[pstricks] Re: pst-eucl (integrating Metapost into Pstricks :-)

[Georges Mariano]

Maurice DIAMANTINI wrote:
> 
> Bonjour,
Hi Maurice,

> I answer to all because I think it could generate a more
> general discution about integrating METAPOST INTO PSTRICKS.
???

> (Excuse my bad english!)

> I've not use your package (no time for the moment)
> but I've read the doc.
idem... (well, I use pst-eucl in a (little) programming project
(read the description in french (sorry) here :
http:www3.inrets.fr/estas/mariano/Euclide2)
 
> I'd like it to be included in the core of Pstricks!
> But if it is, all the command names have to be cleaned
> to be more Pstricks complient.
I would like to follow you on this path but you have to define just one
little thing : 
what does it mean pstricks complient ??
There is no naming policy defined somewhere in the doc (as far as I
know...) !!
And, more important, the naming of pstricks command don't need to be
defined from an implementation point of view (pstricks complient) but
from a user point of view, and I don't know if it is possible or
"compatible" (i.e non contradictory) ...



>   so all commands should begin by \geo... 
yes

>   Defining nodes:
do not agree with keeping "Node" in the names (again, user's point of
view), there is no node in euclidian geometry just "points" and "lines",
so keeping the word is just related with pstricks...
How do you explain to a mathematics teacher why s/he should use names
like \...Node... ??

 
>     \geoMarkxxx{}
>     \geoPutxxx{}
>     \geoSegxxx{}    (for Segment)
>     \geoTrixxx{}    (for Triangle)
>     \geoCirxxx{}    (for Circle)
>     \geoAnglexxx{}
>     \geoDistxxx{}   (all what defined a distance)
>     \geoBoxxxx{}    (for all predifined boxes or point shapes)

>   Also you provide many commands that do both : defining
>   a node and put something: this should be considere as
>   a facility to do two thing in a single step.
Again, you do not have to split a command which is doing two things into
two commands, you just have to check that this command is defined using
two "atomic" commands (which may then be used by developers)...
Users usually need "facilities", they don't understand why they have to
wrote two lines if they "see"/think that one is sufficient...
 
>         drawGuideLines (ligne de rappel en français :-)
(ligne de construction ??)

En français : Certes il est probablement nécessaire de redéfinir
certaines choses dans pst-eucl (mais j'ai pas fais assez de tests pour
juger) cependant il faut faire attention à bien garder une utilisation
intuitive de la chose (c'est, je pense, le premier soucis de Dominique).
Globalement, il ne faut pas confondre facilité/qualité de
l'implémentation en pstricks (=> nécessite des ajustements dans
pst-eucl) et facilité/qualité de l'utilisation (nécessitent d'*autres*
ajustements).
Il faut améliorer les deux simultanément, c'est la difficulté...

PS : Ceci dit, je pense que l'utilisation de tels packages ("grand
public") passe plutôt par la définition d'un langage "abstrait"
permettant de décrire/spécifier les figures en dehors de pstricks et de
n'utliser ce dernier que pour la restitution, c'est ce que nous avons
explorer dans le projet de licence informatique cité plus haut...
la conclusion étant : "ça marche!" ;-)

A+

 
-- 
# Georges MARIANO               tel: (33) 03 20 43 84 06
# INRETS, 20 rue Élisée Reclus  fax: (33) 03 20 43 83 59
# 59650 Villeneuve d'Ascq       mailto:georges.mariano at inrets.fr
# FRANCE.                         
# http://www3.inrets.fr/estas/mariano/OCamlDebian
# http://www3.inrets.fr/B@INRETS/B-Bibliography