Goodbye TCL

[Cameron Laird]

In article <c60i7t$2aff$1 at godfrey.mcc.ac.uk>,
Donal K. Fellows <donal.k.fellows at man.ac.uk> wrote:
>Cameron Laird wrote:
>> 'Different way to say this:  "database", in this sense, is a dirty
>> subject, in that it's far more tied to grungy details of specific
>> implementations than the masses generally realize, or than either
>> Python or Tcl, both of which prize purity in their own ways,
>> generally support. A general database abstraction is about as
>> rarefied as a general GUI abstraction.
>
>I had a quick look into general DB APIs in the past, and the response I
>got from the DB community at the time was "but this doesn't support this
>*vital* feature only implemented in one DB that 0.01% of users really
>need and therefore we could never ever use your proposal for a standard
>API".  It's kind of discouraging.
>
>FWIW, I think some mostly-common subset of functionality is possible (if
>we assume SQL underneath; without that, you can't get anywhere) and it
>would be enough to get most of the low-end and casual users total
>satisfaction. And maybe this would encourage alteration in the high-end
>interfaces to be more like each other too.
>
>Donal.
>
Yeah; I've written database abstractions, going back to before
there was a Tcl.  *I* like them fine.  To me, though, they don't
feel like something that belongs in the core.

-- 

Cameron Laird <claird at phaseit.net>
Business:  http://www.Phaseit.net