[Smwg] Next draft XFDU for TGFT Tech Note

[John Pietras]

Erik,
Based on a long conversation Colin and I had the week before last, I've made some significant changes to the Tech Note, a major one being that the XFDU Package can now contain multiple payload data files. The other major changes are:

a)      the content model is further simplified to eliminate the possibility of having metadata file content and payload data file content embedded within the XFDU Manifest itself. All payload data files are distinct files contained within the zipped/tarred XFDU Package file, and all metadata files are either contained within the zipped/tarred XFDU Package file or external to the XFDU Package altogether.

b)      The OAIS-specific metadata types and their associated classifications and categories have been removed from the "standard" TGFT XFDU. That is, in general, services will simply use the anyMDID attribute to cross-reference to metadata elements. (Services could still use the OAIS-specific metadata types if they are applicable, but such usage would have to be introduced and defined by the service specifications themselves).

I've briefly skimmed through your comments, and I see that at least several of them deal with the one vs. multiple payload file issue. I think some of the others have been rendered OBE  by the most recent version, which I have published to the CWE this morning at URL
https://cwe.ccsds.org/css/docs/CSS-SM/CWE%20Private/Terrestrial%20Generic%20File%20Transfer%20Book/XFDU_for_TGFT_TechNote-v0.4-170424.docx

Unfortunately, at this point I've burned through the time that I have allotted to this topic before the Spring meeting, so I won't be able to issue an update that specifically addresses your comments. But I'll read through your comments before the spring meeting and we can discuss them then.

Best regards,
John

From: Barkley, Erik J (3970) [mailto:erik.j.barkley at jpl.nasa.gov]
Sent: Friday, April 21, 2017 8:03 PM
To: John Pietras; CCSDS SMWG ML (smwg at mailman.ccsds.org)
Subject: RE: Next draft XFDU for TGFT Tech Note

John,

I finally had a chance to look at this.     Attached, are some comments.  In general, I think this is a very good write-up re XFDU and considerations re TGFT.

Re the write-up itself, I am not sure that limiting ourselves to one data object is really what we want, although I agree that is a good starting point.

More in general, in reading through this I had sort of an odd conclusion in that the text descriptions all generally made sense and I think the analysis you have offered is good. But when I put the picture together and take a look at the sample I'm not necessarily convinced that all of the "surrounding" information is that useful.  Of course, a lot of this is because this is what the XFDU offers and the XFDU, as you have noted, is a solution to a much larger problem than that being tackled by TGFT.

CSSM Colleagues,
I think it will be useful in San Antonio for agencies to discuss to use cases (and requirements) to help us decide if and how much XFDU we incorporate. Please see the attached for a pictorial representation of the instance sample of the XFDU (in .emf and .png formats) from John's write-up (I tend to find the picture helps in identifying the types of information in the proposed XFDU for TGFT). The main question I have is does this really help anybody in terms of doing file operations - I'm perfectly okay with the answer being "yes" but I'd like to hear from the membership what "yes" really means for the particular agencies.

Best regards,
-Erik


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.ccsds.org/pipermail/smwg/attachments/20170424/fe30e6b1/attachment.html>