Reply on XDF

[Mauro Cavalcanti]

Paulo Buckup wrote:

>I don't think that the ILDIS XDF format has gone unnoticed.  In fact
>it has even been featured in NSF-sponsored international workshops (e.g. the
>Workshop on Use of Artificial Inteligence in Systematics).  The real reasons
>for it not beeing used are: (1) it is quite complex, and (2) except
>for the ILDIS own software, I am unaware of any other software capable of
>producing output, or even reading their XDF format without a great programing
>effort.

I cannot entirely agree to this, for two the following reasons: (1)
I do not think that "complexity" is an appropriate criterion to judge
an exchange data format for biological data - that are by their very
nature complex ones, especially in the case of descriptive data! For
example, several users consider Mike Dallwitz's DELTA format "complex",
but no other format for storing taxonomic descriptive data with such
richness of details and powerful flexibility has gained so widespread
acceptance as that... (2) Richard White and Robert Allkin are just working
on a general-purpose XDF processor (the XPGM program); they also intend
to make available a library of routines for reading and writing XDF
databases, which could be used by others, in developing their own
software; they should be contacted for details.

>In practice flexibility goes against standardization.  For example,
>a controled vocabulary that incorporates all words currently in use would
>certainly be flexible to convey descriptive data, but would be of
>very little use as a standard vocabulary.  The most sucessful
>standards are those that are simple and restrictive of formats.  The
>Gopher protocols ilustrate the point.  Only seven data types fullfill
>the needs of millions of users.  More flexible, but complex database
>servers are unlikely to be so sucessfull:

Again, I am only in partial agreement to this. Standards "simple and
restrictive of formats" would perhaps be good for specimen-based data
(as, for example, the ITF proposed for the exchange of accession data
from plant specimens housed at arboreta and botanical gardens), but
I think they would not be adequate for taxon-based data (as ILDIS itself
could demonstrate!).

And finally, it is worth mention that the flexibility of a format as
XDF would permit to any particular database project define its own,
standardized, data types (again, as ILDIS have done) - being them
seven or a hundred. Hence the importance of the flexibility offered
by this kind of format.






Prof. Mauro J. Cavalcanti            Phone     : 55 (021) 551-5542 ext 148
Departamento de Biologia Geral       Home phone: 55 (021) 594-4745
Centro de Ciencias Biologicas        E-mail    : maurobio at ax.apc.org
Universidade Santa Ursula                        maurobio at ibase.br
Rua Fernando Ferrari, 75 - Botafogo
22231-030, Rio de Janeiro, RJ, BRASIL