Scribe
- Andrew Berezovskyi (KTH)
Attendees
- David Honey (IBM)
- Harish Krishnaswamy (Software AG)
- Jim Amsden (IBM)
- Martin Sarabura (PTC)
- Nick Crossley (IBM)
- Ian Green (IBM)
- Andrew Berezovskyi (KTH)
- Axel Reichwein (Ind)
- Jad El-Khoury (KTH)
Regrets
Resolutions
Actions
- Martin to fix some references and introductory paragraphs for TRS and CCM - still on list
- All committee members encouraged to read query spec
- Axel to commence editing on TRS spec
- Nick to set up subcommittee meeting to discuss skew resolution
Chat transcript from room: oslc
[07:00] Martin Sarabura (PTC): Previous minutes:
https://wiki.oasis-open.org/oslc-core/Meetings/Telecon2017.10.26
[07:03] Andrew Berezovskyi (KTH): Andrew is scribing today
[07:03] Andrew Berezovskyi (KTH): Minutes approved.
[07:04] Andrew Berezovskyi (KTH): All members are encouraged to read and review the Query specification draft
[07:05] Andrew Berezovskyi (KTH): General feedback is welcome on the mailing list, specific questions in the issues.
[07:06] Andrew Berezovskyi (KTH): David has an action item to set up a subcommittee meeting
[07:06] David Honey (Persistent/IBM): Core issues at
https://issues.oasis-open.org/browse/OSLCCORE-128?jql=project%20%3D%20OSLCCORE
[07:06] Andrew Berezovskyi (KTH): Nick has to set up a subcommittee meeting as well, but next week and the week after are not suitable due to the IBM conference and the Thanksgiving
[07:07] Andrew Berezovskyi (KTH): Subcommittee meetings will be related to TRS and CCM
[07:08] Andrew Berezovskyi (KTH): Jim suggests to offload the TRS spec editorial responsibility.
[07:09] Andrew Berezovskyi (KTH): It is suggested to move the non-normative text removed from TRS into a non-normative note.
[07:10] Andrew Berezovskyi (KTH): Also, guidance on base page is incomplete, esp. how to proceed with the base update.
[07:11] Andrew Berezovskyi (KTH): Access control part should require minimal changes from the previous draft.
[07:12] Andrew Berezovskyi (KTH): TRS patch is also not finished, but will require minimal further updates.
[07:14] Andrew Berezovskyi (KTH): While client guidance shall be minimal, the server shall perform in a consistent way that will not confuse the client. Paged base requires additional guidance for the server.
[07:15] Andrew Berezovskyi (KTH): Creation of the base is a long-running operation, so updates can occur while the base is recalculated. Paged base creates additional problems for handling this.
[07:17] Andrew Berezovskyi (KTH): Servers should be able to detect a client that is half-way through reading the paged base to prevent conflicts.
[07:19] Andrew Berezovskyi (KTH): Paged base cannot be updated atomically with the change log update. We cannot force the linearisability of the change events as it would be a big burden.
[07:21] David Honey (Persistent/IBM): What's the relative priority of these bits of outstanding work?
[07:23] Andrew Berezovskyi (KTH): Example: first, the base is calculated. next, a new base is being recalculated. the recommendation is to keep a considerable part of the changelog covering the events before the cutoff in a new base.
[07:25] Andrew Berezovskyi (KTH): the amount of the events that need to be kept differs depending on how old is the old base
[07:25] Andrew Berezovskyi (KTH): TRS is being in active use, while Query has much less active use.
[07:30] Andrew Berezovskyi (KTH): Axel is volunteering to lead the editorial process of the TRS specs. Jim will create the non-normative committee note and will fill the TBDs in the official spec.
[07:32] Andrew Berezovskyi (KTH): The TRS spec had problems before with the duplication of some guidance, esp. paging-related, should be taken care of.
[07:32] Andrew Berezovskyi (KTH): The meeting is adjourned.
|
|
|
Dr. Martin Sarabura |
