q-lang-users

[q-lang-users] Symbol type

[Burton S. <kr...@gm...>]

Hi Again,

So, after getting over my little problem with the prelude option, I
managed to expand my cat routine in the following way, to handle
multiple arguments:

cat F:String = cat_file (fopen F "r");
cat F:File = cat_file F;
cat F = cat_file (fopen (str F) "r") if issym F;
      = prints "cat: unsupported type";

cat_file F:File = () if feof F;
           = putc (fgetc F) || cat F otherwise;


Since I'm just learning the language, I figured I'd make a cat
function that handled multiple types, like actual File handles,
string/quoted names and symbol names (so it looked more like the Unix
shell).

The only problem that I found with that was that there was no Symbol
type (that I could find), so my code for handling that case had to
differ from the String and File handlers by using an 'if issym' call,
which works fine, but breaks consistency with the other options.

Not that that is a big problem, but one of the reasons I like higher
level languages is their ability to work with actual language symbols
as a type (something about the idea of symbols works very well with
how my brain is wired).  Is the Symbol type left out for a reason, or
is it under a different name that I haven't found yet?  Would there be
any objection to adding it?

-- 
burton samograd                                                kr...@gm...
kruhft.boldlygoingnowhere.org              www.myspace.com/kruhft

Re: [q-lang-users] Symbol type

[Albert G. <Dr....@t-...>]

Burton Samograd wrote:
> Is the Symbol type left out for a reason, or
> is it under a different name that I haven't found yet?  Would there be
> any objection to adding it?

Yes, in fact it would be fairly useful to have a Symbol type (and 
subtypes FunSymbol and VarSymbol) for doing meta-programming, but that 
interferes with Q's current type system which only has single 
inheritance. Hence a constant symbol couldn't be a member of its own 
(user-defined) type, which might have its own supertype, and the Symbol 
type at the same time. So, until Q gets multiple inheritance, you'll 
just have to put up with the issym function and friends.

We've actually discussed the issue of multiple inheritance here a while 
ago and, looking at the code, I found that extending the construction 
algorithm for the pattern matching automaton to handle multiple 
supertypes is not straightforward. So I'm postponing this until I have a 
bytecode compiler written in Q itself (that's one of the reasons I'm 
working on Q Yacc and Lex), this will make it much easier to implement 
such changes.

Cheers,
Albert

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikinformatik.uni-mainz.de/ag

Re: [q-lang-users] Symbol type

[Burton S. <kr...@gm...>]

On 8/23/06, Albert Graef <Dr....@t-...> wrote:
> Burton Samograd wrote:
> > Is the Symbol type left out for a reason, or
> > is it under a different name that I haven't found yet?  Would there be
> > any objection to adding it?
>
> Yes, in fact it would be fairly useful to have a Symbol type (and
> subtypes FunSymbol and VarSymbol) for doing meta-programming, but that
> interferes with Q's current type system which only has single
> inheritance. Hence a constant symbol couldn't be a member of its own
> (user-defined) type, which might have its own supertype, and the Symbol
> type at the same time. So, until Q gets multiple inheritance, you'll
> just have to put up with the issym function and friends.

Makes sense.  Thanks for the response. I'll make due without until then.

-- 
burton samograd                                                kr...@gm...
kruhft.boldlygoingnowhere.org              www.myspace.com/kruhft