[Coin-discuss] Re: Coin-discuss digest, Vol 1 #397 - 2 msgs

[Ronald Hochreiter]

Erling D. Andersen wrote:

>Hi
>
>  
>
>>I have not been working on FLOPC++ for the last year, partly because of lack of time and partly because of lack of motivation (due to low rate of feed-back). If there are some people out there who are interested in developing an AML based on FLOPC++, I would certainly be motived and would find the time to participate. 
>>    
>>
>
>Until I made my own optimization product MOSEK commercial and promised support I got virtually no
>feed back. This was very disappointing. Maybe "free" is not a virtue in itself.
>
>Regards
>
>Erling
>
>
>  
>
Open Source: The problem seems to be that the community for optimization 
in general is a bit too "small" for supporting single open-source 
projects. (Small in comparison to open-source projects for e.g. a 
Database MS like mySQL.) That's why COIN is really a good idea to get 
something open-sourced in the whole area of OR. However, there would 
also be the option (e.g. for the MOSEK solver) to offer two licenses, 
e.g. like mySQL mentioned above. It's GPL, but there are commercial 
licenses available, for those who need commercial support.

FLOPC++: I am not convinced that embedding the modelling language into a 
programming language is generally the best idea. My favorite combination 
is using a modelling language like Zimpl (or AMPL) and if I need complex 
looping I write a small script with some scripting language like Python. 
That way it is also communicateable within courses for e.g. business 
students, who definitely are not going to learn C++ just to model their 
supply chain problems. And what about maintainability?

Regards,
Ronald

==========
Ronald Hochreiter
Department of Statistics and Decision Support Systems,
University of Vienna