Re: Specifying all biz rules in relational data

From: mAsterdam <mAsterdam_at_vrijdag.org>
Date: Sun, 19 Sep 2004 21:52:52 +0200
Message-ID: <414de39d$0$566$e4fe514c_at_news.xs4all.nl>

robert wrote:
> mAsterdam wrote:

>>robert wrote:
>>>
>>>...let's start by acting like lowly biologists: 
>>>define the types. 
>>>you need the taxonomy first. 
>>
>>Yes! ...  a taxonomy (hierarchy) of types might just prove
>>to restrictive, though ... but let's give it a try anyway.

>
> ack! hierarchy??? my dictionary doesn't make that restriction:
> "orderly classification of plants and animals according to their
> presumed natural relationships".

googling for "taxonomy" and "classification", I find

http://www.emc.maricopa.edu/faculty/farabee/BIOBK/BioBookDivers_class.html

where I read:

> Linneus attempted to pigeon-hole (or classify) all known species of his
> time (1753). Linnean hierarchical classification was based on the premise
> that the species was the smallest unit, and that each species (or taxon)
> nested within a higher category.

and:

> Some general rules for nomenclature:
>
> 1. All taxa must belong to a higher taxonomic group.
> Often a newly discovered organism is the sole species in a single genus,
> within a single family...etc.
> 2. The first name to be validly and effectively published has priority.
> This rule has caused numerous name changes, especially with fossil
 > organisms: Brontosaurus is invalid, and the correct name for  > the big sauropod dinosaur is Apatosaurus, Eohippus (the tiny

     > "dawn horse") is invalid and should be referred to as

 >    Hyracotherium. Sometime, however, names can be conserved if a
 >    group of systematists agrees.
 >    3. All taxa must have an author. When you see a scientific 

name > such as Homo sapiens L, the L stands for Linneus, who first

 >    described and named that organism. Most scientists must have
 >    their names spelled out, for example Libopollis jarzenii
 >    Farabee et al. (an interesting fossil pollen type I stumbled
 >    across a very long time ago!).

It looks like taxonomy is closely associated with hierarchies. Not really a problem, is it?

> and i'm still not clear about what is being attempted: a fully relational
> user system (which i find both doable and useful), or fully relational
> code generator (about which i am ambivalent).

Compilers and interpreters do a good job at tuplifying human readable code. Now if there is a lot of clearly tabular stuff in the code it may be handy to factor that out into hand maintained tables. DBMS's come in handy to share (between developers) and manage those - but it's still code. In the used system the values only change at software release time.

> if the latter, i don't have the foggiest idea how to guarantee the
> truth of generated code (not to mention regeneration of the code after
> a change of target language).
Received on Sun Sep 19 2004 - 21:52:52 CEST