Re: Model name question

Adding more complexity to the name scoping features in IBIS-X
may not help. The file level scoping in IBIS works today for
tools that work directly from IBIS files, building a simulation
model in memory and discarding it when the program exits. It is
easy for these to observe the scoping rules, and they may have
little trouble adapting to new scoping.

It is the tools that save the simulation models in proprietary
libraries that need special measures to preserve the
model-to-component linkage, because short "nicknames" for
models become pretty non-unique when they are all thrown into
one big library. An IBIS-X file with flexible scoping may
present even more difficulty for these tools, which usually
encode enough information into the stored model names to keep
the linkage straight later. More complex scoping could lead to
less readable model names.

It would be nice, of course, if no translation were needed.
IBIS-X may be flexible enough to eliminate the need for tools
to have their own proprietary formats, but I wouldn't count
on that happening for some time. The proprietary libraries
can take on a life of their own.

Mike LaBonte

Al Davis wrote:
>
> On Monday 16 July 2001 11:14 am, Scott McMorrow wrote:
> > Lynne,
> >
> > My suggestion is to formalize the concept that a unique model name
> > is described by the following:
> >
> > manufacturer.component.model.revision.
>
> That doesn't work. Model is not in Component scope. Model and
> Component are at the same level as far as scoping is concerned.
>
> You can have a file ...
>
> [Component]
> [Component]
> [Component]
> [Model]
> [Model]
> [Component]
>
> etc...
>
> All models are available to all components in that file.
>
> The only scoping we have is by file.
>
> Moving forward, it may make sense to introduce a [Namespace] to deal
> with these scoping issues, but I don't think this will solve the
> problem either.
>
> Should IBIS-X offer nested [Define]s ??
>
> al.
 
Received on Tue Jul 17 07:55:34 2001