← Back to team overview

yade-dev team mailing list archive

Re: renaming classes

 

Vasclav, Janek, Bruno,

Well, I am here and I am concerned about a good "typology" concerning the class name. When Olivier Galizzi developed YADE first , he had a "mystic vision" about naming classes and it was almost impossible for him to understand that YADE could be used by standard persons. I would like someone (whatever is his background) who starts with YADE, not being lost by annoying and incoherent class names.
Moreover, it will help to write a clear doxygen documentation.
Thank you guys for being so creative and constructive.
Frederic.


Janek Kozicki a écrit :
Václav Šmilauer said:     (by the date of Fri, 13 Jul 2007 09:18:46 +0200)

If you rename classes in two days, many mistakes will be done inevitably.

fortunately we have a week. Or in fact, if it takes longer ... three
weeks. Because that's when I return to poland.


(And making Shape shortened to Shap is ridiculous, it saves 1 (one) letter).

That's also why I prefer to use acronyms :-)

State doesn't exist currently. For the rest; Shape vs. OptimizedShape? Technically, it is SiplifiedShape or RepresentativeShape and that is all too long. For my suggestions, not all of them were serious. The serious ones were State, Shape, Mold, perhaps Material. The rest (Bang, Bump, Smash, Bag) were not. Why YadeSimulation - you should have YadeShape, YadeBodyOptimizedShape etc. for everything then?! Simulation is just fine. The same for BodyVariable, BodyConstitutiveParameters, etc. The cleanest whould be to have Body::State (whatever you call that), Body::Shape, Body::Mold (or something different), ...

SimulationLoopEngine, SimulationLoopDispatcher? LoopEngine, LoopDispatcher would be better perhaps. What about inconsistency between LoopDispatcher and DispatcherFunctor? LoopDispatcherFunctor? I would suggest something like Engine, EngineDispatcher, and maybe EngineFunctor.

Thank you, I have updated wiki with your suggestions. Please edit it as you want.


And for the dispatcher functors: you do the same mistake as you have been doing from the beginning IMO: If you say Sphere2Sphere4SpheresContactGeometry, it is clear that you mean InsteractingSphere, since that is what SpheresContactGeometry uses.

I don't think it's a mistake. I think rather to use acronymys in
naming, that say "it dervies from InteractingGeometry". Bear in mind
that there are other functors also, not related to
InteractingGeometry.

A possible alternative is to make subdirectories for each type of
functor, depending on from where they come, but I don't like it, and
maybe be impossible, because you can't have 2D directory structure.


Remember that you should create a bash/perl/whatever script that will mass-replace the names everywhere (including actors, where the names are quoted). Ideally, we should have something like "convert-yade --from 0.11 --to 0.12 *.?pp" that would do that and be maintained.

of course, we need a good way to convert. I won't do it by hand, forget it.

As a side-note, most classes still lack any doxygen documentation. That is much more serious than their names and Frederic should make you focus on that (Fred, are you there?! ;-) ).

Well, currently Frederic is more interested in submitting an article
(with final versions of class names).



-

_______________________________________________
yade-dev mailing list
yade-dev@xxxxxxxxxxxxxxxx
https://lists.berlios.de/mailman/listinfo/yade-dev

Follow ups

References