On 3/23/06, Steve Ross-Talbot <[EMAIL PROTECTED]> wrote:
> you clearly misunderstand.
Glad we got that sorted out. :)
> I expect you use Java or maybe C++ or somesuch.
A lot of different somesuch, yes.
> Do you have to understand Turing completness or lambda calculus to use
> them? Just because you drive a car doesn't mean you have to be a mechanic.
Well, if you're a 50-year old suburbian who's on first-name with your
local mechanic, that's true. But is this the right metaphor for the
people on this list? I assume we're architects and developers, the
equivalent of mechanics and engineers in the car-world.
I do have *some* understanding of both Turing completeness and
calculus, but I don't feel compelled to make my fellow developers
understand that to the same or higher degree than me to have a
meaningful discussion about SOA design. :)
But the point was more that we were urged to take a serious look at
this, and the only link provided was to the PI-calculus open-source
tool, so it's pretty natural for me to start off in that direction. A
tool that uses PI-calculus to do WS-CDL suggests to me I need to dig
into both.
> WS-CDL using the open source www.pi4soa.org tool suite which is itself
> using pi-calculus does not mean you have to understand it. We have a lot of
> people using it who have not got a clue about pi or lambda or any other sort
> of
> formalism. The plain fact is they don't need to know about that stuff. That is
> the whole point of WS-CDL and the tool suite. It uses the stuff but hides it
> from
> you and so delivers value.
Hmm, as I've seen, the WS-CDL is somewhat married to formal calculus
(haven't checked the nitty-gritty details of PI-calculus, though) with
channels and guards and whatsnot. (The sentance "The Pi Calculus [...]
forms the basis for the W3C's WS-CDL" comes to mind here) WS-CDL is a
model (and abstraction) over more formal calculus solving of given
problems which no matter how much you fiddle on the syntax/model
layer, the underlying stuff always shines through. Which may or may
not be a good thing. Seriously, this isn't criticism of PI-calculus
per se, but as of now given it is the only open-source tool available
to do WS-CDL you would think they are connected. Maybe this is where
you jump in and explain exactly how the two fit together.
Most of the time I reckon it's a good thing to have he underlying
concepts seep through to the model layer; you can teach a lot of old
dogs new tricks that way. Heck, if I was a tad smarter, I'd really
want to jump in there and try to understand PI-calculus; I reckognise
the problem and that this is a possible solution, no doubt, but as a
serious developer I *actually* want to understand how things work.
Sorry. :)
The more I dig into WS-CDL, the more *I* have to understand (wheras
other people might not) certain basics from calculus, a field I've not
looked at for many years. I'm not saying this is a reasonable excuse,
but I assume it is one for most developers out there. It is a fairly
complex standard, unfinished as such. (And perhaps it's a shame you
haven't got the examples section ready; maybe that would take the
sting out of it) I'm a semantic modeller, and the first thing I do is
to try to understand that model, where it comes from and why it's
shaped the way it is.
But hey, we can talk pure WS-CDL separated from it all as well; I'm flexible.
> Rather than issue such a criticism I suggest that you try the tools
> out and then if you have an issue you are at least basing it on fact
> and not fiction, hearsay or the usual FUD (fear uncertainty and doubt).
Hmm, what fact is it that I've fudged? Be careful about telling people
that they're lying; perception is a multi-facetted thing.
> I look forward to a more factual response next time.
I can sense that you're in defence-mode, so I'll just end with a few
personal factual observations that should be taken
tounge-somewhat-in-cheek ;
- complexity : the WS-CDL specification tries to do it all through a
model and language which is new to me, meaning steep-learning curve
- model : as above; why that model? What's the justification?
- syntax : again, unfamiliar, possibly marred to the model. I come
from a Topic Maps world where there is a general model for all types
of relationships, roles, types and ontology stuff on top. Why not use
that?
- Cute name; where's the theme music? :) (yes, that's a joke)
Anyways, there you go; it's not bashing of specifics at all, just
showing you *confusion*.
Alex
--
"Ultimately, all things are known because you want to believe you know."
- Frank Herbert
__ http://shelter.nu/ __________________________________________________
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/service-orientated-architecture/
<*> To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
