swarm-hackers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [swarm-hackers] Re: Host platform dependence (was Naming conventions


From: Bill Northcott
Subject: Re: [swarm-hackers] Re: Host platform dependence (was Naming conventions)
Date: Thu, 19 Nov 2009 19:02:09 +1100

On 19/11/2009, at 6:15 PM, Scott Christley wrote:
>> I really can't go with that.  I don't get this as quantum physics.  IMHO the 
>> MVC paradigm is seriously useful, which is part of why Rails is so sweet.  
>> Violating it by mixing the M and the V leads down a path beset by dragons.  
>> It may be valid, but it is very perilous.
> 
> I understand your point, though you are confounding programming concepts with 
> modeling concepts.

I am not convinced about that.  In this instance, the model is a program.

>  The notion of "observation" can still be part of a model, that is valid.  
> Actually I think its in many models but not made explicit.  For example in 
> Heatbugs, each of the bugs observes its environment and makes decision about 
> how to move.

I am quite happy that there can be observation which is endogenous, but this 
has has nothing to do with the GUI which is exogenous - unless of course we 
have a web interface and some human agents (I'll think about that later).
> 
> Swarm is designed with MVC, Cocoa is designed with MVC.  I'm not saying the V 
> in MVC is part of the model.  But the same Swarm classes that are used by V 
> to run a GUI display for a user, could just as well be used by agents within 
> the model to observe themselves and their environment.

Classes or even simple C functions may be part either the M or the V.  Code 
extracting data from an object for another model object to observe is part of 
the model.  Whereas the same code extracting data for the UI is part of the 
view.  I don't see the confusion.
> 
> Anyways, I'm being pedantic and this sounds more like philosophic banter.

I think pedantic is good at this point
> 
> Your initial point was that the model in Heatbugs should not be using Cocoa 
> classes.  It does not.  In fact the model classes for the OpenStep version of 
> Heatbugs are practically identical and unchanged, what is different is the 
> Observer class or the V in MVC.

Yet again that model word.  I am interested in the Heatbugs app as an example 
of what a modeller might write.  Within that app we should encourage separation 
of model and view code n the programmatic sense.  If the Observer class is in 
the Heatbugs app code rather than in the Swarm library, then I have a problem 
with it containing host specific code.  Just as you had to abstract the 
runtime, I think we probably need to abstract some of the observer type classes 
in the library to keep the Heatbugs code portable.

I think this is all Swarm 3.0 stuff and maybe we should have release branch and 
a todo list aiming towards a 3.0 release.
> 
> This is true with one minor caveat.  Heatbugs know how to draw themselves, 
> thus they contain GUI code within them to do this, the GUI code to do this 
> for tcl/tk/blt is different from the code for OpenStep.

That is horrid, just what I so don't like.

I need to find more time to work on this rather than just prattle about it.

Bill

Attachment: smime.p7s
Description: S/MIME cryptographic signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]