- From: Al Gilman <asgilman@access.digex.net>
- Date: Sun, 13 Sep 1998 11:03:40 -0400 (EDT)
- To: w3c-wai-er-ig@w3.org
- Cc: dwe@arde.com
1. Copied below is an excellent example of a report from analyst to consumer in a complaint-processing cycle. 2. My hopes for the WAI-ER-IG was that we would walk through a "detect, analyze, communicate, synthesize, fix" cycle for actual website problems by hand for a while before we decide which parts of the cycle are stable enough to be done by tools. 3. This relates to my reaction about Dave Clark's question on pass-fail ratings. We don't know if we _can_ give a pass/fail rating without detailed analysis of more examples. This means statistically-significant spread across kinds of websites and disabilities. 4. Once we think we understand the failure modes for a site, when I am talking to the webmaster or website content person, I don't want to limit myself to pass/fail discussions. They may think they want a pass/fail criterion, but they are not best served by leaving them with that delusion. The access failures are generally part of a pattern of site usability problems and they will get better results from their site and we will get better responses to our input if we cast the dialog with the site in terms of "here are things that detract from effective communication (including the things that cause access failures)" and not simply "here are outright access failures for PWDs at your site." 5. In terms of dialog with websites, there are often two people one must involve to determine the right fix or repair for the site: a content person and a technical person. We should honor the ancient rule of "mirror the customer's organization" and think of the W3C as adding a plaintiff technical person to the complaint- originating consumer who serves as the plaintiff content person. And model the contribution of the consumer's technical friend after the excellent example that Dave has provided. Al ----- Forwarded message from Dave Eaton ----- From: Dave Eaton <dwe@arde.com> To: lynx-dev@sig.net Subject: Re: lynx-dev Bad HTML on hotbot site (Was: bug in lynx) On Sat, 12 Sep 1998, Doug Kaufman wrote: > On Sat, 12 Sep 1998, Stephen Lloyd wrote: > > > ok guys, here's how to reproduce the bug. > > go to http://www.hotbot.com > > you'll be prompted as to wether you want to accept a cooky. press a > > ...[clip].... search for the exact phrase ... lewis collins and submit > > Lynx returns "Bad HTML", which generally means that they really have > incorrect html ...[clip]... > I am forwarding this problem to lynx-dev, where someone can probably > determine what the exact problem is. Maybe we can then get hotbot to > fix their website. The Lynx version returned has several problems, but the one that looks like it is hurting most here is an attempt to nest "<FORM ...>...</FORM>" tags. Using -tagsoup didn't halp any, but hacking (what I best-guessed as) the intended nesting, I was able to get a version of the page to operate correctly and go back to get the next set of data. And indeed different pages are returned to different browsers ... including additional forms to send info to their advertisers, and still some nested forms. Looks like they have gotten confused as to where to end one form and where to begin the next as they build the page and the version they hack together for Lynx is worse than others. Trying to use the page returned for Lynx with a Netscape 4.04 on HP-UX I found that it doesn't even show the "next" that the original problem discussed here mentions, so there would have been no way to try to go to the next answers had that page been returned to Netscape. MSIE 4 shows the "next" submit button, but won't take any action either. FWIW, hotbot is running an MS IIS/4.0 server. A run of the form returned to lynx through weblint shows: hbot(2): outer tags should be <HTML> .. </HTML>. hbot(2): <BASE> can only appear in the HEAD element. hbot(45): <FORM> cannot be nested -- </FORM> not yet seen for <FORM> on line 8. hbot(46): IMG does not have ALT text defined. hbot(49): IMG does not have ALT text defined. hbot(52): unknown attribute "BORDER" for element <input>. hbot(81): value for attribute HREF (http://jandmcont.com/tellup/genealogy/discussion/collins_main.html) of element A should be quoted (i.e. HREF="http://jandmcont.com/tellup/genealogy/discussion/collins_main.html") hbot(82): empty container element <P>. hbot(118): </HTML> must immediately follow </BODY|/FRAMESET|/NOFRAMES> hbot(118): no closing </FONT> seen for <FONT> on line 9. hbot(-): expected tag(s) not seen: BODY Hope this helps you when you try to report the problem to hotbot. Dave Eaton -- Artronic Development - Your Web Site Solution 12629 N. Tatum Blvd., Suite 224; Phoenix, AZ 85032 e-mail: dwe@arde.com - voice: (602) 953-0336 - http://www.arde.com ----- End of forwarded message from Dave Eaton -----
Received on Sunday, 13 September 1998 11:03:21 UTC