We Have Chosen Shame and Will Get War.

by Philip Greenspun

Site Home : Research : One Article

"Owing to the neglect of our defences and the mishandling of the German problem in the last five years, we seem to be very near the bleak choice between War and Shame. My feeling is that we shall choose Shame, and then have War thrown in a little later, on even more adverse terms than at present."
Winston Churchill in a letter to Lord Moyne, 1938 [Gilbert 1991]

It all started as a paper I wrote for the Second International World Wide Web Conference, back in August 1994. I explained that "We could have taken a formatting language and added hypertext anchors so that users had beautifully designed documents on their desktops. We could have developed a powerful document structure language so that browsers could automatically do intelligent things with Web documents. What we have got with HTML is ugly documents without formatting or structural information."

This sort of sentiment was so popular with the powers-that-were at that conference that they rejected the paper in favor of a bunch that said "HTML and the Web are so great. I downloaded NCSA HTTPD 1.3 and then got some documents and now I get 5,000 hits every day..." A paper about the Web's deficiencies was about as welcome as a guy wearing a "living with AIDS" T-shirt would be in a singles' bar.

However, the same Web that irritated me enough to provoke a scream of agony against HTML's limitations freed me from having to depend on a conference program committee to get my message out. Since Travels with Samantha was getting 100,000 hits/day, the spillover from there into my research area was actually good for more exposure than my paper would have gotten from the conference proceedings.

The Journal of Electronic Publishing published the screed first. Then SoftQuad came along and put it on CD-ROM. All along, people sent me mail every few weeks with new ideas inspired by my writing.

Anyway, you probably came here for the paper. Here are your choices:

As long as we're talking about stupidity...

I might as well note that I'm forced to link to a special pseudo bboard system from here as a workaround for the moronic robots.txt standard.

Reader's Comments

1.choose a set of 100 documents that represent the spectrum of things we'd like to see on the Web 2.come up with a language capable of expressing the author's intent in 98 of those documents 3.come up with a language capable of expressing the designer's intent in 98 of those documents 4.add to HTML the semantics of the languages developed in the preceding two steps >>

The reason you will likely never see such a logical approach is that there are not enough multidisciplinary designers of HTML. By that I mean, the people who develop the code are not always those types who are actually familiar with the way published documents already look in paper form. (And it takes a really creative type to imagine a new type of document....one that isn't published in paper form...)

May I say, however, that your idea is extremely logical. For that reason alone, no one will probably pay any attention to it whatsover.


P.S. I am a librarian.

-- J. A. Stevenson, April 29, 1997

I completely agree with what you're saying, and it now seems that W3C does, too. The XML standard is approaching fast (MSIE 4.0 ships with two XML parsers), and does exactly what you want, especially combined with CSS/DSSSL. More info at http://www.w3.org/XML/

-- Lars Marius Garshol, October 25, 1997
I was going to mention XML, but the previous commentor did already. From what I have seen, people who propose using XML often get excited about creating DTDs. For general purpose use, I suppose that they are necessary, but for a particular field of endeavor, they should not be, so far as I can tell.

I believe that an XML-based data interchange can have the following properties: 1. No DTDs need to be produced or exchanged. 2. The order of the tags does not matter; data elements in a hierarchical data object are effectively "tuples". (Normalization may force a particular order though.) 3. A data consumer can accept data from a producer even when they do not both understand the exact same set of tags. If a receiver does not understand a tag, that info is simply not available. 4. If a receiver sees an unsupported tag which is above the leaf level in the hierarchy of the data object being parsed, it is still possible to read below that level for other tags that are understandable. 5. If a producer is given a request formatted in XML which omits a level of the hierarchy, it is still possible to fulfill the request. 6. Versioning and updates are not required.

Perhaps I should give some context: I envision a system for exchanging data, where both requests and replies are formatted in XML as "data objects" -- hierarchically structured items that carry their own context by virtue of an agreed on meaning assigned to tags. Three years ago I worked on a project to develop a standard very similar to XML for to the Real Estate industry, known as DxM, the Data exchange Methodology. It died because most people thought that EDI was "better". (And of course XML did not exist at that time.)

-- Scott Rowe, June 16, 1998

Suppose we grant that everything Phil has said about HTML's shortcomings is true, and that even the relative win of combining XML and CSS is less than nirvana. Even so, the true stumbling block is not the specs; it's the implementations.

I was pretty excited when style sheets were finally implemented in Netscape and MSIE. 'At last,' I exulted, 'I can continue to be purist about using HTML only for semantics and still make my pages look pretty with style sheets.' But, alas, one never can eat one's cake and have it, too...

What's the major problem? IHMO, it's the existence of MSIE 3.x, or, as I call it, The Browser of Frankenstein. Had MS left CSS support out of MSIE 3.x it would have been better than the brain-damaged implementation we actually have to plauge us until the next millenium--or beyond, considering that many of my customers still use Windows 3.1 and Netscape 2.x or even 1.22! As it stands, anything like full use of style sheets seems to get so severly mangled by MSIE 3.x that it would be better to not use them at all.

The only "solutions" I have found are either a bastard mix of non-semantic HTML tags and style sheets or serving entirely different pages to MSIE 3.x. The latter is a pretty poor option, since it defeats one of the major benefits of style sheets: ease of editing.

Even if MSIE 3.x could be obliterated from the face of God's good earth tomorrow, we'd still be left with the vagaries of the current (4.0x as of this writing) major browsers' implementations. If nothing else, why didn't CSS 1.0 call for a minimal feature set in compliant implementations? I suppose MS and Netscape wouldn't have paid it much heed anyway...

Or, to put all this another way: We have chosen frames and shall get more.

-- Rob Szarka, July 15, 1998

I can't agree with your stance on poor browser support for HTML features - what if the HyperTeX program had been horribly broken - that wouldn't have meant the overall design was unsound. HTML 4.0's support for generic inclusion improves over the kluge which was IMG, and PNG solves the downloadable gamma issue. I'd claim, though, that the best case is to provide images with a gamma of 1.0

I like forms, for querying databases and searching for objects. I also like Java, for dynamic content (although I wish Java applets could be allowed to modify a Document Object Model in the browser.

Oh, and your Confirm Comment page should use GET rather than POST

-- Tim Bannister, August 23, 1999
I started writing HTML about six to nine months ago. A structural language? I think not.

I consider HTML to be a formatting system something like a Postscript wannabe but with color and links. I use a "one pixel gif" with HSPACE and VSPACE to provide white space. I use tables (and in some bizarre ways) to format the data so that spacial relations are preserved. I have even created a table which can simulate a dictionary entry (bold words OUTDENTED above the descriptive paragraph(s)).

Speaking of nonstandard colors, one site had blue links on one page. The next page showed up in red. I wondered "But how could I have seen these already? I have never been to this site before!". I found that normal links were red on that page. As an example of a no-no, that one was first-rate.

Some of my pages use the standard blue/red, but I usually color my text dark blue, and links use blue/black. The new and visited links are compatible with the text, but visited links are still colored differently--but less distractingly--than the static text. All links are underlined anyway.

I also take pains to accomodate Lynx users. (Well, because most sites shaft them, that's why.)

-- David Bessey, November 23, 1999

I believe that you predicted well the formatting needs of the future. We have passed through shame and perhaps even war. I thank you for the insight, now onward to HTML 5.

I came here looking for a quotation and leave with a grin, I'm thinking back upon the web in 1994. Thanks!

Also, this might help others that follow: Winston Churchill Quotes

-- Harry F, June 23, 2010
Add a comment | Add a link