At 02:28 AM 17-10-00 +0100, Richard Robinson wrote:
>Yes. I think we're both saying "six of one, half a dozen of the other",
>really, and the old "be liberal in what you accept and strict in what you
>output". And maybe that what's generated is more of a key to the situation
>than what's accepted.
I am obviously not getting my point aross correctly. I've always been
slightly suspicious of the "be liberal in what you accept and strict in
what you output" rule, particularly when what you're accepting is, in many
cases, human generated. I agree that what is generated is the key, however
in the case of human generated abc, what is generated will be directtly
influenced by what the programs will accept, so being liberal in what you
accept leads to lots of non-standard abc.
At 10:50 AM 17-10-00 +0100, Laurie Griffiths wrote:
>Right - programs should accept the widest choice but generate standard ABC.
>BUT
>Think about Barfly. It doesn't generate at all.
>If it accepts mangled syntax then it encourages and legitimises such
>mangling
>If it refuses to, then it is "broken" - so it needs careful thought.
>And any other "input only" ABC program (e.g. one that just converts to
>tadpoles, or one that just plays) it has the same characteristic.
>I don't know what the answer is.
This is exactly my point, and why I asked the question about what superset
of abc a new program writer would be advised to support. I notice that
nobody has answered that yet.
Bob
To subscribe/unsubscribe, point your browser to: http://www.tullochgorm.com/lists.html
