Re: Alternative syntaxes for the prolog

I think I already have.

When someone comes up to ask for advice and says "a colleague sent me
this and said it was the answer to my problem -- how do I use it?", it
can be (and almost always will be) useful for the advisor to know (a)
what kind of thing "this" is (the string "ixml" at least gives you
something to look for information about) and (b) what version of the
relevant spec is relevant in this case.

If "this" is a program in a language the advisor recognizes, or a
dataset in a format they recognize, then all is well.  "Oh.  This is a
Cobol program.  You compile it like this [demo], and then run the
compiled version like this [demo]."  Or "This looks like a collection of
comma-separated values.  You should be able to import it into a
spreadsheet program.  But no, I have no idea what character encoding
this is, so you're on you're own trying to figure out which character is
which."

If the programming language or data format is less well known, and not
known to the advisor, the person seeking advice and help is somewhat
less likely to get any useful help.  I suspect that many people will not
know how to execute the following program, or how to figure out what
might be necessary in order to execute it.

   :(s,fi,(
   :(ei,<1>, 1, 0,(
   :(ei,<1>, 2, 1,(
   :(aa, :(ri,fi,:(as, <1>,1)),:(ri,fi,:(as, <1>,2)))
   ))
   ))
   ))`
   :(mw,fi)'

It might be easier if one knew the name of the programming language (and
the version number, if it's one of those languages that has gotten
overhauled every so often).  (Any reader who does recognize the
programming language is invited to ask themselves whether there are any
programming languages in the world which they might not recognize.)

Those who never face the task of figuring out what some unknown file is
will derive very little benefit from any magic number.  Since the world
is a very big place, I suspect that those people do exist.  But I also
think they will suffer very little harm from the presence of internal
metadata, and so I continue to think that at least rudimentary internal
metadata is a good idea.

Michael

Steven Pemberton <steven.pemberton@cwi.nl> writes:

>> Indeed. It’s useful for iXML to have an identifiable “magic number”[1].
>
> Can you give an example of when it would be useful?
>
> Steven


-- 
C. M. Sperberg-McQueen
Black Mesa Technologies LLC
http://blackmesatech.com

Received on Tuesday, 5 March 2024 17:25:22 UTC