[Bioperl-l] pipeline

[Imre Vastrik]

Ewan Birney wrote:

[snip the rest since I agree]

>   (c) It seems to encourage large numbers of objects which actually
> prevent understanding the system well.

This is really up to the programmer/designer who decides to create yet
another object. I'm not quite sure if the manual O<->R mapping
associated "overhead" has discouraged any "object afficinados".;) My
point was actually "opposite": if one has large (whatever that means)
number of objects it would be dumb to handle the O<->R mapping
"manually".

>   (d) it becomes hard or virtually impossible to view the data as pure
> "data" as we do in things like Lite. I used to be against a pure data view
> of the world, but experience over the last couple of years, and watching
> what people can do with the data view (eg, Arek) has changed my mind.

I'm sure that provided that one knows what is the data that needs to be
viewed as "pure data" it would be possible to create the denormalized
"schema" (as well as code which does the denormalization) on the basis
of the "object model".

Rgds.,

i

P.S. I'm in no way suggesting that Ensembl as a whole should move
towards automatic O<->R mapping and suchlike. I'm merely trying to
tickle people's brains with less "orthodox" views.