Re: [Axiom-developer] RE: Another question

[Ralf Hemmecke]

On 08/22/2006 04:20 PM, Martin Rubey wrote:
> Ralf Hemmecke <address@hidden> writes:
>
> > > > [...]  in that chunk, then that corresponds to \atthistype. (And, yes Martin,
> > > > that can be replaced on the fly by the appropriate \adtype{DomainName},
> > > GREAT! could you send me an appropriate patch, please. Then I can document my
> > > guessing package with ALLPROSE.
> > > > but that is only half of the story.)
> > > Why?
> > That is the reason why it is completely unnecessary to replace \adthistype by
> > \adtype{DomainName}. If you look at the generated .as.nw.tex file, you see that
> > after \begin{+++} there suddenly appears an \addefinetype{DomainName} that you
> > have never typed yourself. This command defines \adthistype. So you should be
> > already happy.
>
> No, I'm not. I need this in the generated .spad (or .as) file, i.e., here
>
> +++   \begin{addescription}{Generates all isomorphism types.}
> +++   \end{addescription}
> +++   \begin{ToDo}
> +++     \begin{rhx} 14-Aug-2006:
> +++       It is not yet clear what the type of this function/constant
> +++       should be.
> +++ %
> +++       In general, isomorphism types are equivalence classes of
> +++       structures. It could be reasonable to say that \adthisname{}
> +++       returns (unique) representatives of these classes. (The
> +++       \emph{unique} is probably a though thing, since we might have no
> +++       order on the input set or on $L$ in general.
> +++     \end{rhx}
> +++   \end{ToDo}
> --#line 384 "species.as.nw"
> isomorphismTypes: Set L -> Generator %;

Hmm, I don't understand why? The function name and its type appears at 
the bottom. So the information is there. I don't have any glue which 
(existing) tool you want to use in order to process the +++ stuff from 
the .spad or .as file. Be a bit more precise.

> I'd like to have ALLPROSE instead of \adthisname{} write
> \spadfun{isomorphismTypes}. And, if possible

Would you be happy with the \addefinename{...} thing at the beginning. I 
don't know yet whether this is easily achievable, but I have the feeling 
that it would be easier.

But still, what use is it to have that stuff inside the spad file if you 
have no tool to process it, or have you?

> > > ... change the regexp in a way that instead looking for a semicolon, it
> > > just goes on until it finds a newline without a preceding underscore.
> > I don't want a such a thing, I want a semicolon. ;-)

> I know, Ralf. Me too. And I'd like to have at least a promise from Stephen Watt
> and Algebra.as and SumIt.as

Then you should write to Stephen for compiler and libraries and to 
Stephane Dalmas for SumIt.

Ralf