UNIDATA

[Jim Beach]

> Hi folks,
>
> Have any of you any experience with UNIDATA? Our Palaeontology Department,
> have recently switched from an in-house database to the commercially
> available UNIDATA. This is currently being reviewed for use throughout the
> Natural History Museum here in London. I'd appreciate any comments about
> UNIDATA's performance and networking abilities please.
>
> Brian
> Brian Pitkin
> Department of Entomology
> The Natural History Museum
> Cromwell Road
> LONDON SW7 5BD
>
> INTERNET  b.pitkin%nhm.ic.ac.uk at nsfnet-relay.ac.uk


Brian,

The basic problem with Pick-based databases like UNIDATA is that Pick
derives its strength from non-relational data structures which make Pick a
jewel of a system for database developers.  It really comes into its glory
when data are significantly de-normalized in the data files in order to make
application development relatively easy and quick.

It is a favorite development environment of consultants because prototype
and production systems can be engineered realtively quickly.  But personally
I consider Pick and its derivitives (such as Advanced Revelation) to be the
cocaine of scientific computing.

What you give up is medium and long-term flexibilty for client/server
computing architectures and systems.  Only Pick code runs against a Pick
database.  You won't have any flexibility for third-party report generator
front-ends or QBE front ends, Windows front ends, etc.  You'll be married
(and monogamous) to UNIDATA for a long, long time.

For vertical markets like Doctors' and Dentists' offices it can be just the
ticket - why should they care about data modelling, entity cardinality
relationships, and taking advantange of relational operators and servers?

Scientific computing is another matter, it involves community information
models, data quality and integrity issues, severe human resource
constraints (in the case of most museum departments), and the need to
adopt a database computing strategy which will allow for medium and long
term flexibility in order to take advantage of new software products (in
this case database front-ends and servers).

Unidata, Pick, and Advanced Revelation are very specialized database
environments which are usually implemented in very non-relational, very
de-normalized designs.

I am going to copy your query and this response to the Taxacom mailing list
in the hope of generating some additional responses and perspectives.


Jim


--

James H. Beach                                      beach at huh.harvard.edu
Museum of Comparative Zoology,                      Tel: (617) 495-1912
Herbaria, Arnold Arboretum                          Fax: (617) 495-9484
22 Divinity Avenue
Harvard University
Cambridge, MA 02138, USA







--

James H. Beach                                      beach at huh.harvard.edu
Museum of Comparative Zoology,                      Tel: (617) 495-1912
Herbaria, Arnold Arboretum                          Fax: (617) 495-9484
22 Divinity Avenue
Harvard University
Cambridge, MA 02138, USA