perfsonar development work

[Martin Swany <>]

Hi all,

Comments below.

> > > > So my question is why does this schema not use the base schema  
> > > > (where I pulled my example from)?  Making changes in multiple  
> > > > locations like this is not desirable because of situations (such  
> > > > as this) where what is required works fine in one file, but not  
> > > > in another.  I will make the change to this file, but I would  
> > > > highly recommend this other schema be built using techniques to  
> > > > extend the current work.
> > > Jason, your point is fully clear and understandable, but where is  
> > > in XML_Schema_v03_final.rnc some reference to the base schema ?
> > > If I recall this right, building of stand-alone schema for E2Emon  
> > > was caused by the not-really-network-related structure of  
> > > required metadata and data.
> >
> > Most (if not all) of your schema is mirrored in the other schema  
> > files.  There are elements that are borrowed from nmbase, and the  
> > various topology and event schemas.  Just by looking I dont see  
> > anything that the e2emon schema requires that is not covered  
> > somewhere in the other files.
> > I do not recall why it was decided that a 'stand alone' schema is  
> > needed, but in my opinion it is not a good idea.  Extensibility is  
> > far more desirable than trying to isolate something like e2emon  
> > from changes to the schema files that may be required in other  
> > services.  I would argue that there have not been monumental  
> > changes to these files in a long time; keeping them separate  
> > doesn't really make much sense.
> Jason, as long as schema suits our requirements, it is not really  
> important for us whether its is stand-alone or integrated one (it  
> is really up to you and Martin)

I think it would be ideal if this schema removed duplicate elements and
pulled things in from the base schema.  That would allow fixes and
improvements to be automatically included (certainly we must insure
that changes don't affect backward compatibility, but we are already
faced with that.)  I think we should re-factor this to use the base like
the other perfSONAR ones do.

best regards,
martin