[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[moca] Re: RE: RE: RE : synchronous implementation of Pi calculus with choice?
>To me a business is a system which interacts with its environment through
>Named ports, as simple as that. Input port is a service that this
>organization needs and output ports are services that are advertised (in fact
>this can be advertised on a website). Other businesses can bind to this port
>and get services or offer services. If somehow a system can accomplish this
>then that will be the ERP system. My concern is do we need a language at
>all, Cant we use the PI calculus and describe a ERP System directly.
As I know, the concept of Named Ports was not proposed firstly in pi-calculus.
It may date back to CSP or CCS. Comparatively, the key characteristics of
pi-calculus is its mobility, i.e. to describe the system with a flexible
built-in structure.
To some extent, named ports, or generally speaking, channels are correspondent
with interfaces of services, which are open externally and strictly separated
from their implementations. However, any client who want to access such a service
should get a priori information on its interfaces. For example, anybody who want to join
this mailing list MOCA, should get to know its mail box or web site. The tendency for web application, I think, is to extend the accessibility of
services. Such open interfaces may not be known in advance. They can be
discovered; they can also advertise themselves directly or via some
proxies. As a result, services are bound dynamically. Pi-calculus gives a
graceful abstraction on such characteristics. However, it's not suitable to model systems directly by experience.
Regarding the language design & implementation of pi-calculus, Catuscia
Palamidessi's work may give us some inspiration on what language components should be primitive.
----- Original Message -----
From: "Alappatt,Antony" <Antony_Alappatt@xxxxxxxxxxxxx>
To: <moca@xxxxxxxxxxxxxxx>
Sent: Thursday, January 16, 2003 12:11 AM
Subject: [moca] RE: RE: RE : synchronous implementation of Pi calculus with choice?
> I fully agree with Bill, although i come from a totally different
> background. My background is implementation of ERP systems. My interest was
> the same, there has to be a better way to implement/describe Enterprise
> systems. This led me to Agent technology and then to PI Calculus.
>
> In pi calculus i found the right theoratical framework for both inter
> business and intra business coordination. In my opinion, the business
> implementations dilute some of the descriptive power of PI calculus
> (Languages like XLANG and Biztalk).
>
> To me a business is a system which interacts with its environment through
> Named ports, as simple as that. Input port is a service that this
> organization needs and output ports are services that are advertised (infact
> this can be advertised on a website). Other businesses can bind to this port
> and get services or offer services. If somehow a system can accomplish this
> then that will be the ERP system. My concern is do we need a language at
> all, Cant we use the PI calculus and describe a ERP System directly.
>
> Let me know if there are holes in my thinking. I hope this has gone off the
> subject.
>
> Regards
> Antony
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The "models for mobility" mailing list mailto:moca@xxxxxxxxxxxxxxx
http://www-sop.inria.fr/mimosa/personnel/Davide.Sangiorgi/moca.html