yade-dev team mailing list archive

[Vaclav Smilauer <eudoxos@xxxxxxxx>]

> r2469 modified slightly the doc of 
> Law2_ScGeom_NormalInelasticityPhys_NormalInelasticity, in this sense.
Thanks.

>  I 
> emphasize the fact that I mention (it was already the case before) a 
> python script test which is supposed to illustrate the effects of the 
> law. Me, I consider this as a very good mean to present the contact law.
Ok, that is correct, but script serves a different purpose. That person
asked me what constitutive laws were available in yade (and has never
used yade before), and for a brief overview the mathematical formulation
is the most compact information you can give. You can understand without
running simulations, just reading off the screen in a browser.

> - about the requirement you mention, I wonder if the "understand how 
> they work" relates more to understanding of special features of 
> different laws, or more to understanding of the general way to use any 
> contact law (the definition of the InteractionLoop, with correct 
> functors..). Corresponding needs in doc would not be the same.
Of course for each law, since it is a documentation for each contact
law.

> - you point out the fact that non-academic people will more and more use 
> Yade, and that therefore exhaustive documentation is required. For me 
> this has to be seriously considered. I consider that financial and human 
> required ressources are radically different between a research 
> open-source code, and a general public software. And we all know on 
> which side Yade is...
> I do not wish to anybody the feelings I had when I began with Yade with 
> no docs, but on the other side, it is difficult for me to imagine that 
> someone gets engaged in a customer service for all his life, once he 
> typed one "bzr commit" (few exageration here...)
You have a point, but it is a matter of quantity. Good open-source
project depend on their users/develoeprs and those can be invited or
chased away (see the difference betwee the attitude ofpeople around llvm
and gcc, for instance). You might say this is a marketing, and it is
true in the sense as that is how to get people involved. But in the long
run, it is also what will make the project survive longer. SEE

> To conclude, I agree with your idea of "experimental/undoc" feature 
> (with the risk that this feature will dominate in Yade ?)
Good to have your opinion, let's see how it evolves. 

Cheers, v.