky about what you want, the chance of them having the book is pretty small. They carry books that are being heavily hyped and books that were popular and relevant six months ago. Traditional bookstores can't respond quickly to customer demand for new or newly popular titles. In dozens of cases, friends of mine would go into a store to ask after Database Backed Web Sites. Usually the book had not been ordered and the store had no intention of stocking the title. The front desk clerks had no mechanism to provide feedback to the buyers. If a person did not plunk down his credit card and special order the book, no record would exist of the inquiry.

At an on-line bookstore, however, every in-print book is "in-stock" and available to be ordered, at least in theory. Thus, the on-line bookstore need only query its RDBMS to find out what new books are selling well and should perhaps be featured or stocked for quicker shipment.

Given that on-line bookstores are the right thing for an obscure and/or new title like mine, what are the differences among the on-line stores? First, has reader reviews. If a book is good, the people who are really experts in the field are going to get it first and have an opinion about it first. Amazon captures these peoples' thoughts with their reader reviews. A good book will always get reader reviews months before it gets the attention of dead trees magazines or the bookstores' "expert editors." So the Amazon page worked much better to sell my book than the page at Computer Literacy, a computer specialty shop that recommends particular titles (not mine as of July 12, 1997). Another good thing about Amazon is that they allow authors and publishers to comment on a title (more free content for them; more information for readers) and have a standard interview form for authors.

[Note: You really ought to be following the links from my Amazon page to my author interview. The form is standardized and designed for "real writers." Here's me answering their least apropos question: What authors do you like to read? What book or books have had a strong influence on you or your writing?

P.G.: Oh my God, I guess this question wasn't intended for authors of computer books. Yes... I was reading Annie Proulx and Tolstoy yesterday. Their example inspired me to write "Netscape 4.0 for Dummies". ...

What surprised me most was reading other authors' interviews at Amazon. Amazon had apparently expected to get real writers but most of the authors who'd stumbled upon the form as of May 1997 were hack writers of tech books. You'd think that lying on their college applications would have prepared these people to answer this question with something like "Every summer I sit on the beach and re-read Proust. It recharges my creative batteries." But apparently they'd all lost the skills they'd had at age 17. My favorite interview was with Simson Garfinkel, one of our local professional computer book authors. In response to the above question he said "When I was in high school, I read a lot of Larry Niven [ed: a science fiction writer]. ... Now I only read non-fiction."

Sadly, you can't find this interview with Simson anymore. Someone must have realized that it was an embarrassment and deleted it.

Another thing that Amazon has going for it is Web presence. I would have thought that the cover page for the book on my site would have been the natural spot for readers to leave comments. But even people who did not buy the book from Amazon would go to the book's page there instead and leave their comment.

The most pathetic example of on-line marketing is Barnes & Noble. Engineering the software that sits behind the Amazon site is not that difficult. In fact, I built something very similar for MIT Press in a matter of weeks. However, the people who designed the Barnes & Noble site apparently lacked the tech skill or the publishing imagination to do anything interesting (even though they launched two years after Amazon). They don't have any provision for authors or publishers to comment. They don't have any provision for readers to leave reviews. Their pages look bland and bleak compared to Amazon's. The saddest thing about the Barnes & Noble site is that they have swallowed whole a dollop of Media Lab hype and bought the Firefly system, an example of a whole genre of naive thinking that I diss in Chapter 15. At Amazon, they cleverly query their database to figure out that buyers of Book A also bought Book B and therefore if you are interested in A then you probably should look at B. At Barnes & Noble, they stupidly ask you to log in and rate all the books that you've ever read before they'll serve you up anything more than publishers' marketing copy.


Mission Santa Barbara.  California. A story this long ought to have a conclusion. Unfortunately, I'm writing this two months after my book came out of the bindery. That's little more than an eyeblink in the dead trees world. Database Backed Web Sites is number 474 on the " 500", proving (IMHO) that my low-key on-line marketing efforts have been reasonably successful. The book is often spined in an obscure corner of a traditional bookstore, proving (IMHO) that Macmillan's marketing efforts are still gaining acceptance.

I wrote a book that, once read, appeals very strongly to literate people who don't like computer books and who don't ever go into the computer book section of any bookstore. If Macmillan gets behind it with some serious marketing dollars, the dead trees version may yet have a chance to reach those people as well as the usual computer book buyers. If Macmillan decides it isn't worth it, then I can fall back on my new Web server, a 4 CPU Hewlett-Packard Unix box with 4 GB of RAM and 150 GB of disk space. Running AOLserver and connected to MIT's T3, it is capable of delivering the full text of my book to every English speaker on Planet Earth in four years (assumptions: 650K of HTML; 1 billion people). I guess I'd better finish up the Web edition...

Update: September 1997

Kayaker on the Nenana River.  Just outside Denali National Park (Alaska) The MIT Press bookstore finally got their 50 copies. They stuck them out near the front of the store. My book is now their 8th bestselling title. So the good news is that, if placed in the front of the store, the book is a surefire bestseller. The bad news is that there is only one bookstore in the world where Database Backed Web Sites is out front. killed their " 500" list. So my book is no longer featured there. They still have a Computer 50 list. I'd expected to surely make this list the week that Hewlett-Packard bought 31 copies from amazon, but my book did not appear. Hard to explain since a handful of sales were enough to push it to 474 on the old 500 list (of all books; not just computer-related ones).

The Internet bigshots at Hewlett-Packard may save my book yet. They've been buying boxloads at a time and thrusting them into the hands of employees and partners. Every time HP buys a box, though, it brings home to me the cost of writing the truth. If I'd toned down the book and hadn't skewered Oracle, Netscape, Informix, Sun, and Microsoft then maybe these companies would be buying boatloads of copies as well.

A friend of a friend of mine in Japan decided that he wanted to translate the book. He's got a Ph.D. in computer science from Carnegie-Mellon and he manages R&D for one of the world's largest software developers (a household name in Japan). He's connected to all the industrial and government consortia in Japan and probably could have gotten at least 20,000 copies sold just with a few handshakes. The Macmillan people were unimpressed. They only entertain translation offers from established publishers and generally seemed in no hurry to get the book done in Japanese. I'm not sure what the lesson here is but I guess it is to get translation and foreign editions squared away before the start of a project. Perhaps to reserve the foreign rights if the publisher does not commit to bringing out foreign editions.

After three months on the market, has the world taken notice of my magnum opus? An HP employee down in Washington, D.C. went to his local Barnes & Noble and tried to buy a copy. Not stocked. One of our local high school kids turned hardcore Web/DB developer went off to Yale for his freshman year:

Date: Fri, 12 Sep 1997 14:03:50 -0400
To: philg@MIT.EDU
From: William Crawford <>

I ducked into the Yale Co-Op today and saw several copies of Siegel's
"Killer Web Sites" on a display table right inside the front door. After
heading upstairs, I found a lone copy of your book, spined, in the back
of the networks section.

Continuing our survey of the Ivy League... (October 1997)

From: Rosemary Michelle Simpson <>
Date: Fri, 24 Oct 1997 15:14:34 -0400 (EDT)

I wanted to get the book (I'm webmaster for several CS dept sites) and
called the Brown University bookstore.  (1) The computer store book
buyer had never heard of it and he "buys at least one of everything
that's presented to me", (2) when I attempted to special order it, I was
told that the title is "RDBMS-based Web Sites" and that they'd had a
copy on special order for months and had yet to receive it.  She told me
that apparently the publisher wasn't interested in providing it.

Why I don't just kill myself now

The preceding story has been various described as "scary", "pathetic", "heartbreaking", and "depressing". So why don't I just kill myself now? Because I keep getting nice email messages from people who've read either the Web or the paper edition. I only get a few a day but that's more than enough to make me think that the effort I put into the whole project was worthwhile. Here's a sampling of my personal favorites:
"I have started reading your book, and I'm more impressed than ever by your godlike Web and writing skills. Also, thanks for sending me the laminated Oakland map--that was sweet of you. However, I'm *still* not going to sleep with you."
-- an old friend in California; arrived underneath a subject header of "The Stearns & Foster of Web publishing"

"I've been reading chapters 10-12 of your web whore book from your web site. It's the best, funniest, most interesting, etc. computer book I've ever read. I also liked the random strange photos strewn throughout the text."
-- Michele

"It was the most entertaining technical book I've ever read. Come to think if it, it's the only technical book that I've ever read from cover to cover."
-- J. Piscioneri

"Your book was even better than I had hoped it would be. I bought it mainly because I enjoyed your web sites and figured I'd have a few laughs by reading it, but it contained information that definitely influenced my point of view."
-- David

"Your book is the best about this topic I've seen (and it is the most readable - and amusing - book about computer science I've read in the last years)"
-- Till (from Ulm, Germany)

"your book was (is!) a gloriously informative and funny read."
-- Karen

"Thanks for writing your book. I'm not always a thinking person, but I was while I was reading it. ... the typography on the cover is unspeakably ugly. Gratuitous justification, ugh! If it's your fault, shame on you. Otherwise, I feel sorry for you in your pain."
-- another David

"Despite the few pages at the end of your book I get the impression that you really believe that engineering stuff can really make a difference in the quality of our lives. Once again I'm sure other people have made similar comments to you, but I have to say I'm much more skeptical that this stuff really matters. We live in a society in which its becoming impossible to raise a family on a regular job, work is increasingly being outsourced to miserable third world wretches making pennies a day...these aren't problems that are going to be solved with engineering alone."
-- Bruce (thought he should have the last word with this)

Independence Day: January 7, 1998

On the morning of January 7, 1998, seven months after Database Backed Web Sites hit the streets, I got two overnight letters. The first was from a woman in the remaindering department at Macmillan:
"We are pleased to have published this book and highly value our author/publisher relationship. However, the decision has been made to take 1562765302 DATABASE BACKED WEB SITES out of print."
I could buy any of the remaining 283 copies (out of 8000 printed) at $4.50 each.

The second letter was a signed contract from Morgan Kaufmann for a new book. It will be bigger, reflect a rethinking of some of the big issues over the last year, and reflect improvements in technology (e.g., cascading style sheets, XML). The book will be four-color throughout and filled with photographs from my site, kind of a coffee table computer nerd book.

My divorce from Macmillan wasn't actually as harsh as it sounds. Basically the short story is that Macmillan killed the Ziff-Davis Press division just as my book was coming off the presses. Something like 100 people lost their jobs and many writers found that their contracts had been canceled. I don't know anything about publishing or business, but I think Macmillan actually made the right decision. The book the Ziff folks sent me as an example of their art was Late Night VRML 2.0 with Java, 700 pages + CD-ROM, published February 1997. I was personally acquainted with more movie stars than people who might conceivably have wanted to buy this book or any book like it.

When a big company kills a division, they don't go sifting among the ashes to find something to salvage. So despite my theories about my book's potential, it was lumped in with hopeless marketing cases like Late Night VRML. I nagged Macmillan to either assign the promotion of the book to new people or kill it. They graciously agreed to kill it and give me the print rights back.

If Macmillan hadn't noticed anything special about my book, lots of other publishers had. I'd been getting periodic email from acquisition editors at other houses anxious to sign "my next book". I started returning their emails. Very quickly it came down to a mega corporation eerily similar to Macmillan and Morgan Kaufmann, one of the most prestigious computer science publishers. I turned down the mega publisher's $25,000 advance when the acquisition editor started talking about all the other people in the company whose approval he would have to win in order to do a book the way he/I wanted. Morgan Kaufmann has 35 employees. Mark Stone ( invited me to come out to San Francisco and meet the team. I shook hands with and eventually signed a contract with Mike Morgan, the owner. He cannot say "oh, we reorganized and now your book is being handled by this other division." If he changes his mind about my book, he has to admit that he is changing his mind.

What else did I like about Morgan Kaufmann? They put the author first. They have to because their authors are people like Dave Clark, research scientist at MIT and developer of TCP/IP (i.e., the Internet). They are not going to tell Dave Clark that he has to use Microsoft Word because Dave Clark doesn't have too much spare time and Morgan Kaufmann wants him to be able to focus whatever he can spare on the content. So when I said that it would be easiest for me to write the manuscript in HTML using Emacs, Morgan Kaufmann replied "no problem".

Another good thing about Morgan Kaufmann is that I will be their stupidest and most commercial writer. Seemingly everyone else who writes for them has a Ph.D. and a professorship. They don't have a "DOS for Dorks" series with potential to sell 200,000 copies/years. So when they do get their appointment with the buyer for Barnes & Noble, my book will be the one that gets pushed rather than Atomic Transactions: In Concurrent and Distributed Systems (two of whose authors are tenured profs in my department at MIT).

A final factor in my decision to go with Morgan Kaufmann was that everyone who'd shaken hands with Mike Morgan had good things to say about him. It turned out that many of the folks in the MIT computer science building had written at least a chapter for Morgan Kaufmann. "A real gentleman," was how one MITer described Morgan.

My deal with Morgan Kaufmann

Lower Klamath National Wildlife Refuge My contract with MK is short and it doesn't include a clause that says I can't tell you what it says so I think I'll tell you. They get the print rights; I keep the rights to publish the work on my Web site. They pay me 10% of net on the first 18,000 copies, 12% on the rest (if any). I didn't have an agent and I didn't ask MK to make any structural changes. Yet there are no conditions under which I get paid a lower royalty rate. Nor is there a penalty clause for being late with the manuscript, though I imagine most of their authors must torture them by missing deadlines. I get 50% of whatever MK gets if they sell the rights to a foreign translation (or movie; just imagine!); the big companies' standard contract would have given me 10 or 12%.

MK includes the evil indemnification clause that I complained about earlier in this story, but it isn't as evil as the standard big publisher clause. The standard clause makes you promise not to say anything libelous and you have to pay for a breach of this promise or an alleged breach. So if the publisher gets sued and eventually wins, you still have to pay their enormous legal bills. I only have to bail out MK if they get sued and lose, i.e., if what I wrote is found to be libelous in a court of law.

That's pretty much it. The contract is 6 pages long and probably about as short as it could be and still lay out clearly who is responsible for doing what and when. I could have gotten a bigger royalty percentage if they weren't going to take a huge risk and print this thing on a four-color press with expensive paper. But I'm curious to see what happens and the money is going to be insignificant either way. Oh yes, the advance. They didn't offer one! Most of MK's authors have day jobs and don't need advances to pay their grocery bills. This worried me at first, mostly because I figured they could easily walk away from a project in which they didn't have any cash invested. But really I don't care whether I get money now or two years from now. I'll probably end up getting as much cash or more than I'd get from a big publisher and I expect the book will be much higher quality. It will also reach a higher quality audience since MK's existing readers tend to be very well educated.

The German Edition

Soviet memorial to the victims of the Nazis, on the site of the old Jewish cemetery in Berlin (dug up by the Nazis) Interleaved with all of this is a translation of the old book by Dr. Olaf Borkner-Delcarlo ( Macmillan sold the German rights to Carl Hanser Verlag, a belletristic publisher that is apparently tired of printing Umberto Eco and wants to do something with true literary value. I didn't even find out about this deal until Olaf sent me some email! He turned out to be a very interesting guy with a strong CS background and a literary bent so probably the German translation will be better than my original.

July 1998: I finally got my hands on the translation (thank you, Olaf). The Germans cut the funniest part of the book: the acknowledgments. They cut the napking drawings, the most aesthetic part of the book, and replaced them with ordinary tech-book drawings. They cut the color photo insert (no big surprise; it was expensive and didn't add much). They have lots of footnotes explaining things like "NYNEX is a big phone company". I keep it on my coffee table and my friends get a kick out of it.

Copy Editing with M-K

Copy editors must think that all authors are morons since, from their perspective, we keep making the same mistakes and never learn. All authors, on the other hand, think copy editors are ham-handed. So that I don't vent too much rage at Morgan-Kaufmann's copy editor, I'll put the funniest corrections here:

The New Paper Book

It came out of the bindery early in April 1999. The final product is a tremendous success, mostly thanks to Yonie Overton, the production manager. Yonie did the final copy edit (she didn't introduce any errors and found lots of mine that the real copy editor missed). Yonie camped out at the printer and did quality control. Yonie got hold of 70 lb. glossy paper (this is the heaviest 600-page book I can remember).


During its first week on the market, the new book soared up to number 103 on I was all excited and ready to email my friends but a day later it crashed back to 250 when Amazon cut the discount from 40% to 20%. If this book fails, I won't be able to fault Morgan-Kaufmann's marketing machine. They shipped out 200 promotional copies within days of the book coming out of the printing press. Sheri Dean showed me plans to schmooze the bookstores, run ads in magazines, take the book to conferences.

So that nobody can say I ended an article on my Web site on an upbeat note, let's look at the best case scenario. Suppose that over the next six months Morgan-Kaufmann sells every copy of this book that they printed. And sends me a big fat royalty check reflecting our joint hard work over a 24-month period. The check would almost equal... one day of revenue for my company,

June 1999: The Slashdot Effect

Ellen Spertus, a computer science professor at Mills College, reviewed the book on Slashdot. The result? A sales rank of 92 on Amazon:

Reader's Comments

Thanks for the insights, and for illuminating the publishing process! The humor and irony are priceless, and it's really comforting to know that someone else finds the garbage just a frustrating as I do.

-- Eileen --, July 20, 1997
As someone in a technical job making way more than a book publisher would ever pay him, I have to say that I've been through the same situation. Unfortunately, I wasn't as persistent as Philip in insisting on a skinny book. (I had 200 pages of material, they insisted on 400 filled with illustrations).

Unfortunately, I also had a day job that I refused to put down for the book, so eventually missed enough deadlines ( mainly because of the insane page count) that the book was cancelled.

A sobering experience, but a lesson well learnt. I wondered whether getting a book published would make me feel better, but now that I've seen what Philip's been through, I doubt it.

Congratulations, though. I'm going to buy a copy just out of sympathy!

-- Piaw Na, August 8, 1997

Reading "The book behind the book..." answered most of my burning questions, such as, How did a book with a cover like this (i.e., not a cute animal) get to be so good? I can't help wondering whether the book would have been more successful with O'Reilly--it seems to me that philg's audience would zero right in on it that way, while I consider it a minor miracle that I found the book (in my local Waldenbooks no less!) at all with ZD as the publisher. OTOH, I suppose the O'Reilly folks would want to push Website, eh?

Regarding blurbs, they had no net impact on my buying decision: if anything, seeing Dertouzos listed on the front page gave me pause. Robert Thau's blurb should have been featured prominently, in large type, instead!

Why did the book work so well for me? It's a lot like sitting down with your favorite nerd friends over a cup of coffee and getting the unvarnished stories of their trials, tribulations, and hard-won wisdom. In my experience, that's how most people learn and make decisions. Reading the book was the next best thing to meeting Phil and getting his story. Maybe even better, because I would surely lose the URLs I would scribble down on pieces of paper...

Database Backed Web Sites is a great book--right up there with Sendmail and Programming Perl. Maybe even better, though it doesn't seem fair to compare them. Thanks, Phil, for putting so much of yourself into the book and giving us the inside story, too.

-- Rob Szarka, August 15, 1997

Thanks for a great book, and mostly for illuminating how most computer books are written by people learning the topics contained, something I've always suspected :)

I also like the above comment, were your book is compared to Programming Perl, a book that you denounce somewhere else, must make you feel warm and fuzzy.

-- Barry Robison, August 26, 1997

In the 80's I used to frequent the rather sparse nerd sections in book shops. The last McBook I bought was "The Complete C Referance", and it was. In the last 3 years I have regularly hunted through the mamoth editions of JAVA+++++, NT FOR LOSERS, etc in the vain hope of finding some useful information. I was worried that I'd fallen behind in the industry, as I can't be bothered reading the mindless bland crap which is being dished out. eg "Functions of EXCEL menu icons" Duuh Thank you for sharing your views on this issue. I will now continue to put my moral being at risk by sifting the uncensored, unsanitised information on the web. By the way, I assume you have read from John Steinback "Travels with Charlie" on the above subject.

-- Malcolm May, September 17, 1997
It is not suprising that you are unsatisfied with Macmillan's promotion. They don't know a good book from bad. I don't consider books published under the Ziff-Davis, Waite, and other Macmillan brands because they are almost universally "by idiots for idiots." On the other hand, I at least browse most Adison-Wesley and ORA books on subjects that interest me.

-- Jeff Emanuel, October 31, 1997
I agree with your point about computer book authors not distinguishing between the obvious and well designed parts of software and the hard/buggy parts. Some of the senior physicians who taught me did the same thing - telling medical students that all parts of the physical exam are simple and useful when in reality some are very difficult and/or convey little information. Badly designed software can make even brilliant programmers feel frustrated and stupid. In my eperience people who are good at using software applications have often just spent hours of trial and error. We need books like yours that honestly point out the good and bad parts and tell you when something is is not worth the effort.

-- Hamish Fraser, November 26, 1997
Reading these pages reminds one of how depressing the machinations of the "real world" can be. It reminds me of why I often seriously step aside from this electronic world too and want to give it all up. Then I remember how useful and alive it can be. I suspect my reaction is more also against technology and computers themselves. If only there were more books telling it straight, then more people might read them, then more people would see things as they are, then things might even improve! Well, I'll still keep hoping no matter how vain those hopes. Consider this, this type of stand you are taking is more than just about books/publishing/etc it is about the very foundation of society itself (which changes a lot more and more quickly than we like to allow ourselves to think). Now that you've placed yourself in a unique position to make a difference, please don't give up the fight!

Best Regards


aka SnowmanF

-- Si Bloomfield, December 18, 1997

Thank you for posting "The book behind the book behind the book..." I've just finished writing "The Busy Educator's Guide To The World Wide Web" and tomorrow it comes back from the printers. Originally I had two publishers that wanted me to write a book, but after all the restrictions, hassles and differences of opinion, I decided to do it on my own. Your comments about publishing were all so true. I loved reading them.

-- Marjan Glavac, March 18, 1998

It was heartbreaking to visit Amazon after reading The Book behind the book behind the book and finding 40 *GLOWING* reviews of The Book and yet it is Out Of Print.

The Emperor Has No Clothes.

-- Daniel Allen, October 6, 1998

A few days ago I saw a copy of the german translation of "the old book" in a local store. I'm sorry to say it, but the translation is very bad. I enjoyed reading the english online version very much, both for its technical content AND for the entertaining style of writing. The german edition is like one of those funny translation programs for Windows was used. I'm not in the translating business, but in my opinion there is more to translating a book than putting sentence by sentence and word by word into another language. I recommend the english book to everyone working in this field. I do not recommend buying the german edition.

-- Ralf Eichmann, November 20, 1998
Sorry to say I don't do database stuff, and haven't even read the web version. But one thing I know you're good for irrespective of the topic is a hearty gut laugh. Thanks, Philip.

Chris Grantham:

-- Chris Grantham, March 28, 1999

One big advantage of dead-trees publishing is the ability to autograph one's work. When I was in Boston, I picked up a copy of _Philip and Alex's Guide to Web Publishing_ from a rack at the MIT Press bookstore because it was advertised as "Signed!"

Imagine my disappointment when I cracked the book open on my flight back home to San Francisco: Philip had signed the book but Alex hadn't. I suppose it's typical of Philip's egomania that he thought that this would be sufficient.

(I was glad to have the book despite this. It gave me more good ideas for web applications than I've ever had on an airplane before.)

-- Lawrence Hosken, May 11, 1999

I've been an avid reader of computer-related dead tree carcases for a while, and this article has changed my perspective. I've read the "Dummies", "Idiots", "Dolts", "Jerks", "Bozos", and "Knucklehead" series of books. I've gotten quite a bit of education from them.

Yes, they're silly, and dumbed down, but most of them touch on the basics, enough to give an intelligent reader enough of a start to do some experimenting. However, most of the productive thigns I've accomplished, I developed on my own, or through other resources.

I typically spend $35-$50 for one of those ten pound wonders. I figure the author is making at least $5-$10 of that. Silly me. Now I know.

What do I expect for this $35-$50? Not much. the good part is that I usually receive exeactly that. I figure from each of these 600-1200 page behemoths, I learn 2 or 3 useful tricks. That's about it. At least I can later use these tricks to impress my clients and make a few dollars, hopefully getting back the cost of the book that taught me these tricks. Is this true of every one of these books? No way! Just an average. Some of the books were completely worthless, and others (rarely) are gems.

(I've always wondered why I can breeze through a 600 page computer nerd book in a few days, but a James Michener novel of the same size takes weeks to read. Again, now I know. Michener doesn't use 300 irrelevant screen shots and cute guidance icons, nor does he resort to 1.25 inch gutters and 12 point type... Sneaky. I'll bet there are a lot of really pissed off trees in the afterlife, knowing they gave up their precious fiber for some of the crap served up as computer books. And, no, I'm not an environmentalist. Just a realist. Hug people, not trees.)

I am reading "Philip and Alex's Guige to Web Publishing" right now (online). Philip has oopend my eyes, and at first I was going to simply send him the $1.50 and save a tree. On the other hand, his new publisher seems to be genuine about publishing quality, so I may support that whole chain by buying the paper version.

Either way, Philip, you've given me a far better education in the fundamentals of the web than the three bookcases (and several hundred pounds) of dead trees in my office have. Thank you.

-- Chris Mospaw, May 23, 1999

It didn't occur to me that second edition was just published. Anyway, last Sunday at Berkeley, I walked into a bookstore and seen a familiar photo glancing at me from the shelf. It's like meeting one of your favourite web pages in person. So now I have it, still reading it (3 hours past midnight, ouch!), and truly enjoy it. Great work Philip, you're a real MIT hacker.

-- Dmitry Kohmanyuk, May 26, 1999
I've been a fan of for a long time...First it was Philip's photographs, then his expert guidance and articles. Later, I stumbled onto his web-engineering work (including the online book).

I started reading the book one afternoon while waiting for a compile to finish. I kept reading. I let my project slip. When my wife arrived to drive me home that night, I made her wait while I printed out the remaining chapters. I read all night.

I pre-ordered my tree-carcass copy from Amazon that night.

The next day (in fact most days since then), I've pushed this book on everyone around me who thinks they can or want to design web pages.

Of course, I read the "book behind the book" story and was distressed at the treatment Philip's first book received.

How exciting, then, to flip over to Amazon just today (10 June, 1999) and see that the book is ranked #570. It's good to know that a well-written book by someone I respect can climb to a respectable spot without appearing to sell out.

Kudos, Philip!

-- Dan King, June 10, 1999

Just so this section doesn't entirely reflect a spirit of happiness and light ...

I've pushed this book for the past two months to everyone I know in the computer and web business. No one (execpt for me) has bought it.

I've mentioned the web site to all my computer and photography oriented friends. Only one has actually visited and used it to any extent, and most of his contributions center around darkroom and enlarging techniques -- a topic that does not really cover.

I followed the link to Creating Cool Web Databases just so I could see what a bad book on this topic looks like. The customer reviews on were either "great" or "bad." One of the "great" reviews posted the name of a web site built using the techniques espoused in that book. The result seems fairly attractive and workable.

I wanted to use Philip's Uptime service and software to monitor my web servers both locally and remotely, but it seems that there is no easy way to get them to pull a page from an SSL encrypted link. I wound up "rolling my own" software to monitor those servers.

I read and enjoyed a fair amount of, and, in gratitude, I decided to make a donation to Philip's favorite charity. In return, Philip sent me a gift of couple of lovely prints. Now a financially astute friend tells me that the contribution is not deductable since the value of the gift (based on the quality of similar photographic art) may be more than the contribution.


Still, I bought the book, read it, enjoyed it, and learned a lot from it. Highly recommended. Even more highly recommended is the web site: it's great. However, you're here already, so you know that.

-- Frank Wortner, June 14, 1999

Here is a description of my interaction with another publisher. For some reason they reinforce the notion that Publishers believe their Customers to be absolute idiots who are not to be cared for under any circumstances.

How Not To Design Web Sites OR How Not To Ignore Customers

I bought the book "The Practice of Programming" by Brian Kernighan and Rob Pike. I have liked reading Kernighan and Pike ever since I learned Unix from their famous book of a while ago -- "The Unix Programming Environment".

At the back of the book, there was a URL that promised to have a site dedicated to the book. The site was To call it hideous would be a gross understatement. All around the start page, there is a perimeter of useless graphic buttons about 3 inches thick. The meat of the matter is buried in the middle. The page took about 5 minutes to download over my modem line.

Along the lower boundary, I see an error message "No matching titles (catid() is empty)". As a buyer I wonder what catid() is and why it is relevant to my mission of learning more about this book, and finding a way to discuss it with the authors. I find it difficult to read a book if I do not first understand what message the authors intend to convey through their work.

I first sent off a missive to their webmaster. I got the webmaster's address from the "CONTACT US" button. It said that comments about the website were to be directed to The email I sent was:

=====BEGIN===== As a long-time customer of Addison-Wesley, I must tell you that your web site is dreadful. Here are some problems I see:

(a) Way too many graphics. There are way too many graphics elements all over the place. If your web designers work on a LAN, it might seem very good to them. Ask them to try loading their own web pages over a 28.8K modem line. None of these graphics elements (other than the cover photo) actually add anything to the user experience. They are totally superfluous and only serve to annoy potential customers.

(b) Broken Links: I tried searching for "The Practice of Programming", a recent AW release. The link your search engine sent me was broken. =====END=====

About a day later, I got an autogenerated email. It said that the account was getting far too many emails for them to process, and gave me a list of email IDs that I could try sending email to. Notice that they did not bother forwarding it themselves -- they asked their customer to retry the email.

I then thought that 'corporate communications' was a viable place to send email to. Accordingly my next email was to Imagine then my consternation when I received an email from a certain 'Susan Stockman' which identified itself as an autogenerated message and admonished me not to respond. Apparently dear Ms. Stockman no longer works at Addison-Wesley.

By this time I had tired of sending emails. I decided never again to buy a book from a publisher that was so careless about customer service. First I am made to see an utterly clueless web page for a new book written by highly respected authors. Then the webmaster says that he/she is too busy to receive my feedback. And then I receive Susan's email.

A web site for a book should have the following:

A more elaborate discussion with the authors than a preface in the book would allow Contact information for the authors No graphics unless the graphics are somehow relevant. As a practising programmer, I don't have patience or time to download useless graphics A place to discuss the merits of the book (or lack thereof) in a free and frank manner A collection of hyperlinks where I can find more information about the ideas expressed in the book

I think AWL did itself a major disservice by printing the URL of the book's website on its back cover. Potential buyers might see the website, and dismayed by its poor quality, refuse to buy what is a fairly decent book.

-- Jagadeesh Venugopal, June 28, 1999

I read Philip's commentary on creating the book about six months ago, and it answered a lot of questions I had about computer books. In particular: why are they so useless? Since then I have thought of this commentary many times when I considered buying technical books.

For the most part there is no way around teaching yourself new skills by looking up related specs, rfcs, web pages, faqs etc. and then just writing practice code. Sometimes you need a broad overview, in which case a dead trees book is the right thing. In general you are out of luck though - there are few well written overviews of this kind. Recently I found one of the rare winners, "Java Cryptography". Not remarkably, the tone of this book is almost identical to Philip's.

In general, the value of Philip's writing is in his tone. Informed, not condescending, no hyperbole or salesmanship, thorough, honest, blunt. That is what technical writing is all about.

-- Lucas Gonze, September 20, 1999

I recently arrived in Korea from the states, and have quickly become jaded with the whole military hangout scene composed of beer-can-on-the-forehead, tobbacco-dipping meatheads and whore house strips posing as topless clubs. To keep me from getting depressed, I've resolved to use this coming year for learning. Luckily, even though the phone lines don't work half the time, the camp library's computer section is useless garbage from the 70's, and I was told at the video store, which is replete with nothing but various MS 2000 books, that it would take four months to get "JAVA for Dummies" (I feel silly about trying to order that now that I read this page), I have a dedicated line to the internet 24/7. This provides a medium for acquiring the information and resources that I need to build my on-line magazine, my learning project for the coming year. It was such a great surprise finding Greenspun's site, after weeks of fumbling through cyberspace trash trying to find anything useful. I read "the book on the book..." to see if I would be interested in reading the book, and now I am not only interested, I am looking forward to it. I have a lot to say but I'll keep it short....I feel better having read about Greenspun's experiences with book publishers. Some years ago I tried publishing my poetry to no avail. I got replies from all the publishers saying they were not interested. One even sent a letter with some hilarious poems featuring the god Thor and some Hai Ku (useless, IMHO) which was meant to be taken seriously. This was discouraging given that I had spent a good chunk of what should have been college exam study time organizing and putting my poetry in WordPerfect 5.1 (does anyone know how to get a copy of WP 5.1, BTY? I had no foresight to save the files as text...Word does a bad job of transporting these files). I haven't written but two or three poems since then. It is even more discouraging to find Jewel's poetry, which reminds me of my own (don't laugh), on the front of the bookstore displays. Reading this page, however, has raised my enthusiasm for my decision to publish it on my own. Distributing art, which should be free, through established, mechanized channels is a contradiction, as is publishing a work about the web on a static, non-interactive medium. The most important people are my friends and family, who are spread out all over North America, and if they are the only ones that read it, it's o.k.

As I was saying, I am glad I found this site, which is just what I need to get me started in the right direction. I feel like I can learn from this guy without snoring up a storm. I hope his books are as insightful and entertaining as his articles on photography and book publishing. (I also liked The Game, which I scored something like a negative thirty-something the first time and won on the second try. Lesson learned: lie to blonde bimbos when you want something from them...actually, I knew that already, it was just the first time I got any type of result by doing so). I would be willing to buy the ZD hard copy of the Web Whore (that's the title as far as I am concerned) with the bad cover and napkin graphics (after I read the on-line book), just to have around as a symbol of Individual Expression vs. Infrastructures of Greed. In time it will probably be a collector's item anyway.

-- David Guarneri, December 17, 1999

You mention that you did not want a cdrom included with your book. For those of us with slow net access (33.6 modem) I'd go buy your book to save me the 100MB download for Oracle8i. (Assuming Oracle would let you include it with a cdrom). Also, I have found Oracle's web-site to be one of the hardest to navigate. It took me about an hour to find where to download Oracle8i for Linux, maybe they ought to read your book.

Now the praise: your book, webSite and ideas are all very good. Keep up the good work. I'll be spending the next few weeks devouring the content of your pages.


Image: HarryOrg.jpg

-- Mick Timony, January 9, 2000

Everytime a friend or acquaitance talks about the possibility of writing a book, I refer them to this excellent source of infomation on what it is like to deal with publishers or editors from an author's point of view. There is nothing like honest first hand experience to make them aware of the possible pitfalls.

Steven A. DuChene

-- Steven DuChene, January 9, 2000

What a magnificent attitude! The level of technical expertise, future insight and wry sense of humor, actually caused me to stay up for 2 hours to read "the book behind the book." Your humor, both in person and on the internet, is refreshing, very entertaining and actually keeps me inspired. ArsDigita and both are great examples to follow. If you have a chance, see Phil LIVE AND IN PERSON. You will get to meet Alex and if you are lucky, you may even be the guy from Harvard who lends Phil a lazer pointer while being surrounded at MIT. I enjoyed your lecture and continue to enjoy the pages and humor you and your HP serve. Thanks Terrence Boylan

-- Terrence Boylan, February 17, 2000
Having read the first few chapters of your book online, I like it a lot. I also found the story of the paper version fascinating - if I'd been thinking of trying to write technical books for a living, you'd have saved me a lot of trouble!

One point regarding, whom you mention frequently. I've bought books from them in the past, but due to their abuse of the patent laws, no longer do so. I understand that not everyone regards their behavior as abusive, but many do, and I think it should be pointed out so that people can make their own decision on the matter.

-- Russell Wallace, June 19, 2000

I am going to keep my comments short this time.

In November 1999, I accidently stumbled upon while search for "SQL + Oracle + database + websites + help" on Yahoo! After spending more than three days nonstop trying to crank through all the readings(there were more to than just Phillips and Alex's guide to Web Publishing), I have decided to buy the hardcopy.

I went to the local Borders Book Store to find Alex's book(or Philips) before attending the one-day bootcamp in Berkeley. After searching relentlessly through every single section, I finally asked clerk for help. Together we found the book (i.e. one single copy) under the Web and Internet section. The outcome was no better than the B&N bookstores.

And that was only half of the story.

After the slash-dot effect, I went back to the local Borders bookstore again because my brother is flying back from Carnegie Mellon University, and I wanted to give him a book that will enable him to build a successful database-backed website, hire a Stanford M.B.A. to be the CEO, go public, make his millions, all in the time it takes for a non-geeky guy to have a decent relationship(one year and a half). After all, I am his brother.

Once again, I can't find Phillips and Alex's guide to Web Publishing. I went to the aisle, and asked politely, "I was wondering if you guys have a book called Phillips and Alex's guide to Web Publishing." The clerk fooled around on his terminal a bit and answered, "Yeah, we used to have this book. It's sold out. Do you want us to order it for you?" I did not want to wait. I was thinking about going on fat brain to get the book anyway. However I was amused someone else is interested in Philip's book. So I asked the clerk, "you know... I bought the same book not too long ago. I was wondering when was the last copy you sold?" The clerk answered, "we sold the last copy on November 26 last year."

It turned out I bought the last copy in the store 8 months ago, and they haven't order since. So much for the slash-dot effect.

-- John Lin, July 16, 2000
a student put me onto this saga, and boy do i wish i'd read it about 5 years ago. even then it was clear that direct-WWW publishing was a viable alternative and it is more true all the time.

my book is Finding Out About, which i have been trying to push thru Cambridge Univ Press. for my detail, see

anyway, misery does love company. authors unite!

rik belew

-- Richard K. Belew, October 24, 2000

a kind of a coffee table computer nerd book.

Yes it is and very enjoyable too — after reading the web-page I went straight out and bought the book. Which proves that online self-publishing and dead trees can coexist.

But in that context not enough of a coffee table book in my opinion. The text is available on-line, and in any case the content is to be valued for more timeless qualities that a technical reference; so what I am paying for is the pleasure of holding a well made book in the hand.The cover is too thin and the pages are ‘perfect’ bound. A real coffee table book would be cloth bound at least. Even if Morgan-Kaufman can't be persuaded to do that the paper binding could be better. (it dosn't make a significant difference to the production cost, its a marketing problem; the public think the paperback should be cheaper.) For an example consider Photoshop 6 for Photographers (Martin Evening, Focal Press, 2001) same format and similar size, same need for quality illustration, similar price, but it is bound in signatures and has a heavy cover.

In the future people will go to the web for content, but continue to buy books for the pleasures of owning artefacts: but only if they are worth keeping.

-- David Clark, August 20, 2001
Bravo! I co-authored a book a couple of years ago about Oracle running on NT, and ran into the same clueless incompetence on the part of the publisher. People seemed to like the book (or at least they said they did), but everything you said was true (e.g., the desire to fill up pages with screenshots, the carping about deadlines, the revisions that made no sense, Word templates that felt more like a straitjacket, the wasteful CDROM, etc.).

One of the first things they said prior to me signing the contract was that the book needed to be at least a couple of inches thick, so it'd look good on the bookshelf next to all the other books for sale. Sheesh. They used really big type, lots o' screenshots (against my better judgment), and even the paper the pages were printed on was extra-thick!

Now the book's out of print, which is probably for the better. I'm happy to have helped some people, but the finished product was far inferior to what it could have been.

That was the first and probably the last book I'll ever write. And I didn't get paid nearly enough for my trouble either. They did give me a nice Cross pen when I finished - yippee.

-- Mike C., January 25, 2002

When I saw the Philip and Alex book in my local Waterstones bookstore, I thought to myself, who is this arrogant guy who thinks his book should be filled with strange photos and not screen shots and charts. A week later my curiosity got the better of me and I bought it. Philip, I've read your book from cover to cover 7 times ( and even a couple of times I've re-read my favourite chapters on the web at work), it really is that good. A few months ago I sold half my useless old computer books on ebay, when I get round to selling the rest I'll do it on the website I built using the ideas and technology you introduced me to.
Image: backyard.jpg

-- John Holroyd, April 14, 2002
I saw the comment trashing Addison-Wesley's site for The Practice of Programming. I don't know how many of the past four years it took to fix it, but the site is now ... ok.
The links all link, the code samples are there, the graphics are just the usual commercial boilerplate text-in-images, dumb icons, corporate wallpaper, and the book cover. There is a table of contents, part of the preface, and a sample chapter (I). There is a Web Resources item that points to the authors' own book page (the url I give above), which is only 2 clicks from actually emailing the authors.
And this comment may be worn thin with overuse, but I did actually hear it, and it does summarize pages of experience:
[Novice Tech Author] I am going to write a book.
[Experienced Tech Author] Oh, you want to make dozens of dollars!

-- Ephraim Cohen, October 17, 2003
Let's compare average lifetimes of CDs and URLs. If you'd put the CD in, it would still be working today, unlike your URL. If you'd given the keywords needed to track the page down with a search engine, that might have helped (c.f., but I'd still prefer the CD.

-- Tom L, January 25, 2004
I will leave most of what Simon Hayes said to be responded to by Dr. Greenspun ( and hopefully he does ). However I must say that the conversation he replicates sounds more like one of those episodes of a TV show where the characters fight over what happened and give their perspectives tainted by what they want to believe.

His representation that "books without CDs in the back" was an idea way ahead of it's time, ( so there were no books published before CDs were invented? ) seems strange, given that virtually no book published by Addison-Wesley at the time had a CD in back. Many of these were considered "must read" books serious preogrammers. Authors include James Coplien, Scott Meyers, Bjarne Stroustrup. One of the books was "Design Patterns".

Finally, when he decided to compare Dr. Greenspun to Scott Hacker ( sounds like the programming equivalent of a porn name to me ): "There were some amazingly good other authors at that party, including Scott Hacker, who turned out to be a fabulous author and friend and went on to write some equally wonderful and far better organized technical books, also quite ahead of their time, ", I decided to check him out.

I couldn't remember any books by Hacker, so I checked Amazon where he was hard to find ( Amazon does not have a good Author search so I got hit with 5000+ books ). I then checked with Illinet, which is the online catalog of all books in most Illinois university libraries ( the notable exception being the University of Chicago ). It's URL:

I couldn't find even one reference to a book he wrote. Not one book in any of the libraries of most universities in Illinois? Strange for a writer "ahead of his time".

-- Thaddeus Olczyk, May 23, 2004

I've since come back to writing An Engineer's Guide to Silicon Valley Startups, and handling entirely myself. This lets me keep the page count low, not clutter up the book with junk, and it turns out that it makes more than a few dozens of dollars. The entire experience is also much more satisfying, since you actually get positive feedback about the book. So much so, I'm writing Independent Cycle Touring the same way. It turns out that you can make money writing and selling books, but you have to self-publish.

-- Piaw Na, June 8, 2010
Add a comment

Related Links

Add a link