Minutes of the SystemVerilog errata meeting

August 6, 2004

9:30am-12:00pm PDT

Mountain View, California

Agenda: http://www.eda.org/sv-bc/hm/1869.html

Participants

--------------

Dennis Brophy

Surrendra Dudani

Vassilios Gerousis

Matt Maidment

Francoise Martinolle

Swapnajit Mittra

Mehdi Mohtashemi

Karen Pieper

Brad Pierce

Dave Rich

Rishiyur Nikhil

Arturo Salz

Bassam Tabbara

Tesh Tesfaye

Yong Xiao

Eugene Zhang

Action Items

1) Dennis to talk to David Smith about the e-mail capabilities of the database system.

2) Dave Rich to factor the non-enhancement portions out of erratum 23 ( http://www.eda.org/svdb/bug_view_page.php?bug_id=000023 ) into new errata

3) Karen to disposition errata 105-157 as 'immediate'/'low' (=must-fix/delayable) and update database

4) Create 'developer' accounts for the volunteers listed below

5) Ask Ralph Duncan if he wishes to volunteer as a developer

Additional developers

Vassilios requested additional volunteers to act as 'developers' of proposals for the errata. Volunteers were

1) Francoise Martinolle

2) Bassam Tabbara

3) Matt Maidment

and Swapnajit suggested Ralph Duncan.

The JEITA SystemVerilog task group submitted 32 errata. Karen entered them into the new database ( http://www.eda.org/svdb ). They were assigned id numbers 126-157.

The agenda is to examine and summarize the errata in the database so far, and to discuss the process and operating procedures that will allow us to preserve and enhance this information.

We operate under the IEEE rules and procedures. We're performing a task on behalf of and at the request of its P1800 committee.

There are approximately 150 errata recorded in the database, ranging from trivial typos to significant enhancement requests. A few of the errata have already been resolved according to the votes of the existing SystemVerilog meetings, and have been marked as such. These need to be incorporated into the next draft.

A few of the nonresolved errata have formal proposals and somewhat more have informal proposals. Swapnajit moves the following –

"Submitters are encouraged to provide a formal or informal proposal to fix an issue. The relevant committee or committees will adjudicate."

Karen seconds. No opposed. No abstains. The motion passes.

Vassilios notes that we will not consider enchancements as part of this activity. Eugene asks about the process for enhancements. Dennis says there is no room in the P1800 work for enhancements, e.g., AMS. Eugene says, "so nothing brand new can be considered". Swapnajit asks if missing items are enhancements. Dennis says, yes, unless they were already expected.

Swapnajit suggests that people continue to open errata, even if they are enhancements that cannot be addressed in the current activity, so that the information is captured and preserved. There is concensus in support of Swapnajit's suggestion.

Questions are asked about the structure of the P1800, but after various speculations Dennis reminds that this should be left to the August 9 meeting, because it's a P1800 concern. He also speculates that it will adopt a "consent agenda" mechanism that gives any entity the ability to force discussion of particular issues, but does not require the working group to step through each and every issue that is resolved by its technical committees.

Swapnajit says that the new database is not closely coupled with the e-mail archive system and that it would be a good idea to send e-mail to the reflector whenever a bug is filed. Dennis says earlier feedback had suggested that a digest mechanism would be preferred by the community. Dave would like a report on the capabilities of the database system, and Dennis agrees to talk to David Smith about that.

Matt says that he wouldn't mind getting all of that mail because he can do filtering and is concerned that the database does not seem to have a powerful enough query mechanism, especially for change tracking. Brad suggests a new archived reflector that would be sent messages about each change, but that would need to be separately subscribed to.

Vassilios points out that only two of the four levels of accounts are currently in use, namely, 'reporter' and 'developer'.

Francoise asks if there's a way for people without one of these kinds of account to report a bug through the system. Answer: Only indirectly via e-mail to the reflector. Karen points out that eda.org insists on this policy because they had problems in the past with illegitimate file distribution exploits.

Francoise asks if this database is permanent. Vassilios says, yes, it is the database that will be used by P1800. Dennis points out that it complies with the IEEE openness rules. Francoise raises the issue of the 1364 database hosted and maintained by Stefen Boyd -- don't they need to be merged? Swapnajit asks if there's a possibility that all of the issues from 1364 will need to be refiled. Vassilios says that the IEEE SystemVerilog standard needs to be completed before considering potential merge issues.

After a short break, the first 100 errata (with id numbers 4-104, because there are no 1-3 or 28) are dispositioned by looking at each erratum individually and marking it 'immediate' or 'low', according to whether it must be fixed in the current standardization effort or is delayable to some potential future version of the standard.

The scheduled meeting time is nearing its end, so Vassilios requests that Karen disposition the remainder of the errata according to the same reasoning that has already been used on the first 100. There is consensus in support of Vassilios' request, and Karen agrees to take on the action item.

At 12pm, Vassilios moves to adjourn.