[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

(FC-Devel) draft model-interchange policy



<disclaimer> 
[drw: If anybody would get the impression that I'm an expert 
on this subject, he or she is wrong. I'm not. 
I just happen to have some experience that leads me to
think that it is very important to state a policy on the
matter early on.] 
</disclaimer>

==========================================================
<h1>draft model-interchange policy</h1> 

FreeCASE will have to do a lot of interfacing, both
internal and external. We can define internal interfaces
from both sides and adapt them (to a certain level) to our
changing views as to what is needed, as we go along. 

Any external interface however is restricted by the other
party's support of format, semantics and protocol.

Volontary adherence to emerging standards is one of
the key elements of the barreer breaking power of 
the internet. It makes sense to say: Let's do that!
Let's make interface zyx RFCxyz-compliant. 

Well, we would. Convergence of formats and languages 
we will have to deal with is actually happening right
now. The direction of this convergence is clear, 
however it has not reached a stage where we can 
simply take the conclusion and build accordingly. 

Yet, we have to make a model-interchange choice for now.
This may sound like premature design to some. 
We think that it is a crucial step in narrowing down the otherwise
explosive problem we decided to deal with: distributed cooperative
development.
 
Summarizing an email-discussion with Uche Ogbuji and 
Junichi Suzuki [drw: Actually they discussed, I listened.
They are far better equipped to monitor current evolution
than I am.] gets us 

   3 predictions: 

1. XML will be important to us.

2. The outcome of the SMIF RFP of the OMG will very likely 
   set a standard that will be accepted by most if not all
   vendors. The outcome may be an XMI (which is in XML)
    -variant or a CDIF variant. Whichever comes out, it
   will be MOF-based. This suggests that it will be rich
   and complex. 
   
3. Directed-graph navigation widgets and underlying W3-RDF 
   (more generic than both XMI and UXF, but like them in 
   XML) will be important to FreeCASE. They will mature at
   a rapid pace.

This is our recomendation:

<policy>
long term:  
        Support the outcome of the OMG SMIF-RFP in 1.0.0
        How much work this will mean depends on the outcome.

mid term (as soon as RDF matures):   
        Have a UXF-RDF-converter in FreeCASE 0.1.x 
        The converter itself won't be very complicated.
        Adopting the widgetset may be a different matter.

short term: 
        Use UML eXchange Format (UXF) in FreeCASE 0.0.x, 
        the most readable format, in XML, available now. 
</policy>
==========================================================
Amendments before [reasonable deadline, let's say
within 10 days. Not ironclad so if the discussion
is still vivid at the deadline, we'll think of 
something] 22 july 1998 GMT 23:59
==========================================================

Info 

XML:
http://www.w3.org/XML and http://www.ucc.ie/xml

MOF:
http://www.omg.org/library/schedule/Technology_Adoptions.htm#tbl_MOF_Specification

SMIF-RFP:
http://www.omg.org/library/schedule/Stream-based_Model_Interchange.htm

RDF:
http://www.w3.org/RDF and http://www.mozilla.org/rdf/doc/index.html

UXF:
http://www.yy.cs.keio.ac.jp/~suzuki/project/uxf

further study:
DOM: http://www.w3.org/TR/WD-DOM