---------- Forwarded message ----------
Date: Thu, 9 Mar 2000 07:09:35 -0800
From: Ken Weiss <ken.weiss@UCOP.EDU>
Subject: [UPS] Re: academic press policy

At 9:42 PM +0000 3/8/00, Stevan Harnad wrote:

>On Wed, 8 Mar 2000, Ken Weiss (California Digital Library Technologies) wrote:
>>  I'm not convinced that we've given enough thought to the implications
>>  of a large number of small author-managed repositories with regard to
>>  the issues of access, availability and persistency of scholarly
>>  publications. It's very cheap and easy to set up a server. It's much
>>  more difficult and costly to keep it running over the long haul, with
>>  good security, reliability, performance, and consistency.

>
>The solution (again, given the genie/bottle facts Ken confirms) will of
>course be for an archive-of-archives that harvests and backs up
>individual open-archive contents cumulatively. To get and keep the ball
>rolling towards the optimal and inevitable, though, subversive local
>archiving NOW is preferable to waiting for any other archiving LATER.
>

I applaud Stevan's enthusiasm, but I'm afraid he's proposing to trade
a link management problem for a version control problem. Either of
those problems is preferable to stasis, but in some ways the version
control problem is more insidious than the link problem. If you get a
404 error and no information, you know where you stand. If you get an
out-of-date version of paper from a harvesting archive, you don't
know it unless you chase down the copy from the archive of record.
Dienst has some hooks to help resolve the multiple copy/version
control problem. We need to make sure that the OAI policies require
scrupulous adherence to good management practices, and we need to run
the 'archives of archives' in such a way that they can route around
problematic downstream sites.

--Ken