Re: [Axiom-developer] hyperlinked algebra

[root]

whew.... 

hey, it's open source. change it any way you'd like.
you have nothing to fear but the feedback :-)

>  > i've hacked src/doc/endpaper.pamphlet to add hyperlinks.
>  > most of the links go to catdef.spad.dvi because they are
>  > categories that live there. i don't know how to implement
>  > #tags in latex yet.
> 
> I just had a quick look at your file. It's cool that hyperref can also 
> refer to dvi files by just saying something like
> 
> \href{catdef.spad.dvi}{abelsg}
> 
> I was always looking for something like that. I've never tried it out 
> though. Thanks for showing that to me.
> 

glad to be of service.

> HOWEVER, I would not have written the links in endpaper.pamphlet as you 
> have done. OK, I understand that endpaper.pamphlet is just a quick hack.

we could take the endpaper example, make it dynamic, and put it on the wiki.

this is much less of a quick hack but much more useful.

> I would have proposed just to write \adtype{AbelianSemiGroup} (sorry 
> that I borrowed the \adtype tag from ALLPROSE, but it is just a tag 
> after all). In the definition of \adtype one would have to take care 
> about the file in which AbelianSemiGroup is defined. That frees you from 
> changing anything in endpaper.pamphlet (or any other file that refers to 
> AbelianSemiGroup) if someone ever thinks of moving that definition to 
> some other file. One just had to redefine \adtype (and even that could 
> be done automatically).

\adtype? since this appears to by your macro i can't comment.


> Next, don't use the abbreviation in the \href command. That could be 
> automatically extracted from AbelianSemiGroup, if you really want to see 
> the abbreviation. (I would rather be happy with the full name.)

the original end papers on the inside cover of the jenks book used
full domain names. however they had the luxury of 2 pages of space.
my attempt at reproducing the information ran into the edges. to
save space and still keep the information i was forced to use 
abbreviations. full names wouldn't fit.

> Furthermore, if you put ".dvi" into the link you basically restrict 
> yourself to dvi output (unless you redefine \href to swallow the 
> extension ".dvi"). Why not having everything, dvi, ps, pdf, html, ...?
> With ALLPROSE I now can produce the first 4 reasonably well.

well i tried http urls to the wiki but those failed (at least in xdvi)
so i played around more and found i could get to local .dvi files.
since i was only trying to illustrate an idea i was advocating the
local files were sufficient. bill has already taken it a step further.

> Every information is already in "AbelianSemiGroup" some scripts should 
> be responsible for making the links correct. I am very much against 
> hardcoding references to files if you actually mean a reference to a 
> domain/category.

the endpaper.pamphlet was an example to show what i was talking about.
you're welcome to modify it any way you'd like.

actually the latest "grand scheme" is related to Bill's question about
parents/children. we could certainly hack up a routine in lisp or
aldor that, given a domain, could spit out the parent/siblings/children 
and then dynamically create the file containing the graph. thus a user 
could point at a domain and ask for the graph.

the layout is a challenge though. laying out endpaper took about 4 days.

t