[Axiom-mail] Troubles with subscripts and solve()

[Соловьев Александр Александрович]

Hi again, list!

Some history.

Polynomial has been made like this

> f:= reduce(+, [ reduce(+,[ subscript(a,[i,j])*x^i*y*j for i in 0..2]) for j 
> in 0..2 ]) ::HDMP([x,y], ?);

Built series of equations 

> concat([[ eval(f, [x=i,y=j])=subscript(m,[i,j]) for i in -2..2 ] for 
> j=-2..2]);

and solved by

> solve(%)

Ok, result is List List Equations with expected  solutions and some trash. If 
resolving variables pointed explicitly by

> solve(%, concat([[subscrpit(a,[i,j]) for i in 0..2] for j in 0..2]) ) 

solver just eats CPU time for a long time and nothing gives. But in case of 
coefficients of form a00 and m11 (i.e. without subscripts) solver gives 
expected results without trash ( btw trash is redefinition of a_i_j variables 
via some intermediate ones %AA %AB and some additional equations like m_1_1 = 
%AA + %AB + ...). 

First question, is there some workaround in my problem? I'm expecting clear 
result like in second case.

Second. In case solver gives clear result, how it possible to substitute 
founded values aij into source expression f?

Thanks.