Thread Subject: subs in R2008b; matrix concatenation impossible?

syms a b
m = [a b]
 
m =
 
[ a, b]
 
syms a1 a2 b1 b2
a = [a1 a2]
 
a =
 
[ a1, a2]
 
b = [b1 b2]
 
b =
 
[ b1, b2]
 
subs (m)
 
ans =
 
[ a1, b2, a, b]

"Joerg Buchholz"
> syms a b
> m = [a b]
> syms a1 a2 b1 b2
> a = [a1 a2]
> b = [b1 b2]
> subs (m)
> [ a1, b2, a, b]

yet another bug in a long list due to an inconceivably ill-considered move by TMW to prematurely exchange a well-working engine with an inferior product (i cannot remember that users were consulted ahead of such a far-reaching decision)...

this does not make it better

     syms a b a1 a2 b1 b2
     m=[a,b]; % as above
     a=[a1,a2];
     b=[b1,b2];
% now
     subs([a,b])
% ans = [ a1, a2, b1, b2] % ok
     subs(m)
% ans = [ a1, b2, a, b] % as above, works as expected in r2008a
     m=[a,b]
% m = [ a1, a2, b1, b2] % ok

joerg, did you already create a bug-report?
us

"us " <us@neurol.unizh.ch> wrote in message <ghtf25$avm$1@fred.mathworks.com>...
<snip>

> joerg, did you already create a bug-report?
> us

I did. (#8 of R2008b ;-))

"us " <us@neurol.unizh.ch> wrote in message
<snip>
> yet another bug in a long list due to an inconceivably ill-considered move by TMW to prematurely exchange a well-working engine with an inferior product (i cannot remember that users were consulted ahead of such a far-reaching decision)...

After a few hours of playing around with the MuPAD engine I get the feeling that the new symbolic engine is not really the problem. It seems to be slower but more accurate than Maple. The problem is the lousy MuPAD-Matlab interface; which is good news, because I am optimistic that TMW developers will soon be able to debug the interface without the need to touch the MuPAD kernel. ;-)

"Joerg Buchholz" <buchholz@hs-bremen.de> wrote in message <ghtlpn$e44$1@fred.mathworks.com>...
> "us " <us@neurol.unizh.ch> wrote in message
> <snip>
> > yet another bug in a long list due to an inconceivably ill-considered move by TMW to prematurely exchange a well-working engine with an inferior product (i cannot remember that users were consulted ahead of such a far-reaching decision)...
>
> After a few hours of playing around with the MuPAD engine I get the feeling that the new symbolic engine is not really the problem. It seems to be slower but more accurate than Maple. The problem is the lousy MuPAD-Matlab interface...

yes, that's why i used the word ENGINE rather than MUPAD - thus referring to the whole wrapper/sym-package - which is what the ML user experiences, after all...
i don't want to play with mupad, i want to use ML's tbx, which costs a lot of money...

us

The bug has not been fixed in R2009a.