On Aug 12, 2009, at 9:04 AM, Mark Martin wrote:
Elaine Ashton wrote:
What information do you suspect we have? :) Getting to the urls
programmatically from inside the db is not as simple as it might
seem. Also, I'll have to revisit the redirect algos for apache and
tomcat to see how and when it scans the remapping table, but when
you're talking about 10-20,000 pages to be individually remapped,
it will not be without an expense to the performance of the site as
a whole. And if you maintain the broken links, you commit to
maintain them in perpetuity whereas if they are broken, people fix
them and move on. I believe most links will get you to the top-
level project page which is a reasonable compromise.
1) You don't have to maintain in perpetuity. You can set reasonable
expectations.
What makes you believe that breaking them either today or in 2 years
is going to be met with a discernibly different response? I fully
expect that it'd be the same flurry of unreasonable expectations as
there are now.
2) People don't "fix them and move on". Look at the fiasco that
happened with the ARC move. No notice. Not enough decency or
respect to even respond to questions or suggestions. And in the
end, a large body of work (case artifacts, emails, etc) which is
either difficult or impossible to correct links within. At least in
this case, we've been giving something resembling a notice, which I
suppose is an improvement.
I think the ARC stuff is a special case and is mostly a problem from
within the ARC group itself.
Links change constantly and we can't control who points to what in
various media and this has been an endemic problem in hypertext
since there were more than two pages on the internet which linked
to each other.
And thus our forefathers had the foresight to grant us the 30x
response.
I suspect this is just another case of the website "community" group
marching to the beat of a different drummer (i.e. serving other
interests), and essentially ignoring the needs of the users.
You know, it does get pretty discouraging, even though I am in the
business of being unappreciated in the boiler room, where people who
are technical enough to appreciate and understand the problems with
the scale of what migrating all this content entails do little but
gripe and sling insults.
Our forefathers did not give us sentient webservers with self-healing
URIs.
I AM thinking of the end users as if we would do a URI rewrite for
10-20,000 pages, this would entail the server scanning /every/ rule to
match against before finally serving up the page. In the battle of
glacial server vs. broken link on a zippy server, the zippy server wins.
That being said, if there is someone out there who does happen to be
familiar with large-scale production-level real-world kind of site
migration issues, I'd be interested in hearing about your first-hand
experiences.
e.
_______________________________________________
website-discuss mailing list
[email protected]
