Re: [pgsql-advocacy] Gborg: announcement by 404 - Mailing list pgsql-www

From Magnus Hagander
Subject Re: [pgsql-advocacy] Gborg: announcement by 404
Date
Msg-id 20071114085025.GA10465@svr2.hagander.net
Whole thread Raw
In response to Re: [pgsql-advocacy] Gborg: announcement by 404  (Michael Paesold <mpaesold@gmx.at>)
List pgsql-www
On Wed, Nov 14, 2007 at 09:15:29AM +0100, Michael Paesold wrote:
> Gregory Stark wrote:
> >"Greg Sabino Mullane" <greg@turnstep.com> writes:
> ...
> >>http://gborg.postgresql.org/project/slony1
> >>
> >>...which now gives a very unfriendly 404 message from pgfoundry. Granted,
> >>the first Google hit is the authoritative one, but this is probably
> >>just the tip of the dead link iceberg. Google on 'pljava', 'dbdpg',
> >>or 'pgsphere' for some even scarier examples.
> >
> >I think the problem here is that if there's an error Google will notice
> >that
> >and stop returning search results. If there's a page saying "I'm not a
> >page"
> >then Google will assume that's what's supposed to be there.
>
> That's what redirection using "301 Moved Permanently" is for. It will tell
> Google that the new page permanently replaced the old one. Concerning
> Google, this is a better solution than either a 404 or a human-readable
> page about the move.

Or you can construct a human-readable 404 page. Google will notice it says
404 and not index it, but a browser will show a nicer error msg than just
"not found".

IIRC, custom error pages in apache does that. If not, it can certainly be
done with simple PHP - we do this for wwwmaster.postgresql.org to deliver
"not found" pages inside the graphical framework. (originally they didn't
have a 404 code though, which caused the mirrorer to pick them up, and
google to index them, which wasn't a very good idea :-P)


//Magnus

pgsql-www by date:

Previous
From: Gregory Stark
Date:
Subject: Re: Gborg: announcement by 404
Next
From: Magnus Hagander
Date:
Subject: Re: pgfoundry is down