cocoon-docs mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Conal Tuohy" <con...@paradise.net.nz>
Subject RE: Cocoon docs link check
Date Thu, 11 Jul 2002 08:58:12 GMT
It's bizarre sometimes how serendipity works. I'm just trying to rewrite
some code I lost for traversing websites, as an example. Perhaps we could
use it as a link-tester? This might make a good concrete example.

> -----Original Message-----
> From: David Crossley [mailto:crossley@indexgeo.com.au]
> Sent: Thursday, 11 July 2002 20:42
> To: cocoon-docs@xml.apache.org
> Subject: RE: Cocoon docs link check
>
>
> Conal Tuohy wrote:
> > Hi David
> >
> > When you say "more automation is needed" and "does anyone
> have a clever
> > solution", what kind of thing did you have in mind? i.e.
> What in your
> > opinion is deficient about the link-checking tools that are
> currently used?
>
> One deficiency that i see is the issue of tracking the
> external breaks. If we knew that a particular link was
> continually broken (say on 3 runs), then we remove it.
> Otherwise let it pass.
>
> LinkAlarm is good, but Cocoon cannot rely on that
> because its use is an external donation to Cocoon.
>
> Perhaps the Cocoon LinkStatus Generator can be expanded
> to also do external links. However, link robots are an art.
>
> For the next while the LinkAlarm reports are available
> and so we could build upon the new Cocoon Sample which
> retrieves the concise listing. LinkAlarm only stores one
> set of reports, so Cocoon would need to store its own
> copies to retain a history of broken external links.
>
> That is no problem, we can manually place them in CVS.
> Then we would need a Cocoon Generator to compare each
> links.broken-YYYYMMDD.txt file and alert about the bad ones.
>
> This would be a great tool. We use LinkAlarm on our
> own website. Always too busy to fix them all. With such
> a tool we could attend to just the bad links.
>
> We also need a better way to track which broken links
> have been fixed, to reduce duplication of effort.
> --David
>
>
>


Mime
View raw message