Re: 3DLDF

[Hans Aberg]

At 21:31 +0200 2004/08/14, Laurence Finston wrote:
>> You do as you want, as it is your program. But you sounds as Akim Demaille
>> before he implemented M4 into Bison.
>
>If it ever seems useful to me, I'll  do it.  I appreciate the suggestion in
>any case.

Only you, of course, can decide what is useful in your programming. :-) I
merely want to provide some inputs.

>> I think that these discussions started with your problems of using Bison to
>> implement a MetaFont-like language. But if you have no problems there, just
>> go ahead.

>No, it started with a discussion about `%union' and the relative merits of
>using `void*' and a base class.  It then moved on to a discussion about
>`static_cast()' versus `dynamic_cast()'.
>
>I haven't had many problems with Bison, it works a treat.  I don't consider it
>a major problem that it's apparently not possible to implement all of the
>features of Metafont's parser using Bison.  I wouldn't expect it to be.

But here you say that not all MetaFont stuff can be implemented using
Bison. My own view is that I do not want to stick my head into the
questions on my own, but only discuss the problems here, if brought up.

>> The point (no pun intended!) with the intermediate language is that
>> you no
>> longer need to have that coupling:
>
>I see this coupling as an advantage rather than a disadvantage.

Well, you must judge what is useful in your programming. I think though
that an intermediate language might be useful if that tie becomes
bothersome, or if you wants to produce a machine language later on.

>Different people have different approaches to programming.  That this is
>possible is one of the things that makes it an art, in my opinion.  I
>originally thought that the title of Knuth's  _The Art of Computer
>Programming_ was a joke, and that it wouldn't be a serious book.  Then I
>looked inside.
>
>It was largely Knuth that made me realize that programming could be a creative
>activity and that it could be pursued on a high level --- as high as one is
>able to go.   But one doesn't need to be a great computer scientist or
>mathematician to program something useful and interesting.  It's also possible
>to make a lot out of a little.

I was initially very impressed with TeX, and tried do a couple of things in
it. But then I found that it has sneaky trap: it invites you to do advanced
things in elitistic manners, but programming features to carry these things
to the ends are not really there in the language. So I was by that put off
by the Knuthian ways.

I then started programming in C++, and I found that to be an exciting
language to get things done. But eventually, I exhausted its capabilities
in a project. For that reason I started to write on a language using Bison
to automate the things I want in my C++ code.

So therefore, my hunch is that focusing too much on a single path, like you
do with MetaFont-like syntax, might become a bit too restrictive in the
long run. But there is nothing wrong dipping down into such a path fully
until one realizes it has become exhausted. At that point, one can truly
set forth doing ones own, creative work, building on those experiences.
(This should not be construed as that I suggest or know what turns your
future work might take :-).)

  Hans Aberg