Also, it isn't a matter of whether or not the "should" be opaque. They are. The spec. says they are and that is then end of the story. The only fields which are allowed to be "interpreted" by the client or anyone but the "menting" service are the Address field (which is a URI), and the Metadata section. The first of these is used to target future web service requests, the latter is optional "hints" that clients may choose to use or ignore. Further, you can write a spec. profiling what the meaning of various fields in the EPR are, but you are not allowed to constrain what service mentors choose to put into the other fields. For those of you who program in C or C++, telling someone how they are allowed to create their EPRs is like telling them that only prime numbers are allowed as pointers -- it doesn't make any sense. -Mark On Mar 11, 2010, at 9:06 AM, Andrew Grimshaw wrote:
All, I cannot make the call, I have another meeting. However, if one is using web services, then EPRs are the way to go, unambiguously. It is important to keep in mind that they are not intended for human consumption! They are XML data structures that can potentially carry quite a bit of information beyond the address:to field useful both to clients (metadata) and service implementations about context (reference parameters).
That information can include a unique ID as was being discussed in the email thread, security information such as public keys, or policy information, or even information on how to rebind the service if the service appears dead.
A
-----Original Message----- From: pgi-wg-bounces@ogf.org [mailto:pgi-wg-bounces@ogf.org] On Behalf Of Morris Riedel Sent: Thursday, March 11, 2010 7:09 AM To: Pgi-wg@ogf.org Subject: Re: [Pgi-wg] [gin] Reminder & Agenda: PGI Telcon Today 2010-03-11; 3:30 pm CET
Hi team,
gLite, GENESIS and UNICORE agree on using EPRs and for today we agreed that members of ARC (i.e. Balazs, Aleksander) fill in some drawbacks of EPRs in the WIKI.
Currently I don't see any point against EPRs in the wiki: http://forge.gridforum.org/sf/wiki/do/viewPage/projects.pgi-wg/wiki/Executio nServicePoints
Please do so before the Telcon - otherwise we can say YES to EPRs because the majority votes for it, has experience with it, and we have no counter statements.
We need to make progress - so an agreement on EPRs here is a good step forward and would be a good outcome of the call today.
Take care, Morris
-- -----Ursprüngliche Nachricht----- -- Von: gin-bounces@ogf.org [mailto:gin-bounces@ogf.org] Im Auftrag von Morris -- Riedel -- Gesendet: Donnerstag, 11. März 2010 13:03 -- An: Pgi-wg@ogf.org -- Cc: gin@ogf.org -- Betreff: [gin] Reminder & Agenda: PGI Telcon Today 2010-03-11; 3:30 pm CET -- -- Hi folks, -- -- today will be a PGI telcon as planned. -- -- (1) -- Numbers: -- -- we are using: -- via Skype call +9900827049931906 (free of charge), ordinary phone numbers -- (local rates) with the 9931906 conference number: -- -- Austria 0820 401 15470 -- Belgium 0703 57 134 -- France 0826 109 071 -- Germany +49 (0) 180 500 9527 -- Switzerland 0848 560 397 -- -- -- -- (2) -- Screen-Sharing: -- -- http://webconf.vc.dfn.de/ogf-pgi-wg/ -- -- No login is required. Just enter as 'guest' with your real world name. -- -- -- -- (3) -- Proposed agenda: -- -- (a) -- Review Minutes last week -- -- (b) -- Review action items -- -- (c) -- Continue EPR discussions and agree on addressing approach -- -- (d) -- AOB -- -- -- High participation is appreciated -- -- -- Your co-chair, -- Morris -- -- ------------------------------------------------------------ -- Morris Riedel -- Distributed Systems and Grid Computing Division -- -- Jülich Supercomputing Centre (JSC) -- Info: http://www.fz-juelich.de/jsc/JSCPeople/riedel -- --
_______________________________________________ Pgi-wg mailing list Pgi-wg@ogf.org http://www.ogf.org/mailman/listinfo/pgi-wg