Re: [igraph] Language remmendations

[Gábor Csárdi]

Tom, I will be short, because I think there is not much to say, Tamas
already wrote what he thinks is relevant about the Python interface
and I don't know any better than him.

Personally I think that if you want to do math and you know R fairly
well, then there is no point in switching to Python. Both R and Python
are nice languages with extensive sets of third party packages. The
main difference I see is that the R functionality is shifted towards
math and statistics. What Python offers is a more diverse set of
extensions (e.g. if you want your program to talk to web-servers
directly, then Python is a better choice) and a somewhat cleaner
language (IMHO). If you want to create sophisticated figures, then R
has an advantage in this, I believe.

In summary, I would say go with the language you know better. But of
course learning Python is fun,
Gabor

On Fri, Jan 30, 2009 at 9:55 PM, Tom Backer Johnsen <address@hidden> wrote:
> Gentlepersons:
>
> I need a tool for the analysis (and perhaps simulation as well) of social
> networks, in particular small social groups (10+- 2 members or smaller) in
> respect to concepts like balance (Heider, Cartwright & Harary etc.).  In
> other words, the graphs I will be working with are signed (valued) and
> directed graphs.
>
> It seems to me that a scan of the literature reveals one serious alternative
> in two different versions, igraph for R and Python.
>
> I am fairly familiar with R, much less familiar with Python (but I
> anticipate few problems in learning a new programming language), and the
> latter one in any case looks like a nice and useful language.
>
> I have downloaded the documentation for the R version of igraph package.
>  However, I have not been able to locate anything specifically for the 2.5
> version of Python.  Does that mean that the calls on functions are
> essentially the same?  Is it possible to find sample Python program snippets
> on the net?  Also, the R documentation I have is at least half a year old,
> which may make a difference, given the fair amount of activity on this list.
>
> For the possible (type agent-based) simulations I am thinking of using
> Netlogo.  R should be an execellent tool for any data analysis of the
> results from any simulations.  So, what I am looking for is some kind of
> advice (or opinions) in respect to the most efficient combination of these
> tools.  For instance, would it be advisable to drop Python altogether and
> exclusively work with a combination of Netlogo and R?  Or is the versatility
> of Python useful?
>
> Tom
>
>
>
> _______________________________________________
> igraph-help mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/igraph-help
>



-- 
Gabor Csardi <address@hidden>     UNIL DGM