[Mpi3-hybridpm] External interfaces chapter updates

Bronis R. de Supinski bronis at llnl.gov
Sat Oct 30 16:29:08 CDT 2010


Doesn't it seem like it would be useful to partition
the changes into the issues that motivate them? It
would seem we could use the ticket number as a way
to do that.

Otherwise, it will be hard to tell minor wording changes
that can be considered individually from those that need
acceptance of a more significant overall change.


On Sat, 30 Oct 2010, William Gropp wrote:

> We don't need tickets for edits to the document - tickets were the
> appropriate way to manage minor corrections to the document but don't
> help the process of developing a new or greatly revised chapter.  The
> RMA group has two proposals on its wiki, both of which are in the
> context of complete chapters.  A drawback to tickets was they made it
> possible to think that by reading the ticket, the issue was
> understood.  This caused several problems in the prior rounds, and
> would be more serious here.
> So, no ticket is needed.  Following the written instructions, a zero
> (0) may be used.
> Bill
> On Oct 30, 2010, at 3:32 PM, Bronis R. de Supinski wrote:
>> Pavan:
>> Re:
>>>> 2. You need to annotate your proposed changes with ticket
>>>> numbers. I suggest that we need separate tickets for each
>>>> of the major additions. Isee that ticket 217:
>>>> https://BLOCKEDsvn.mpi-forum.org/trac/mpi-forum-web/ticket/217
>>>> covers the helper threads proposal. We will need ticekts
>>>> for the shared memory extensions and endpoints.
>>> I thought the process was changing out of using tickets for the
>>> changes
>>> for MPI-3. If not we can update this once all the tickets have been
>>> created.
>> The change tracking macros include a ticket number field.
>> My understanding is that for simple wording changes for
>> things like grammatical issues on which the chapter
>> reaches consensus then we can use "ticket 0". For more
>> significant changes, like major additions, my understanding
>> is that they are still supposed to use a separate ticket.
>> However, the revised text should be in LaTex so that
>> we vote on it as part of the ticket.
>> Bill and/or Rich should confirm or correct.
>> Bronis
> William Gropp
> Deputy Director for Research
> Institute for Advanced Computing Applications and Technologies
> Paul and Cynthia Saylor Professor of Computer Science
> University of Illinois Urbana-Champaign

More information about the mpiwg-hybridpm mailing list