(j3.2006) (SC22WG5.4552) Comments on the technical content of the coarray TS

[John Reid]

Bill,
>
> Is N1835 intended to be included in this list by reference, or do the
> items there need to be repeated? I note that there are references to
> N1835 in the new document, but I missed seeing a clear statement that it
> is included entirely.

This was one of the resolutions at the June meeting:

G10. Content and processing of the TS on further coarray features
WG5 directs its convenor to invite the public to submit, before
23 September 2011, requests for items to be included in the TS and
directs PL22.3 to construct a draft requirements document by the end
of its meeting in February 2012.

The aim in WG5 is always to reach consensus so I was hoping that J3 
would take account of any past WG5 paper that is relevant to this aim.
I see no point, for example, in your repeating in N1883 the case you 
made for revised collectives in N1835. We need also to discuss the old 
features, set out in N1858.

Cheers,

John.


>
> Cheers,
> Bill
>
>
>
> On 9/30/11 11:04 AM, John Reid wrote:
>>
>>
>> Tobias Burnus wrote:
>>> On 09/30/2011 02:51 PM, Bill Long wrote:
>>>> I think an important question that needs to be decided is whether the
>>>> goal of the TS is to include only those features that are already in
>>>> demand (from users) or that users through (having failed to carefully
>>>> read f2008) were already standardized, leaving more general expansion
>>>> to the next revision, or whether the TS should be a comprehensive
>>>> extension of coarrays that would unlikely be modified in the next
>>>> standard. I would argue that the current TS draft is somewhere between
>>>> these tow options. Going with the minimal TS would allow for quicker
>>>> completion and sooner start on f201x. The larger version would
>>>> possibly result in better long-term integration of the feature. At
>>>> this point, my personal preference is for the minimal approach.
>>>
>>> I cannot say anything about the meeting agreements, but as a Fortran
>>> user and Fortran-compiler developer, I am also in favour for a smaller
>>> TS. Given that the number of compilers well supporting coarrays (as
>>> defined in Fortran 2008) is very low, the practical experience with the
>>> newer features is still rather low. Thus, adding the most important
>>> missing features as TS allows to draw on user experience for additional
>>> features during the F201{3,8} development.
>>>
>>> For me, the collective/broadcast feature is the most important omission;
>>> having some subdivision (teams) or collective I/O seems to be of lesser
>>> importance (for my projects at least).
>>
>> I have taken the liberty of editing your words very slightly and adding
>> them at the end N1883. Is this OK?
>>
>> Best wishes,
>>
>> John.
>>
>>
>