[LON-CAPA-users] square roots in formula response

[Peter Riegler]

Hi,

two issues I want to add.
o I am very pro the generic xml structure. Even more, I imagine using a 
stronly typed xml-structure which reflects the semantics of mathematical 
expressions (e.g. <symbol>pi</symbol> vs. 
<areaofunitcircle>pi</areaofunitcirle>. Horrible, but very useful!).
o Besides all this I fear we will never achieve in abstracting the CAS 
away. They all have their peculiarities which actually are useful. 
Abstraction only can cover the intersection of functionalities of all 
CASes.
Moreover, two different CASes might come with identically named commands 
doing the same things according to the manual pages and yet do not do 
exactly the same. trigreduce on maxima and Mathematica is an example of 
this.
If you want to use a CAS in your problem, you have to know about CAS.
If you want to use the old-style formularesponse, you have to know about 
how it works.

As usual these are competing issues. The challenge is to find a suitable 
compromise. Maybe all this discussion is better located at LON-CAPA-dev? 
(Before I become to philosophical ...)

Peter

Gerd Kortemeyer wrote:

> Hi,
> 
> On Sep 4, 2007, at 8:12 AM, Peter Riegler wrote:
> 
>>
>> Also this issue here brings up my uneasy feelings that lonmaxima's
>>    is(trigsimp(trigreduce(RESPONSE-ANSWER))=0);
>> is not a good idea. It should be
>>    is(RESPONSE=ANSWER);
>> If an author wishes/needs to have simplification it should be his  
>> task to put in the proper/desired command using the 2nd interface  
>> called mathresponse.
> 
> 
> Well, at the moment we are trying to mimic the old behavior of the  
> sample-based <formularesponse>. Most authors would expect that  
> <formularesponse> recognizes sin^2+cos^2=1 or sin/cos=tan without  
> explicitly being told to do so.
> 
> <formularesponse> is a crude tool.
> 
> Several points:
> 
> * we don't want to completely bind ourselves to one computer algebra  
> system, so on general purpose tags we need to abstract the details of  
> MAXIMA
> 
> * general purpose tags need to stay generic and do what most authors  
> expect
> 
> * as much as possible, what the tag does needs to be reflected in the  
> XML, not in writing some sort of code
> 
> * we can add more specific tags easily. For example, we *could* (I am  
> not suggesting right now that we should) add a tag for  
> <mathequivalencyresponse>, which could in its XML definition:
> 
>  - specify the simplifications that should/should not be done
> 
>  - specify mandatory and prohibited structures
> 
> I am for keeping <formularesponse> the generic "catch-all" way it is,  
> and reducing authors' dependence on the cryptic <mathresponse> and  
> <customresponse> by making a more generally configurable mathematical  
> equivalency tag in between the two ends of the spectrum.
> 
> - Gerd.
> _______________________________________________
> LON-CAPA-users mailing list
> LON-CAPA-users@mail.lon-capa.org
> http://mail.lon-capa.org/mailman/listinfo/lon-capa-users

-- 
Peter Riegler
Fachhochschule Braunschweig/Wolfenbüttel
Salzdahlumer Str. 46/48, 38302 Wolfenbüttel
Tel. +49 5331 939 6314, Fax. +49 5331 939 6002
email: p.riegler@fh-wolfenbuettel.de
http://public.rz.fh-wolfenbuettel.de/~riegler