Unknown HTML tag test page

Motion, even in the direction of the horizon which presents itself to a person standing erect with open eyes facing forward, is not necessarily progress.


Below are some thoughts I had about XML during the summer of 1998. I still think they may be worth reading and express valid concerns. After attending the Graphic Communication Association's XML 98 conference (Chicago, 14-20 November 1998), however, I feel that the issues are much bigger, and that my earlier concerns and thoughts do not point in the direction(s) where limited energy will likely be most productively expended -- although I still do believe it is important for grandmothers to be able to author their own web pages.

After the conference, I no longer see XML simply as "dumbed down" SGML. I now see XML itself more as a "hook" on which a lot of powerful and complex supporting technologies will be hung -- like a pile of garments way overloading a wobbly coat rack (could these technologies have been hung on the SGML hook? For "political" reasons, that now seems an idle question). I now anticipate that XML will likely become far more complex and difficult than SGML, when considered in the light all the adjunctive technologies and standards surrounding it (e-commerce technologies, XSL, Namespaces, Xlink, Schemas of whatever kind, Link maps, XIML(?), GKW(?) etc.).

These developments (e.g., schemas replacing DTDs) may result in XML becoming far more powerful than SGML ever was in practice. Some of the challenges XML is tackling (even such seeming details as handling non-Latin character sets, e.g., Kanji) may immensely add complexity. There may be "consequences" (as in, e.g., "unanticipated side effects"). Beyond observing the truism that complexification should itself be cause for concern, I have no idea where XML is headed in its impacts on society, technology or technological workers. I am, however, now pretty sure that the things about which I wrote below are not, as I thought when I wrote them (except for that concern about grandmothers!), "key issues" even in the "regional ontology(ies)" directly involved, not to mention more inclusive humasn[e] perspectives.

Also, I would like to respond to a snide remark about HTML, made by one of the XML 98 conference presenters. This person denigrated HTML as "just anything Netscape or Microsoft decided to support in their browsers, and that HTML web pages were never validated in any way". I agree that HTML is only formatting markup, and it is obvious that even major corporations' and academic institutions' websites are generally composed of pages which don't even have "doctype" declarations. However: almost all the HTML pages on this web site (http://www.users.cloud9.net/~bradmcc/) are validated, and the pages that are not generally have good reasons for being that way, along the lines of another conference speaker's "defense of invalid SGML" in situations where the invalidity of the marked up text -- as in the present page -- is part of a point made by and/or the use to be made of the text.


[ Get Internet Explorer 5! ]May 1999: New development. Microsoft's Internet Explorer 5.0 supports XML pages. I have some sample pages which you can check out if you have IE5: xmlcsst.xml (uses CSS), xmlxslt.xml (uses XSL), and xmlisland.html (uses an XML "data island").

This page tests what Netscape 3.04 (and other web browsers...) do with unknown markup (e.g.: <gkw>?</gkw>). It suggests a way that HTML could be extended with content-describing (SGML-type) tags, that would grow out of existing HTML coding, rather than replacing it with something else (esp.: XML/XSL).

This would allow grandmothers and others to continue to produce their own web pages without needing to use web-page construction programs (e.g., Microsoft FrontPage), while still permitting enhancements to the structural articulation of the text. Both format and structure describing tags would coexist, instead of formatting being split off from content descriptors in a separate stylesheet that uses a new arcane syntax.

I would urge: Where there is serious concern for structural markup, use real SGML (SoftQuad Panorama, e.g., can be used to format it). XML is neither grandmother-friendly (like simple HTML), nor "industrial strength" (like real SGML). Why introduce new compromised clutter that doesn't really satisfy anyone's needs but forces everyone to learn and do something other than what they're accustomed to, into an already messed up world that has serviceable ways of accomplishing all the relevant purposes?

[Vice-president of marketing, P.G.] Bartlett at ArborText [a key supplier of SGML and XML software] has no doubts about what's going on. "XML," he says, "will prove to be one of the top ten technological innovations of the first century of computing." (New Scientist, 30 May 98, p. 37)

If a "dumbed down" form of SGML is really needed, it is already available in the form of a person using only a limited subset of SGML's existing functionality. If this is "too difficult" for people, maybe the problem is not the difficulty of the SGML itself, but the inadequacy of documentation how to use it. (The inventor of SoftQuad Panorama, Yuri Rubinsky, had a vision of "SGML on the WEB", which, alas, died along with his own premature decease.)

To see the source code for this file, with its undefined markup, please click here. If anyone wants to go beyond HTML in substantively empowering ways, I would suggest they start with:

  1. Figure out a way to implement CGI scripting that would enable it to be a normal part of everybody's website (e.g., GeoCities and AOL homepages). CGI is a lot more powerful than a lot of more confusing and browser-dependent workarounds, like Javascript and Java. (Do computer science wizards shy away from this challenge because it is too difficult? Or is such a service to the user community beneath their self-image of suitable employment of their skills?)
  2. Implement a multiple-target hyperlink construct. E.g., clicking on: <a hrefs="http://www.nytimes.com/ 'The New York Times', target2.html 'Some web page of mine' ..., ftp://more.blahblah/targetn.ftp 'Somebody else\'s ftp site'">let me choose where I want to go</a>, would popup an annotated menu of destinations from among which the user could choose which one they wanted to link to. (Note that SGML HyTime implemented this multi-target linking facility; I once maintained an HTML application that implemented multi-targeted links thru JavaScript -- see screenshot of example multilink, below.)

Thank you for considering my thoughts. Your thoughts? [ Email me! ]


Go/Return to SGML document introducing you to SGML.
Go/Return   to another intro to SGML: *Darwin Among the Machines* (Susanne Langer and SGML).
 
Read Laurent Sabarthez' notes about XML (SGML's "replacement").
[ ] [ Learn about SGML! ]
 
What's new on this website?
Go to website Table of Contents.
Return to Brad McCormick's home page.
Return to site map.
[ ] [ Go to Site Map! ] [ ] [ Go to website Table of Contents! ] [ ] [ Go home! (BMcC website Home page!) ] [ ]
[ ]

http://www.cloud9.net/~bradmcc/gkwtags.html
Brad McCormick, Ed.D.
bradmcc@cloud9.net
10 April 2006 (v01.29)
This page is not
valid HTML 3.2, 4.0
or anything else!
 ]
[ ]
[ Multilink example: Clicking link pops up menu of URLs to go to ]
[ ]
Example multilink. When you click on "Native American Indians", a window pops up with a list of places to go to.
[ ]