Howdee Prof. Lee and hackers,
Aha! I will definitely revisit that one again tomorrow. It sounds familiar.
I'm not
sure from your description if this is restricted to DDF - I'lll see...still
getting up to
speed on what is general and what is specific.
> It is certainly possible to do this.
> It is done in the DDF domain (see the towers of hanoi
> example, which uses a recursive construct to dynamically
> build a dataflow model as needed). There is also the
> MobileModel actor, which accepts a MoML string at the
> input and instantiates it as a model to process
> input data. This latter actor should be viewed as
> a concept demonstration.
MobileModel seems like a pretty cool idea; as it is experimental I might
give it
a try to see how it works, but we'll probably stick to the well-established
stuff
(esp. as we are learning). BTW - I'm about 1/3 through the "Operational Sem-
antics" paper you referred. It's been a *while* ('94) since I last looked at
"DSP-
ish math", and some of the number theory stuff I've not formally studied, so
it is a
tad slow going, but very interesting reading nonetheless.
> There is the whole potential pitfall of self-modifying
> code to worry about here... I think there is a very good
> research topic in identifying the best higher-order actors
> (actors that operate on actors). The recursive technique
> used in DDF is an example of a highly structured (read,
> displined) way to do this. But it also may be more restrictive
> than we want.
This is certainly an interesting subject - meta actors; in fact my mind goes
to the
whole OMG MDA concept - at least if we can consider model transformations
to be carried out by some sorts of such meta actors. Have you considered the
MoC concepts within this area? OMG's official stance on MDA seems (to me)
to be a bit vague, discussing "marks" and "transformations", but it seems an
intrig-
uing notion to consider MoC as part of going from "PIM" to "PSM" - the
target
semantics are certainly part of the specific platform. "Managed semantics".
There's
no shortage of academic research in MDA; neat stuff - including some
attempts to
formalize or otherwise identify taxonomies of model transformations. Yet, I
haven't
seen MoC concepts such as made explicitly in Ptolemy, as far as I know.
Whether
UML itself will be around in 10 years or not (i.e. largely as it is today) I
am not sure,
but I can see some of the MDA concepts being interesting in the long run.
Goodnight, Ptolemy hackers!
Chuck
----------------------------------------------------------------------------
Posted to the ptolemy-hackers mailing list. Please send administrative
mail for this list to: [EMAIL PROTECTED]
