Re: [mmox] Open Grid Protocol -- Maybe need a redifinition for MMOX scope ?

[Kajikawa Jeremy <[e-mail]>]

If the Open Grid Protocol needs redefining ...

Here is a proposal for consideration...

In which case the OGP itself becomes a defined framework with modules
  and each module has a preset function...

<OGP> being the root access of the Open Grid Protocol

<OGP>/Session -- Module for maintaining "Client" connection to the VW,
  *all* additional modules use "handle" for the session module in-use.

 The OGP library only define a standard access library for code modules,
 where individual clients may provide an "own session" module for
protocol

There is two "global" methods used for accessing modules of the OGP
library

  OpenModule()  Load and call an Init Function for the Module named
  CloseModule()  call an Exit() function also allowing the module to be
free'd

 The "Session" module is defined with the following functionality,

  Login - Creating an Active Connection and Identifying to a User
Account
  Logout - Ending an Active Connection
 
  Send() - Extension Modules call this to send anything,
  Recv() - Extension Modules call this to read anything within a set
time,

  ErrorString() - *Optional* provision of error code strings for user
messages

<OGP>::OpenModule(name) returns "handle" for module reference
<OGP>::CloseModule(handle)

<OGP>/Session::Login(server,user,pass) returns token
<OGP>/Session::Logout(token) -> return always success
<OGP>/Session::Send()
<OGP>/Session::Recv()
<OGP>/Session::ErrorString()

The above lacks any reference or allowance for incoming messages to
trigger events,
  and I am assuming everyone here is capable of setting up a mechanism
for read()
  calls to trigger as required or occur at extremely frequent timing
intervals.

Im allowing for the protocol itself to have multiple versions installed
and current code
  to be used side-by-side with new code... if this is feasible...

or am I missing something about the way python works different from C ?

Jeremy