Re: Note on Rev

On Friday 06 July 2001 01:35 pm, Dagostino, Tom wrote:
> What is the backward compatibility problem. If this applies to Ver
> 3.3 or later files what is the issue?

Here are a few reasons:

1. Some of the old tools will be used to generate new style models,
by first generating the old style model, then adding the one new
feature. They would then need to bring the whole file into
compliance.

2. Some simulators will be using the new technology (macro language
based) software to process even the oldest files. Probably all tools
newly adding IBIS support will be in this category. Some of the old
tools will switch to the new method even for the old files because of
lower maintenance costs. For this reason, the new software must
accomodate the old files. This is a basic principle in the design of
the macro language. This one is easily solved by a "compatibility
mode" in the software, that accepts the implied block ends.

3. It is my opinion that the new work should concentrate on what does
not work well in the old standard. I have not heard many complaints
about the header section. It is also my opinion that we should not
make gratuitous changes.

Having said that, some of the "decisions" were based on what we
thought public opinion would be. Now that we have some comments from
outside the committee, it really is arbitrary and could go either way.
 
Received on Fri Jul 6 14:15:28 2001