Re: MathML-Reduce Web Interface

Hello Robert, 
Thanks for having tested my program and made the effort to comment on it. 
On the first point you arise, I have had some problems. 

> 1.  You seem to be generating <relation> ... </relation> tags instead
>     of <reln> ... </reln>.  For example,
>     tay(x*sin(x),x,5);
>        returns
>     <math>
>        <relation><eq/>
>           <apply><plus/>
>              <ci>x</ci>
>              <cn type="integer">6</cn>
>           </apply>
>           <apply><plus/>
>              <apply><fn><ci>y</ci></fn>
>                 <ci>x</ci>
>              </apply>
>              <ci>y</ci>
>           </apply>
>        </relation>
>     </math>
My program use to
generate the <reln> tags, which seem to me the good solution. But for some
reason one day, I was looking at the spec and in section 4.4.4 of the
online spec the <relation> tag is used. I believe you are right, but I am
a bit confused with the tag on the spec. So which one do you think I
should really use?

> 3. The previous example also points out another problem, though in
>    this case it isn't clear to me if the culprit is your code, WebEQ,
>    or the MathML spec.  I am inclined to blame the spec.
>    The problem is that the <minus/> element indicated by the *****
>    above is followed by a single child.  WebEQ throws an error, based 
>    on the fact that </minus> is listed in section 4.2.3 of the spec as
>    a binary operator.  Oops.  Where does that leave the unary minus
>    operator.  As far as I can tell, it isn't there...
Concerning the <minus> element, me too I have some doubts. When I started
the generator in september, there was an appendix F on the mathml spec,
which has now disappeared. And on it was explained every operator with a
lot of information concerning it. The <minus> operator was qualified as
both binary and unary. I used that definition. When this appendix
disappeared, I faced the problem that unary <minus> was no longer
existant, so I just kept it for its usefulness. But maybe I should
consider asking some MathML expert what the official answer is...

Thanks a lot once again for your helpful comments,

Luis Alvarez
Konrad-Zuse-Zentrum fur Informationstechnik Berlin (ZIB)
Optimisation Division

Tel: (+49 30) 84 18 53 34

Received on Wednesday, 13 January 1999 04:36:03 UTC