Announcing: CSIR web service

[David E. Bernholdt]

Dominic Ryan raised some very valid questions about the relationship
between CSIR and other services.  My general philosophy about this is
that CSIR is intended to help people discover software and information
that may help them with their research, teaching, etc.  We don't
produce the software or the mailing list traffic (except for cases
like this :-) -- we rely on its existance and just try to help people
find and use it.

To deal with his points more specifically...

Dominic Ryan wrote:
> Doesn't this sound a lot like QCPE, but bigger?  This, to me, begs the
> question, is QCPE now redundant, even in the latest inception on the CCL web
> pages?

QCPE has a venerable history in our community, and I definitely do not
think it is redundant.  The problem is that QCPE doesn't meet
everyone's needs -- for whatever reasons, some prefer to offer their
software from their own web or FTP site or some other archive such as
CCL, while others prefer to sell their software.

As a result, we end up with scores of sites (including the web sites
of commercial vendors) where you might look for software to meet your
needs of the moment.  How many of these sites do you know of?  Are you
missing useful software because you don't know the right site to look
at, and because no one mentioned it in answer to your query on CCL (or
wherever)?

This is where CSIR comes in.  We want to provide a _virtual_ software
repository -- we don't have to hold the software ourselves (though
that is an option), we can point you to its source.  That source may
be QCPE or a commercial vendor or someone's web site.

I would expect that if we had QCPE's holdings in our catalog
(something I plan to discuss with QCPE and CCL people), we could only
increase their business by helping more people find out about the
software they offer.  The same for any other "repository".

We want to offer breadth of information and leave the depth (the
software itself) mainly to others.  In this way we limit CSIR's scope
and therefore the cost to offer this service. (Yes, we can act as a
physical archive and will do so as far as resources allow.  But if it
gets too popular, and we begin to run out of resources we will probably
seek collaborators to develop sub-discipline repositories and off-load
some of the archive.  In fact, CSIR uses a standard data model for its
catalog, so even the catalog itself can be distributed and still
operate seamlessly, if it becamse necessary or desirable.)

> [...description of the AskNPAC Mailing List Archive...]
> More direct overlap with the CCL.  I would love to see these two
> directions sorted out, particularly since $$ are getting put into
> both.

Once again, we're relying on existing resources, but trying to bring
them together and make them easier to find and use.  Yes, we archive
CCL.  We also archive this list and dozens of others.  You can find
most of this stuff on the Internet somewhere (though not all list
owners archive their lists), but where is it?  Can you search the
archive easily, or do you have to retrieve a series of month-long
files by email from a Listserv system and then search them on your own
machine?  Unles they happen to be hosted by the same site, how easily
can you search _all_ the mailing lists which might have information
that interests you? I like occasionally to look for interesting
threads in a number of newsgroups, but I can't afford the time to
process all of the messages I would receive if I actually subscribed
-- especially if my interest is a narrow one.  CSIR is merely offering
a single, uniform point of access to the many lists that are already
out there.

As far as CCL in particular goes, Jan Labanowski (the proprietor of
CCL) and I are very much aware of each others efforts, and have
disccussed them in some detail.  Quite happily, Jan has received and
NSF grant recently to help develop CCL.  Not to diminish his work in
any way, but it centers around a single list and a single (though
broad) topic.  CSIR cannot hope to work at that level of detail --
we're trying to bring a wide breadth of information to the community.
In due course, I expect we will offer _links_ to many of the things that
Jan is developing to help computational chemists who utilize CSIR to
locate additional information that may be valuable to them.  The
efforts are complimentary rather than competetive.

It looks like this ended up as a pretty lengthy reply.  You can
probably understand why I didn't want to do into this detail in the
initial announcement.  Our Trends in Analytical Chemistry Internet
Column also discusses these kinds of questions, in part, and its even
longer :-)
http://www.elsevier.nl:80/inca/homepage/saa/trac/resource.htm

I'm happy to discuss this further or answer other questions.
--
David E. Bernholdt                      | Email:  bernhold at npac.syr.edu
Northeast Parallel Architectures Center | Phone:  +1 315 443 3857
111 College Place, Syracuse University  | Fax:    +1 315 443 1973
Syracuse, NY 13244-4100                 | URL:    http://www.npac.syr.edu
-----
chemweb: A list for Chemical Applications of the Internet.
Archived as: http://www.ch.ic.ac.uk/hypermail/chemweb/
To unsubscribe, send to listserver at ic.ac.uk the following message;
unsubscribe chemweb
List coordinator, Henry Rzepa (rzepa at ic.ac.uk)