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

[glen e. p. ropella]

Paul Johnson wrote:
> I don't even understand why this discussion is going on.

Steve Railsback wrote:
> (Then, feel free to provide feedback on the development projects
listed above.)

This discussion is going on because the SDG asked for feedback on those
4 projects.  I'm providing my feedback.

> If somebody wants to do some work, go ahead.  Otherwise, not.

Agreed.  If someone wants Swarm ported to 64-bit architectures or the
Cell, then go ahead.  Otherwise, not.

> Now, as for the claim that Swarm itself is over-engineered, I would
> respond this way. When I started with Swarm, before Marcus came on the
> scene, I could not get Swarm built in Linux. Building Swarm was a
> completely non-standard experience, and none of the GNU concepts of
> software packaging were used.

That's simply not true.  Swarm had always been a fairly standard unix
download, make, run package.  When we added the Java layer, things
started to go wonky.  I think what you see is really the fact that when
you started with Swarm, you weren't a programmer.  Now you are.

> Swarm.  It is not over-engineered.  It is engineered to actually work
> and build.  And the design is clear enough so that people can come along
> and add things.

And it's _not_ clear enough that people can come along and add things.
Can you subclass from List?  Can you implement phased objects?  (I
suspect you can because you're one of the rare users who has invested a
great deal of time and energy into Swarm.  But, step outside your own
situation for a minute and consider if you were new to Swarm and wanted
to add to the Swarm codebase?)

> But maintainers seem to quit maintaining, we can't put
> those into the core of Swarm unless someone is willing to maintain them,
> and nobody is.

The smart thing to do would be to look at _why_ nobody is willing to
maintain them.  Why do you think that is?  I have three primary
hypotheses: 1) Swarm is rapidly becoming useless (for whatever reason),
 2) Swarm is too perverse for many programmers, and 3) Swarm's
development culture is dominated by 1 person.

> Same is true for the applications in the
> swarmapps-objc.  There could be more there, but nobody steps up to
> maintain old programs like "market".
> It may be that there are many better ways to build models, if so, tell
> me how. Give some code, show a product.

How?  Use Mason.... or Repast, or SimPy, or any number of other projects
being executed by people responsive to their user communities.

-- 
glen e. p. ropella, 971-219-3846, http://tempusdictum.com
And yet doubt is a good servant but a bad master; a perfect mistress,
but a nagging wife. -- Aleister Crowley,  "The Book of Lies"