[Pgi-wg] PGI requirements list

[Morris Riedel]

Hi Oxana,

 minutes will be available soon covering many aspects of your points.

Take care,
Morris

>-- -----Ursprüngliche Nachricht-----
>-- Von: pgi-wg-bounces at ogf.org [mailto:pgi-wg-bounces at ogf.org] Im Auftrag von Oxana Smirnova
>-- Gesendet: Mittwoch, 27. Oktober 2010 23:51
>-- An: pgi-wg at ogf.org
>-- Betreff: Re: [Pgi-wg] PGI requirements list
>-- 
>-- Thanks, Johannes!
>-- 
>-- For those who could not attend, I think it's useful to know that the prioritization was made by counting
>-- appearances of requirements in 8 use case mappings; those that collected more than 4 counts got labelled green.
>-- 
>-- Counting was done in a break before the last session, so I personally had no opportunity to digest the result
>-- during the meeting. Maybe others were luckier :-)
>-- 
>-- I can't see why 50% is the in/out margin: it can be 66% or 75%, really. Or 33%. Depends what do we want to
>-- achieve: meet all reasonable requirements, or focus on those which are truly common. No requirement scored 8, by
>-- the way, so 5+ is not exactly 50%, actually.
>-- 
>-- I must also admit I'm still confused whether the requirements are made *by* middleware providers, or *for*
>-- middleware providers. My impression was that it is the latter - or at least most people tried to present use
>-- cases, not middlewares. Meanwhile the spreadsheet rather suggests the former, having providers' names in column
>-- titles.
>-- 
>-- The difference is obvious: e.g., when a use case needs, for example, bulk operations, it will need this feature
>-- to be implemented in *every* middleware, naively - because then the scientists will be able to run their tools
>-- across *all* middlewares, and our interoperability dream will be achieved. However, from the spreadsheet one may
>-- (mistakingly?) conclude that it is needed only in some middlewares, and thus didn't manage to get enough points.
>-- So, is our interoperability only a luxury for some users who are lucky to have applications that can be squeezed
>-- into our standards ?
>-- 
>-- And another note from the Brussels meeting: already before Munich we discussed a possibility of grouping
>-- requirements from narrowly scoped, like "I need a resume after hold", to more generic ones. This was done in a
>-- way too generic manner on Tuesday, when e.g. one big common requirement was "Glue2". Did anyone take a snapshot
>-- of that table drawn by Morris? I have to repeat: I simply had no time to digest the "green" requirements to
>-- assess what are the high-level commonalities between them, if any. Well, Glue2 was there, yes, but I didn't have
>-- to write 3 use cases and map them to 173 requirements to get *this* wisdom as the bottom line ;-)
>-- 
>-- I wish we had more time to discuss these things on Tuesday.
>-- 
>-- Cheers,
>-- Oxana
>-- 
>-- 
>-- 26.10.2010 18:52, Johannes Watzl пишет:
>-- > Dear all,
>-- >
>-- > we are just finished with the PGI sessions here at OGF30.
>-- > The requirements list including a prioritization based on the use cases
>-- > can be found on GridForge:
>-- > http://forge.gridforum.org/sf/go/doc16080?nav=1
>-- >
>-- > If you have any objections, please send a message to the list before the
>-- > next PGI call on Thursday, 4 November 2010.
>-- >
>-- > Best,
>-- > Johannes
>-- >
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3550 bytes
Desc: not available
Url : http://www.ogf.org/pipermail/pgi-wg/attachments/20101028/ab47744a/attachment.bin