[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: SRFI-22 -- scripts.
From: |
Neil Jerram |
Subject: |
Re: SRFI-22 -- scripts. |
Date: |
11 Feb 2002 18:47:57 +0000 |
User-agent: |
Gnus/5.0808 (Gnus v5.8.8) Emacs/20.7 |
>>>>> "thi" == Thien-Thi Nguyen <address@hidden> writes:
thi> From: Neil Jerram <address@hidden>
thi> Date: 08 Feb 2002 11:06:58 +0000
thi> Surely the purists would argue that it doesn't matter which one of
thi> them you get -- isn't that, like, the whole point?
thi> well the purists must neither value separation of
thi> architecture and implementation,
Sorry? Surely it's precisely _because_ they value this separation
that they think a standard access point is a good idea.
thi> nor comparitive analysis, then.
Agreed. But sometimes the user just wants to get a job done, and
isn't worried about comparative aspects such as performance. (Plus,
at least theoretically, interface should specify _all_ observables
anyway, including `comparative' aspects such as performance.)
thi> a real purist would explicitly specify a shell script with
thi> overridable environment dependencies. if you're going to
thi> have a monarchy, let the king be a puppet rather than a
thi> self-serving duke.
I'm not sure I get the analogy, but I agree that such a script would
be better. I think a script (or set of scripts) like this would best
be a separately managed project that saw it as its business to keep
abreast of how to invoke standard behaviour in all its supported
implementations.
thi> i wouldn't mind not thinking about srfi-22 for now -- maybe
thi> there will be a better approach in the future.
Fair enough. (And, anyway, I should probably go and read it before
making any further statements...)
Neil