Kevin Thiele writes:
Date: Mon, 14 Aug 2000 15:28:20 +1000 From: Kevin Thiele kevin.thiele@PI.CSIRO.AU To: TDWG-SDD@usobi.org Subject: Re: Trying it out
Bryan - great! ... by the objective. The advantage of this is being able to bundle individual treatments in different ways (for different projects);
We try to teach programming students that their abstractions should not forclose the use of data and code in presently unforseen applications. Software developers should never be permitted to utter "Why would anyone want to do X?" for any value of X. (FWIW, the most common bugs in languages such as C which do not do type checking arise from buffer overruns because a programmer said something like "Why would anyone ever want more than 256 <something> ? "
the disadvantage is having to make sure that such cross-bundling works - ie you may end up bundling together several taxon treatments with different character lists.
As I frequently rave, this is an argument in favor of object oriented data representations, which are ipso-fact self-describing. XML is a simple way to accomplish this, but so are object oriented databases such as Object Store and its competitors. The book by Abiteboul et al. "Data on the Web" describes---albeit sometimes with some deep theory---how one can combine data whose objects have different descriptions.
Bob Morris