Jim Henderson nous apporta ses lumieres en ce 12/05/2006 14:15:
> On Fri, 12 May 2006 20:02:27 +0200, Sven Littkowski wrote:
> 
> 
>>But that also applies if humans check each link. And since humans do it
>>more seldom and not so frequently, the time for correcting a wrong
>>decision is much longer.
> 
> 
> I don't know how Chris and Chris are doing their validation (I'll let them
> speak to it), but the way I'm validating the links I'm working on is to
> try the link - if it doesn't work, I use Google and web.archive.org to
> locate information about the site.  With web.archive.org, I look to see if
> the last update includes a pointer to a new location.
> 
> I only mark a link for deletion if I can find no trace of the program any
> more.  If I find a copy of the referenced program or site that's valid, I
> update the link with the new link information.
> 
> That sort of process is not something that would be easily (I don't think,
> I could be wrong) coded into a link checking program.
> 
> Jim
If a link don't work, flag it. If it don't work for more than a week, kill it
definately. Have the 
script make a log of such dead link. You can also create a broken links page. Monthly,
look at the 
log, or broken links page, and do searches for those pages to see if they have moved
to a new 
address or are realy gone.
At least, you save the work of finding the broken links.

-- 
Alain
-------------------------------------------------

Post a reply to this message