- From: olivier Thereaux <ot@w3.org>
- Date: Tue, 26 Sep 2006 14:25:55 +0900
- To: Nicolas Krebs <nicolas1.krebs3@netcourrier.com>
- Cc: public-unicorn@w3.org
Bonjour Nicolas, On Sep 21, 2006, at 19:08 , Nicolas Krebs wrote: > An idea for a web validator, could be a future version of unicorn, > or a > successor : > During the check of an html document, the validator does not check > only > this one (markup, wai, embeded css etc), but also the linked document. > I did give some example : rss, atom, css, rdf (all linked by > <link>), svg > (linked by <object>). I test if the specific valildator check the > file (and the > result is "checked" for the 6), and next i test if unicorn check > the file > if i submit only the (uri of the) embedding file (and the answer is > "not checked" > except for css). I see now, thanks a lot for clarifying your thoughts. I agree it is an area of work which could be interesting in the future. > Next step in the road : even, coupled to a web site copier/mirrorer > (like Httrack > http://www.httrack.com/), capacity to check a whole web site by > just clicking > one command (such i can miror a whole web site with Httrack). My experience with e.g the W3C link checker makes me lean toward disagreeing with you here. Not that I disagree with crawling and batch-checking a site, I think that's an excellent idea. But I disagree with the idea that the Unicorn tool should take that on: the Unicorn, at this point at least, is a web-based tool, and hence produces results in real-time. Crawling and checking a whole site is, I believe, something more adapted to an asynchronous, scheduled job, and hence more adapted to a tool such as the log validator. http://www.w3.org/QA/Tools/LogValidator/ -- olivier
Received on Tuesday, 26 September 2006 05:26:10 UTC