- From: Mike Meyer <mwm@contessa.phone.net>
- Date: Wed, 15 May 1996 13:24:10 PST
- To: www-html@w3.org
> From: Benjamin Franz <snowhare@netimages.com> > In the last 20.3 million (actual logged hits since Nov. 1st) on a popular > site I maintain, there have been ZERO hits by lynx 2.5 (75,437 hits by all > other versions of lynx), ZERO hits by UdiWWW (all versions) and 150 hits > by Emacs-W3 (all versions). By any rational standard, features in > supported only in UdiWWW, Emacs-W3 and lynx 2.5 cannot be identified as > deployed. You're overgeneralizing. I can think of at least two rational standards for "deployed" which pretty much ignore the number of hits on any given popular web page. They are: 1) Percentage of web pages that use those features. Especially FIG, as it degrades gracefully on browsers that don't use it. This one isn't very good, because the people who take the time to learn enough about HTML to know about those features will also learn enough to relize why using them is a bad idea. 2) Percentage of AUTHORS that use those features. This is better, because it lets you count people who have a captive audience and can take advantage of features that aren't supported by the mass-market browsers. > "But my homebrew browser no one has ever seen supports the > SPININPLACEWHILEBLINKINGINTWENTYCOLORSINTHREEFOURTIME attribute and it > must therefore be considered deployed. Ne?" Sounds like the percentage of both web pages and authors are going to be zero for all practical purposes, so no. That isn't necessarily true for various HTML 3 features. In particular, people have asked about ID. Since Navigator has (undocumented) support for it, you would have it considered deployed and a failed experiment, and it should vanish from the specs forever, right? <mike
Received on Wednesday, 15 May 1996 16:32:01 UTC