- From: Martin Salo <salo@laborint.com>
- Date: Tue, 20 Apr 2004 21:42:59 +0300
- To: <www-validator@w3.org>
- Cc: "'Liam Quinn'" <liam@htmlhelp.com>
Hi You could save bandwidth (and raise limits) by only displaying links to sites that are invalid. When clicking on the links it shows you the details. Displaying all errors is pretty intensive bandwidth wise. It is pretty messy too. Another solution is to make a report in html and zip it up and then send the result. When going large scale (up to 1000 pages) a ticket system or something similiar has to be thought up so you could see the results when finished because they would take up a lot of time (affected by the server speed the site is running on, the validating server speed etc). Another problem would be the bandwith when getting the pages from servers to validate them. Some servers use gzip to zip up the content they are serving, i don't know if the validator uses that technique to save bandwith, but it should be a pretty good idea to use it. I could help in development of the new validator, i don't know about the hosting though. Martin Salo -----Original Message----- From: Liam Quinn [mailto:liam@htmlhelp.com] Sent: Tuesday, April 20, 2004 9:27 PM To: David Dorward Cc: Martin Salo; www-validator@w3.org Subject: Re: (Entire) Site validation On Tue, 20 Apr 2004, David Dorward wrote: > On 20 Apr 2004, at 15:32, Martin Salo wrote: > > What do you think about an idea to develop a validator that validates > > whole sites, not only pages. It could search for links in pages and > > when a link is in the same domain (a subsection of the page) it > > validates it. It should be limited (to about 200 pages or so) to avoid > > abuse. > > Such a validator already exists, > <http://www.htmlhelp.com/tools/validator/>, although its limit is 50 > pages (a fact which, last week, led me to finally getting around to > setting up the w3c validator locally). I just raised the limit to 100 pages. (We recently moved to a new host with cheaper bandwidth, so I was planning on raising the limit soon.) > I have some concerns about the load this would cause on the w3c server; > it has a rather higher profile then the WDG. What is the load > (bandwidth / CPU) like at present (assuming there are no issues with > making that information public)? The load depends a lot on the server that handles it. On our old, underpowered server, we had trouble with badly-behaved robots overloading the server with too many CGI requests (sometimes involving the Validator, but often other parts of the site as well). I've changed the limit on number of pages a few times in the past, but the reason for lowering the limit was always to compensate for the extra bandwidth needed to handle Microsoft mass-mailing worms at our mail server. -- Liam Quinn
Received on Tuesday, 20 April 2004 14:43:16 UTC