Subject: Re: FW: arguments on removal of "static"
From: Dave Rich (David.Rich@synopsys.com)
Date: Thu Dec 12 2002 - 09:35:53 PST
Steven,
If the SV-BC charter is extended to fix problems in the 3.0 spec, you
still have not identified what problem having "static" declarations
creates.
The decision of the IEEE1364-2001 committee is irrelevant as
long as we remain backwards compatible. The IEEE 1364-1995 committee
rejected "generate" in favor of the much simpler arrays of instances. As
soon as the 2001 committee added multi-dimensional arrays and arrays of
wires, it became easier to pass "generate" through the committee.
I'm always against adding new keywords to the language if there are
reasonable alternatives, which may exist in this case. However the SV-EC
has a good reason for using the "static" keyword in classes.
SystemVerilog 3.0 has all the semantics of ANSI C except pointers, extern, and
the goto statement. There is tremendous value in reusing as much as
possible from the ANSI C standard. It provides a firm reference point
for both the implementors and the end users, even if it means inheriting some of
the not-so-nice features. Those pitfalls are well known to the community at large.
Take the issue with bump operators, ++/--. As you have mentioned, there are some
problem areas that can create ambiguous behavior. We may want to limit its
usage, but removing it from the language is not necessary.
So what are the problems you know of with the implementation of "static" as it
is current defined?
Dave
Steven Sharp wrote:
>I don't believe that Peter has raised any issues that were not already
>discussed by the group before making its decision.
>
>There needs to be a process for fixing significant problems in the 3.0
>specification. This group has agreed that this change should be made.
>Many of this group are representatives from companies that are voting
>members of Accellera. Decisions within Accellera are supposed to be
>made by representatives of the member companies.
>
>I would suggest that everyone working for a member company speak to
>their board representative about a solution for this problem. I can
>see two possible solutions. The first would be to get the Basic
>Committee charter extended to allow us to fix any problems in the
>3.0 specification. If the board is not willing to delegate that power
>to us, a second solution would be to create another committee to make
>the final decision on our recommendations. Such a committee should have
>fair and equal representation from the member companies.
>
>Steven Sharp
>sharp@cadence.com
>
>
>
>
-- -- Dave Rich Principal Engineer, CAE, VTG Tel: 650-584-4026 Cell: 510-589-2625 DaveR@Synopsys.com
This archive was generated by hypermail 2b28 : Thu Dec 12 2002 - 09:37:55 PST