[Maxima-discuss] indexes and summation conventions / subscripts etc

[Richard F. <fa...@be...>]

Just because a notation is in use in paper/pencil applications
does not mean it is ideal or even suitable for automated computation.

For example Leibniz notation for integrals would have us write
integral_sign  f(x)dx

but we use integrate(f(x),x)

the notation for differentiation  as   f'(x),  f''(x)   or totally 
bizarrely    d^2f(x)/d^2x  or  f with dots above,
or d_x{f}  etc ..
  could also be used,

but we use diff(f(x),x,2)

It works, it isn't too confusing for humans or computers.

The Einstein convention may be handy, and it may have been more-or-less
reasonably implemented in Maxima programs (I don't do relatively for a 
living)
but in general it's not necessary to mimick exactly what works for 
paper.  Humans
are more flexible than that, and a notation that doesn't play well on a 
computer
should not be used by (physicists?) to insist that programmers should
  hack endlessly on the program.  Especially if another, more
suitable  (unambiguous in the context of other supported mathematics, say)
is available.
A mathematician of my acquaintance insisted that he couldn't use 
Macsyma/Maxima
because the computer terminal I offered him was lacking in keys for 
Greek letters.
Well, he was right in that indeed he couldn't use Maxima.

Anyway,  I just wanted to point out that the notation of super and 
subscripts is already overloaded.
If Maxima is to include 10 different (usually distinct) parts of 
mathematics in one unified
system, each of which has some semantics associated with _scripts, there may
not be a system definition that makes it ALL work at the same time in a 
universal
context.


RJF



On 12/16/2015 7:45 AM, Martin Marmsoler wrote:
>
> In my version it works with ctr+space and ctr+shift+space but thats 
> good it's like in eclipse
>
> Am 16.12.2015 12:14 nachm. schrieb "Michel Talon" 
> <ta...@lp... <mailto:ta...@lp...>>:
>
>     Le 16/12/2015 06:13, Gunter Königsmann a écrit :
>     > wxMaxima has autocompletion: Ctrl+Tab autocompletes a command,
>     Shift+Ctrl+Tab shows its parameters.
>
>     At least with the version of wxmaxima which come with my LTS
>     version of
>     Ubuntu, these commands do nothing (in fact insert a Tab). I have other
>     problems, for example Ctl-C Ctl-V don't work properly between wxmaxima
>     and other windows, etc. Some menu entries crash the program.
>     I am under the impression that wxmaxima is mainly tested under
>     Microsoft
>     Windows, and probably Mac OS, but free operating systems are second
>     class citizens.
>
>     To come back to the position of indices, i think it is ill advised to
>     denote x[i] as x_i when i belongs to [0-9] as seems to be the case in
>     this commit to wxmaxima. Indeed the index is frequently a letter
>     and can
>     be either upper or lower (covariant or contravariant). To see typical
>     usage, see for example:
>     https://en.wikipedia.org/wiki/List_of_formulas_in_Riemannian_geometry
>     If you look for example at the Riemann curvature tensor R, the first
>     index is upper and the three other ones are lower, R^l_{ijk}. This
>     is a
>     property of R and not of ijkl. For example the Ricci tensor is
>     R^l_{ilj}
>     with implied summation on index l which appears both upper and lower
>     (the so called Einstein contraction) so the position cannot be a
>     property of the index, but of the main symbol R. Hence i like much
>     better the proposition of Robert Dodier which describes the
>     position of
>     indices as property of the main symbol. Of course his way of
>     describing
>     this position requires much typing, which i don't like. Some
>     abbreviated
>     form like in the TeX usage would have better chance on concrete
>     use, for
>     example uddd for the R above, one could imagine
>     d|d for one one pre index and one post index, both down, etc.
>
>
>     --
>     Michel Talon
>
>
>     ------------------------------------------------------------------------------
>     _______________________________________________
>     Maxima-discuss mailing list
>     Max...@li...
>     <mailto:Max...@li...>
>     https://lists.sourceforge.net/lists/listinfo/maxima-discuss
>
>
>
> ------------------------------------------------------------------------------
>
>
> _______________________________________________
> Maxima-discuss mailing list
> Max...@li...
> https://lists.sourceforge.net/lists/listinfo/maxima-discuss