[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-kuml-devel
Subject:    Re: Status report
From:       Gerard Flynn <gerard.flynn () free ! fr>
Date:       2001-04-05 4:47:13
[Download RAW message or body]

Chris Moore wrote:

> Hi Gerad,
> 
>> I would have uploaded it already except
>> that before doing that I want to make sure that the code is basically
>> correct and that requires testing it on an actual MOF model.  Therefore
>> I'm currently working on the reverse MOF->IDL mapping.
> 
> 
> There's always the MOF model of UML (XMI format).  See 
> http://www.celigent.com/omg/umlrtf/artifacts.htm
> Shouldn't be too hard to read it in with an XML parser.
> We will need to write this code at some point anyway.
> 
> <snip lots of good stuff I agree with>

   Yes but I already had code for parsing IDL and haven't done anything 
yet with XML.  I'm hoping to derive the XML writer/reader for a MOF 
model from that MOF model somehow but that requires already having the 
MOF model implemented (circularity problem).  IDL importation seemed 
like the easiest way to break the loop.

> 
>>    Of course the easiest way to implement the specific interfaces will
>> be as a wrapper around the Reflective implementation.  However it will
>> always be possible to create an optimized implementation later.
> 
> 
> Interesting.  I was aiming to do the specific interfaces and use the
> implementation repository to do some of the reflection.  The different
> approaches are top down vs bottom up. 

   Yes, I think that's the way the OMG specification assumes the 
implementation will be done.  It seemed to me however that by directly 
implementing the Reflective interfaces most of the code could be written 
generically once and for all.

 I look with interest to see how things turn out in the end.  
   So do I.   

      Gerard

> 

[prev in list] [next in list] [prev in thread] [next in thread]