All, This email announces the LAST CALL for the HPC File Staging Profile (HPC-FSP). The latest version of this document can be found at http://forge.ogf.org/sf/go/doc15024. Last call will last for one week (ending on 2/11/08). Glenn
Glenn, I've just read thru this version. I'm in general agreement with it, but there are number of minor things which should be cleaned up prior to public comment. /Blair Notes: Introduction - In 1st sentence replace [ref] with [HPCP10]. - 2nd paragraph "...used as the basis of interoperability testing between systems....". I assume the point of this profile is not just testing. Should it read "...used as the basis for interoperability between client and service implementations claiming compliance"? Section 3.1 - I'm confused by 3.1.1, 3.1.5, 3.1.6. If the JSDL FileName, Source, and Target elements are required to conform to the JSDL v1.0 spec referenced, then how is this defining a JSDL data staging element modification as stated in the first sentence under 3.1? I suspect this is just a wording issue in that sentence. Section 3.2 - You define the Credential element but don't say what compliant implementations need to do. I believe you should expand on the text to state: 1) Compliant implementations must be able to recognize and parse the 'Credential' element 2) Compliant implementations must be able to recognize and parse the Username Token element encoding and X.509 Certificate Token encoding defined in [WS-Security] and profiled in [WS-IBSP] when appearing as Credential-element content. (You should add refs for WS-Security and WS-I BSP.) 3) Compliant implementations MUST support either the Username or X.509 Token type as a file staging authentication credential, but MAY support both. 4) If an implementation does not recognize and/or support the type of credential conveyed in the Credential element, they MUST return an <UnsupportedFeatureFault> fault. Section 3.3 - References to the standards/specs defining FTP, HTTP, and SCP should be added so there is no question as to which protocol version was intended. - I'd remove the last sentence in this section and cover it in 3.2 per my comments above. Section 3.5 - It says "The later in particular may have different answered for the stage-in....". Couple of typos here. How about "The latter, in particular, may have different answers for the stage-in...." Also, need to replace [insert names] in Contributors and Acknowledgements plus provide complete information for all normative references.
-----Original Message----- From: ogsa-hpcp-wg-bounces@ogf.org [mailto:ogsa-hpcp-wg- bounces@ogf.org] On Behalf Of Glenn Wasson Sent: Monday, February 04, 2008 2:25 PM To: ogsa-hpcp-wg@ogf.org Subject: [ogsa-hpcp-wg] LAST CALL for HPC File Staging Profile
All,
This email announces the LAST CALL for the HPC File Staging Profile (HPC-FSP). The latest version of this document can be found at http://forge.ogf.org/sf/go/doc15024. Last call will last for one week (ending on 2/11/08).
Glenn
-- ogsa-hpcp-wg mailing list ogsa-hpcp-wg@ogf.org http://www.ogf.org/mailman/listinfo/ogsa-hpcp-wg
Glenn Wasson wrote:
This email announces the LAST CALL for the HPC File Staging Profile (HPC-FSP). The latest version of this document can be found at http://forge.ogf.org/sf/go/doc15024. Last call will last for one week (ending on 2/11/08).
I think you're not quite ready for Last Call yet. You have a number of issues still to fix, most of which are minor. In section 3.1.2, the wording prohibits a future profile from extending this profile to support <FileSystemName>. It also means that any implementation which *does* support <FileSystemName> cannot be compliant with this profile. Is this intended? You're using 'must' (in lower case) in a fair number of places. You are aware that that is non-normative? Now for the minor editorial stuff: In the Copyright Notice on the front page, the copyright year need to be extended. :-) Sections 5, 6 and 7 are incomplete. In section 8, the full copyright notice is wrong. (Inconsistent too.) Similarly for section 9. In section 10, the URL for JSDL 1.0 is missing (and the reference should not be just to a URL anyway; it's a published document, a comment which also applies to the BES 1.0 reference). The reference to HPC Profile 1.0 is incomplete. In Table 2-1, it would be a good idea to shade the title row differently to the other rows. You're inconsistent about whether references to References are bookmarks or not. In section 3.4, it would be nice if the paragraphs describing particular protocols were indented to make them more visually distinct from the surrounding text. It seems there were more nits to be picked than I expected. :-) Donal.
participants (3)
-
Blair Dillaway -
Donal K. Fellows -
Glenn Wasson