Bibliography project

Rodger Whitlock totototo@mail.pacificcoast.net
Thu, 30 Oct 2003 13:41:56 PST
On 27 Oct 03 at 8:25, Boyce Tankersley wrote:

> What software application the database uses is less important than
> it being a relational database. 

The proper form of a relational database for a bibliography is a
very difficult problem to solve. The data does not lend itself
readily or obviously to the relational model. Rather than try to
devise a home-brew solution, those interested are well advised to
conduct a careful and thorough literature search and determine what
solutions have been devised by more experienced people.

If the documents were all available in digital form, then a googloid 
search engine would be considerably more useful than a relational 
database, I think. But as the premise is not true, this conclusion is 
not useful.

It also seems to me (if I may speak with full candor) that this plan 
for a bibliography has gone off somewhat half-cocked. Even more 
important than the database design is a specification of the intended 
uses of the system: precisely what kinds of questions and queries is 
it intended to answer and what kinds is it NOT intended to answer.

Also more important than the database design per se is some kind of
realistic planning around the issues of

(1) how complete is this bibliography to be? 

(2) who is going to provide the data? and

(3) who is going to exercise data quality control?

(4) in the long term, how are you going to ensure continuity of 
support and function? The bulb literature is not static and any such 
compilation will require ongoing work to keep it current.

A database (sensu latu) where everyman and his sister can input data
is only as good as the input of the most careless keyboarder. It
remains a fact of human nature that lots of people are notably
careless in such matters.

There's no point in everyone running off half-cocked until these 
issues are sorted out. I'm sure we've all run across web sites that 
were started with best intentions but were left half-constructed when 
the originator ran out of steam. No point adding another to this 
melancholy list.

It is also worthwhile remembering that such indexes as Botanical 
Abstracts exist, and no doubt some institutions (the RHS and Kew, in 
particular) may have in-house indexes. What is the relationship of 
the proposed bibliography to such existing databanks?

-- 
Rodger Whitlock
Victoria, British Columbia, Canada
"To co-work is human,
to cow-ork, bovine."


More information about the pbs mailing list