axiom-developer
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Axiom-developer] steps towards better TeXmacs interface


From: C Y
Subject: Re: [Axiom-developer] steps towards better TeXmacs interface
Date: Fri, 26 Aug 2005 11:46:35 -0700 (PDT)

--- Martin Rubey <address@hidden> wrote:

> Andrey G. Grozin writes:
> 
> > ... if this C code is used, it is absolutely impossible (and never
> > will be possible) to copy a (part of) output by the mouse and 
> > insert in into an input field (because it is not possible to
> > reconstruct the mathematical meaning of an expression from this 
> > obfuscated LaTeX). And this goal is highly desirable.
> 
> Although this goal may be highly desirable, it will work only for
> some domains. For others it is doomed to fail, simply because not
> all the information is in the output. In fact, it will probably 
> work only for those domains that are ConvertibleTo InputForm...

I am not surprised at this, but I suspect handling it well will be
critically important to the acceptability of Axiom for a wide variety
of applications.
 
> So the probably better way to do this is to capture the label of the
> output. The object is stored there, you don't need to reconstruct it.

[snip]  

Does this imply that for a significant subset of mathematical
expressions, it should not be legal to select a subset of the
expression piecemeal?

I think I know what Andrey is thinking about here.  Let's say we have
the following expression:

                                  2
                                 b  + 2 a
                                 -------- + c
                                    d
                                 ------------
                                    f + e

and we want to select:

                                    2
                                   b  + 2 a
                                   -------- 
                                      d

as our input in the next line.  The natural thing to want to do in a
graphical mathematical document environment is graphically select the
desired subexpression and paste that to the input line.  I suspect the
objection is that there may be hidden or non-hidden constraints within
Axiom's knowledge of the original expression that make selection of the
subset problematic, and won't be preserved?  Perhaps one way to do this
"safely" would be to "flatten" the expression to the basic types you
would get if you just typed in the displayed variables, and warn that
all original mathematical knowledge of the parent object has been lost
in the copying process?

This, however, is a violation of Axiom's philosophy of wanting to "Do
It Right."  The more interesting, but far more challenging problem, is
to map the constraints of Axiom's mathematical knowledge into a
graphical interface environment.  I rather suspect there will be a
fundamental clash between the "beginner mentality" that is often
associated with the use of such environments and Axiom's philosophy,
but that just makes it more interesting :-).  The ideal case would be
to have the ability to make assumptions about one's environment, and as
those assumptions are added or removed the interface behavior changes
according to what is allowed in the new "environment."  E.g., there
will be operations that might be legal if certain assumptions are made,
that are not true in general, and if your environment is less or more
general the interface knows that and acts accordingly.

Not sure that's even doable in the case of TeXmacs - IIRC it's rules
for mathematical interaction are those of a document editor, and not
really responsive to mathematical constraints. It would be difficult
even with something like a lisp interface, because interaction
behaviors would have to be defined, and some kind of mapping made
between those mappings and mathematical behavior.  What a nifty
research project :-).

Cheers,
CY


                
____________________________________________________
Start your day with Yahoo! - make it your home page 
http://www.yahoo.com/r/hs 
 




reply via email to

[Prev in Thread] Current Thread [Next in Thread]