Re: [Swarm-Support] development priorities (was Re: Membership inSwarm Developmen Group)

[Bill Northcott]

Marcus wrote:
> > That's not necessarily true.  Optimization of particular solutions  
> > and
> > particular problems are well-covered by R and/or other packages.   
> > But, I
> > haven't seen any tools that help optimize ABMs.  The amount of  
> > work to
> > do what Scott describes inside R is huge.
> It's a simple matter to control foreign code from R.  Of the  
> hundreds of contributed packages for R, many of them physically  
> include compiled C++ and Fortran.    R can run MPI codes, or SOAP  
> (RPCs over http/XML), and so on.   The point is to leverage the  
> expertise of people that know actually know about these algorithms.

For the little that my opinion might be worth, I am right behind  
Marcus on this

> glen e. p. ropella wrote:
>
> > >  Basically what Mono (http://www.go-mono.com)
> > > or .NET does.   That way there could be a smooth transition to
> > > autonomous development groups we'd get a way from the Objective C  
> > > lock,
> >
> > And it wasn't to "get away from" objective-c.  It was simply to  
> > support
> > our users.
> Indeed, Objective C is not a commodity technology.  As a platform  
> this has meant time that could have gone into creating modeling  
> features that didn't (instead struggling to make the compiler do  
> what we needed).  Some of this complexity was the Swarm project's  
> doing (across personnel) and some of it was just the on again off  
> again nature of Objective C, esp. at Apple.   As an unusual  
> language this meant confusion for users.  Neither is desirable and  
> users have said that.

With the resurgence of Apple from a basket case to one of the most  
profitable companies in the computer business, Objective-C is really  
not marginal.  It is very much alive. The compiler problems really  
are history.  Also the convergence between Apple and GNU versions has  
made it a lot more portable.  Phases might be seen as an issue  
because they will be outside the ken of thousands of new Objective-C  
programmers brought up on Cocoa/GNUStep.  Of course there is the  
small matter of messages to nil, but I promise not to go there again.

I have no idea what the discussion about ports to new CPUs is about.   
Swarm runs quite happily on 64bit machines and if we can complete the  
transition to a standard Objective-C runtime it will be extremely  
portable.

The current build system is very functional, but I agree with Paul  
that when first encountered, it was a bit of a nightmare.  To me the  
problem is lack of documentation.  We needed a 'Swarm porting guide'  
or similar that explains to potential packagers how the build system  
works.

Finally a mea culpa for having made very little contribution  
recently, but I need to finish off a Phd in the next few months and  
the ABM stuff is outside the scope of it.  However, it is very much a  
follow on and I have collaborators who are interested and can get  
funding.

Bill Northcott