Re: [MACCAWS] Bobby gets mean and stingy
MJ Ray <[email protected]> Wed, 09 Oct 2002 19:40:15 GMT
| Newsgroups | gmane.comp.web.vkit |
|---|---|
| Organization | A poorly-maintained Debian GNU/Linux InterNetNews site |
| Message-ID | <[email protected]> |
[email protected] <[email protected]> wrote: > What I'm looking for is a package I can download onto a local IIS dev box > and tell it to crawl our site (using wildcards to restrict to only our > subdirectories, we have numerous subdirectories under a main site that is > out of our control) and run an analysis on each page it finds. This allows > it to get the dynamic content, and in the report I want it to list which > page had the problem and what the problems were, so I can then e-mail the > reports to the affected site administrators here. Well, you'd probably have to write your own file tree crawler (or I could write one for a modest fee ;-) ), but it should definitely be possible. Actually, I already have something similar written for another project, that traverses trees of XML files to index them. > Of course, I want that with an HTML validator and CSS validator as well... >:\ Well, aren't the xhtml and CSS validators already out there? The first version of this will definitely include a primitive sort of XML validation -- if it's not valid XML, it probably won't be parsed. Later versions can add support for parsing mangled HTML, but trying to do that in version 1 will be biting off more than we can chew. It's something I know how to do in *theory* but have never used in practice. Anyway, WCAG checking looks quite hard enough without adding more complications to it. > Open source is great. I'd think there would be a market for projects like > this, in multiple languages (Perl, Java, COM, etc) so we could pick and > choose implementations. I think it's more practical to make connectors to allow it to be called as a black box from other languages, than to try implementing it in less powerful languages. Recompilation to target their runtime worlds (JVM, CLR, Parrot one day, hell maybe even the VM someone wrote in PHP...) is always an option. Anyway, if there's a market, why are people so backwards in coming forwards with pledges of support? Can Bobby really command $99 from enough people to make it worth their while, but this can't find 10 people or companies interested enough to fund it? Or are people holding back, waiting for something more? > What are the difficulties involved? Is there a way to encapsulate the logic > so it could be called from other programs, e.g. make it command-line, so it > could be done the way I described? If we can come up with sensible parameters and output format, then that is definitely possible, maybe for version 1. [...] > Would this be a SourceForge project so it would be available at all times? I'll set it up at Savannah.nongnu.org in a few days. It will be available forever, as far as I can ensure that. I do not agree with SourceForge's use of proprietary software, so it will not be hosted there. -- MJR| v ---|--[ Luminas internet applications http://www.luminas.co.uk/ ]-----| `--[ http://mjr.towers.org.uk/ ]---------[ slef at jabber.at ]-----'