generalized indices, was: greek letters



  Hello,

Robert Dodier <robert.dodier at gmail.com> writes:

> On 2013-02-09, Aleksej Saushev <asau at inbox.ru> wrote:
>
>>  - support for indexing expressions in some way, say
>>     {\frac{\partial G} {\partial {n_k}}}_{n_{i\neq k}};
>
> For the record, Maxima allows general expressions to be subscripted:
>
>   ('diff (G (n [k]), n [k]))[n [i # k]];
>
> I don't know what Maxima can do with such expressions ....

In thermodynamics you _live_ with such expressions.
You don't have any complex variables and problems associated with them,
but you have _a_lot_ of partial derivatives, constraints on functional
dependencies and independencies, a lot of linear constraints arising
from conservation laws, and some convex constraints arising from
stability conditions.

>>  - support for other indices, e.g. mu^{sol}, mu^{liq}, mu^{st};
>
> Indices can often be superscripts, as in this example, as well as
> subscripts and superscripts written before the indexed symbol. It would
> be interesting to allow indices in any or all of four positions (sub /
> super, before / after). An index written as a subscript after is handled
> well by Maxima at present, and as a superscript after (although not as
> well, as it should not be treated like an exponent). What are some
> possible notations for indices written before? How about [sol]mu for
> subscript before and [sol^]mu for subscript before? (Obviously we'll
> have to convince the parser to go along with that.) How about mu[^sol]
> for an index superscript after, to distinguish it from an exponent? If
> someone likes some other notation better, by all means, let's hear
> about it.

After some reflexion I think that this is of very little importance.
I'm fine with writing all of those as mu_sol, mu_liq, mu_st as long as
really essential problems are addressed.

>>  - some auxilliary things like declaring functions homogeneous on a
>>    set of arguments
>
> Yes, this is interesting too.

I think that it is relatively easy to generate necessary properties for
partial derivatives. Thus this may be a minor problem when compared to
general support for partial derivatives.


-- 
HE CE3OH...