[sv-bc] DataTypes: 11/11/04 Meeting Minutes II

From: Kathy McKinley (mckinley@cadence.com)
Date: Mon Nov 15 2004 - 09:42:01 PST

  • Next message: Kathy McKinley: "[sv-bc] DataTypes: Meeting on Thursday"

                        11/11/04 Meeting Minutes (12:00 pm PST)

    Attendees:

        Mark Hartoog
        Kathy McKinley
        Brad Pierce
        Steven Sharp
        Stuart Sutherland

    Summary:

    The picture is helpful, and will be included in the informative
    introduction to the proposal. However, the "constant" box in
    the set of "Data Objects" will be removed since it is a category
    rather than a specific kind of object. This picture represents
    different characteristics of a data object, it does not imply
    any syntax rules. So, for example, an attribute has a data type,
    even though you do not declare it with one.

    It is unfortunate that "variable" is the only kind of data object
    that does not have special syntax associated with it. Most have
    keywords (e.g. "parameter"); an attribute has the (* ... *) syntax.
    The keyword "var" is reserved by SystemVerilog, and the idea of
    formally adding it to the syntax has received support in the past.
    For example:

        var logic my_variable; // another way to say "logic my_variable;"

    In addition to making the data object/data type relationship clearer
    in the documentation, adding "var" would allow a user to make the desire
    for a variable explicit in situations where the rules may be confusing
    (like in port declarations). We unanimously approved the motion to pursue
    extending the syntax to allow "var", and for putting it on the agenda
    for the next Thursday meeting.

    We approved Steven's latest version of the changes for section 18.

    Brad will send Kathy new wording for section 7.16 (that incorporates
    the change from erratum 254), to be put in our proposal.

    We decided that the proposed changes in Annexes E, F, and I should
    be handled by the CC group, not the BC group. Kathy will forward them
    to the chair. Both Mark Hartoog and Steve Dovich have looked at the VPI
    extensions needed for new data types on nets, and feel that the changes
    required are small (basically a modification to the object diagram
    for nets). The VPI changes also belong in the CC group.

    We approved new glossary entries for "data object" and "data type".
    We approved minor changes to the glossary entries of "Aggregate",
    "Bit-stream", "Dynamic", "Enumerated type", "Integral", and "Singular".

    Kathy will put all of the proposed changes and the introductory matter
    into a text file and mail it out for review right after the meeting.
    Once the people at this second meeting agree that the changes have
    all been made correctly and are complete, she will pull it into MSWord
    along with the modified picture. Stu will turn it into PDF, and the
    proposal will be sent to the BC/EC/CC/AC chairs so that they can
    distribute it to the members of their respective groups.

    Our next meeting is Thursday, Nov. 18 at 8:30 am PST.

    This archive was generated by hypermail 2.1.4 : Mon Nov 15 2004 - 09:29:52 PST and
    sponsored by Boyd Technology, Inc.