data sharing

[Hugh Wilson]

On 10 Dec 98 at 12:33, Dave Vieglais <vieglais at UKANS.EDU> wrote:

> Hmm.  You have just described a DBMS....

Not really.  I am describing a process that allows the DBMS - any
flavor of DBMS running on any sort of machine - to be managed and
*controlled* at the data *source*.

> simply put the data in a text file?  Is it not easier to keep the data in a
> database (access, dbase, whatever) and have your query posted directly
> against the database?  Are you suggesting that it is easier and less
> expensive to go through many steps of creating a text file and indexing it
> each time you update the database?

Yes, that is what I am suggesting.  The ASPT Treasurer, like many
systematics collections managers, is not about to establish his PC as
a network node.  In addition, if he was willing and able to do this,
he would not want global internet access to the base version of the
ASPT membership directory, i.e., the local source will want local
control and local determination of content for both local and web
access.

Agreed, translation from the ASPT treasurer's DBMS to an indexed file
for web access may not be the most efficient or logical path for web
access from the 'information management' point of view but, given the
practical, 'on the ground' reality of information flow from ASPT
members to the web, it is the path of least resistance, at this point
in time.

The same dynamic applies to the notion of using full text indices as
query sources for data being *shared* by multiple sources, i.e.,
museum or herbarium collections.  This path allows DBMS activity to
be distributed among data sources and *controlled* at the points of
data production.  Data providers are able to work in an environment
that *they* define and, via consensus among those contributing data,
establish whatever (minimal) data exchange format is needed.  The
next step could involve setting up each local DBMS as a web node
with each node responding to a query to the 'merged', 'shared' or
'distributed' data.  This sounds good but it will not work - at
least it has not worked and its not likely to work given current 'on
the ground' realities.  Or, maybe all contributors could send their
data to a big, server-based DBMS to web express the shared data?
This might work but it adds a DBMS processing step, inserts DBMS
computation between the query and the response, removes contributors
from a significant processing step, and inserts a large, complex,
expensive element to the enterprise.  Thus, the notion of using full
text indexing, using public domain software, becomes - to me at least
- an attractive option.  It might not be best solution.  However, I
think there is only one system on the web that provides a view of
multi-herbarium collection data and this site at:

http://www.csdl.tamu.edu/FLORA/ftc/ftphsb.htm

employs text indices for both text returns and map displays.  It is a
flawed, prototype site BUT it works, it is working *now*, and it
provides a foundation that could be expanded and enhanced..




Hugh D. Wilson
Texas A&M University - Biology
h-wilson at tamu.edu (409-845-3354)
http://www.csdl.tamu.edu/FLORA/Wilson/homepage.html