Re: [Axiom-developer] Re: article "standard" header/footer

[Ralf Hemmecke]

Hi Martin,

Could you set up a page on MathAction of this discussion thread? Or 
better, could someone tell me how to create a new page.

What you state here are some visions that should be recorded and not be 
forgotten in the mailinglist-archive.


Martin Rubey wrote:
> Ralf Hemmecke <address@hidden> writes:
>
> > Maybe I should learn about the term "executive overview". I am not quiet sure
> > what "executive" stands for in this phrase. Isn't it a kind of "Introduction"
> > what Martin wants? So what could possibly be the difference between
> > "Introduction" and "Executive Overview"?
>
> I'm not quite sure what I wanted some months ago, but I suggest the following:

> Many contributions consist of several packages, categories and domains, which
> are somehow related.

You see, I also somehow realised that. The Axiom Library would grow in 
bigger steps, not just one cat/dom/pkg at a time. We should start to 
think in Aldor terms. A certain part of mathematics is covered by a 
number of packages, categories and domains. These form one small library 
and are documented together. ALLPROSE could be the environment to 
describe the whole library. And that could form a (refereed) paper in an 
Axiom journal.

> * Maybe we also need an indroduction that *roughly* describes the mathematical
>   background.

Of course together with references to the literature.

>   This would be everything up to "Using the package" in my Guessing
>   package example and would not contain details. The details should probably go
>   along with the program. This would not contain code, only mathematics.

I agree that it is sometimes easier to read the mathematics, but when it 
comes to the algorithmic part... why not writing real code instead of 
pseudo-code?

> I think that the way these bits of documentation are presented in HyperDoc is
> just perfect. I don't think that it would be too difficult to make ALPROSE do
> the same. In principle, HyperDoc is a collection of hyperlinked pages -- just
> as ALPROSE when viewed in a browser and not a dvi-viewer.

My problem with HyperDoc is that there are quite a number of commands 
and no good documentation what they are for. I would clearly prefer my 
aldordoc.sty even though it does not cover all things HyperDoc can do.

> I.e., we would need to devise a tex4ht interface for ALPROSE that suitable
> splits the input into a tree of subpages, which can then be accessed by
> selecting an item from the index...

I've already read in the tex4ht documentation that it can split 
according by \section and \subsection etc. So the most important thing 
is to produce HTML at all.

> Finally I'd like to stress that I strongly believe that AxiomUI should use the
> same file format as the documentation. There is no semantic difference between
> a worksheet and a literate program, except that a worksheet will be
> interpreted, not compiled.

Sorry that I am a bit behind, but I have tried to install jsmath and 
somehow I remember that my Firefox claims that some font is missing. So 
I gave up and thus I have never really seen AxiomUI. On which files does 
it work? Is it a replacement for HyperDoc? Or what is the relation? Is 
there a link to a documentation about AxiomUI.

Ralf