- From: Irene VATTON <Irene.Vatton@inrialpes.fr>
- Date: Thu, 04 May 2000 10:19:28 +0200
- To: "P. T. Rourke" <ptrourke@mediaone.net>
- cc: "Pierre Fortin" <pfortin@SkyBest.com>, "Hugh Sasse Staff Elec Eng" <hgs@dmu.ac.uk>, www-amaya@w3.org
> I got 86 errors when I ran the Inforworld page through Tidy. > > It has sometimes seemed to me that an ideal browser/editor would have a > "tidying" feature that would re-render loaded invalid pages according to a > best approximation of a valid page (with an alert in the status bar), > flagging errors in an optional extra window or file, and allowing the reader > to view the source of the existing page or of the resulting tidied page. > But I expect that this is something the Amaya team doesn't have the time / > personnel resources for (I guess it would have to run the loaded page > through Tidy or some analog first and then parse it into a tree for Amaya to > display and edit; my programming skills are definitely sub-par so I haven't > looked at the source code), so I haven't mentioned it. I guess that could be done outside the Amaya team by a student. Amaya gets a temporary copy of documents before parsing them. Tidy could launched on these temporary files. No well knowledge of the Amaya code is necessary for doing that. > P. T. Rourke > > Pierre Fortin wrote: > > Hugh Sasse Staff Elec Eng wrote: > > > > > > On Wed, 3 May 2000, Pierre Fortin wrote: > > > > Hi, > > > > directly accessing the first frame at > http://www.infoworld.com/pageone.html > > > > since Amaya only displays basic frame info unless "/pageone.html" is > appended. > > > > Once the page displays, it is obviously very different from what > Netscape or > > > > Mozilla present. > > > The W3C validator considers this page to be seriously ill :-) :- > > > > http://validator.w3.org/check?uri=http%3A%2F%2Fwww.infoworld.com%2Fpageone.h > tml&outline= > > Thanks for the pointer! > > > The question arises how far the very small Amaya team can go to prevent > > > pathalogical responses to pathalogical input. Ideally it should not > > > loop forever, but... > > ...but, this raises another question: should the validator just throw > errors, > > or should there be a "proposed/concept" option, or second validator which > takes > > rule-bending into account and helps evaluate future possibilities...? I'd > be > > the last to propose a M$-like "embrace and extend" attitude; but standards > don't > > always stop when agreed upon. Many evolve to new levels... to wit: HTML > > *4.0*... But; by all means, flag M$ extensions as strict violations... > :^) > > > Hugh > > > hgs@dmu.ac.uk > > Pierre > > -- Irene.
Received on Thursday, 4 May 2000 04:23:04 UTC