While I'm still quite unfamiliar with Maxima, I believe the tension
between short/cryptic and long/explicit can be minimized, as with
Emacs/Slime's C-h-k, where you type a key chord to get a description,
and C-h-w where you type a description to get a key chord. Does
something equivalent currently exist?
As for Multics' naming convention, this sounds interesting. Reminds
me of Mathematica, where most functions have at least 2 forms,
Fullform and a shorthand notation (e.g., List[...] and {...}), minus
Multics' compositionality.
Olivier Drolet
On Jul 20, 2007, at 09:07, Stavros Macrakis wrote:
>> I agree that "ilt" is too cryptic. I like "laplace_inverse" better
>> than "inverse_lapace" on the general principle of big-endian
>> naming (i.e. most significant part of name goes first).
>>
>> Now this naming question is something I do feel more strongly
>> about, and I would like to hear from others whether or not
>> they support renaming ilt to laplace_inverse (or whatever).
>
> I don't have strong feelings about naming, but I guess it would be
> good to have some consistent principles. There is always a tension
> between long, possibly unwieldy, names which are however clear and
> explicit, and short, possibly cryptic, names which are convenient.
> Maxima tends to the cryptic, which is perhaps reasonable for the most
> common operations (diff, rat) but probably not for the less common
> ones (cf, ilt, genfact).
>
> Still, once you start becoming a heavy user of
> rational_to_continued_fraction or inverse_laplace_transform or
> generalized_factorial or even factor_polynomial, that quickly becomes
> annoying. Maxima does have the "alias" function which allows defining
> alternate names.
>
> I am also not sure that very systematic, big-endian names, going from
> most general to most specific (e.g. transform_laplace_inverse or
> conversion_rational_to_continued_fraction) are going to be intuitive
> or convenient for many people (military supply clerks?: Uniform, Navy,
> Officer, Dress, Summer), though the principle is appealing.
>
> Multics had an interesting naming convention, which we may want to
> consider. Commands had a full name (e.g. change_working_directory,
> list_directory), and then standard short forms for the *components*,
> e.g. change => ch, working_directory => wdir, so that there would be
> predictable short forms, something like ch_wdir, ls_dir, etc. (The
> Unix commands ls, cd, etc. are a distant echo of this, but as often
> with Unix, throwing out half of the good ideas so it would run on
> small machines.) So we might have laplace_inv, rat_to_cf, etc. And
> then there was the ultrashort form, cwd, ld (?), etc.
>
> Just some thoughts....
>
> -s
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima