[kepler-dev] modeling question about producing f(A, x), f(A, y)...

[Timothy McPhillips]

Hi Norbert,

I understand.  Because the COMAD framework is not in distribution  
yet, I agree it's not a great solution for a problem that needs to be  
solved by 5 PM today.  However, in projects where most  workflows  
involve these kinds of issues (as all of mine do), it may be worth  
looking at even now.

One thing Daniel, Shawn, and I've talked about is writing a simple  
API for actors that would allow their source code to be used without  
modification (although code might need to be auto-generated at build  
time) to yield either conventional actors or collection-oriented  
actors as needed (in most cases).

Cheers,

Tim

On Aug 3, 2007, at 12:04 PM, Norbert Podhorszki wrote:

> Hi Tim,
>
> None of the above. The metric is that someone like Bruce, trying to  
> build
> his first workflow, how easily/heavily figures it out, what to use  
> and how
> to use.
>
> COMAD is "heavy" for the actor/workflow developers who intend to  
> give such
> solutions, so it becomes "light" for Bruce. I would welcome that.
> I meant heavy here, that probably it would be a lot of work to
> write/rewrite actors to work in COMAD mode. However, I may be terribly
> mistaken with this statement.
>
> Norbert
>
> On Fri, 3 Aug 2007, Timothy McPhillips wrote:
>
>> Hi Norbert,
>>
>> Just out of curiosity, by what metric do you consider a collection- 
>> oriented
>> (or any other) solution "heavy"?  Number of actors?  Number of actor
>> interconnections?  Number of tokens passed?  Fraction of CPU time  
>> spent
>> managing data? Time spent composing the workflow?
>>
>> Cheers,
>>
>> Tim
>>
> _______________________________________________
> Kepler-dev mailing list
> Kepler-dev at ecoinformatics.org
> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/ 
> kepler-dev