Open Stack

On May 8, 2014, at 11:33 AM, John Dickinson <me at not.mn> wrote:

> Thanks Troy. That does help. I'm looking forward to seeing the descriptions you have, and I think others will find them helpful too.
> 
> I think the 5 objectstore capabilities are woefully incomplete (at best) or just wrong (at worse). One is generic, one is about auth, one is a feature, and two are about quotas.

A bit about this. The generic category was used to collect the base set of functionality that isn’t useful to consider separately for inclusion or not. At one point, things were broken down more granularly but it didn’t seem to make sense that you would choose one and not the other. That analysis may still be incorrect because I’m far from being an expert on Swift. Auth is one that was separated because it felt like we should decide if Auth through swift (vs. auth through Keystone) should be in core. 

The other items flowed out of extensions and perceived options - again I don’t claim they are correct. So, your worse case assumption could be true.

> 
> What's the best way to make this list better?

I would suggest to give us your take on the right way to consider Swift capabilities give some of the background I’ve supplied. The only additional caveat is that we need to be able to connect existing Tempest tests to capabilities. We can then work on redoing that section of the scoresheet.

> 
> --John
> 
> 
> 
> 
> On May 8, 2014, at 7:58 AM, Troy Toman <troy at tomanator.com> wrote:
> 
>> 
>> On May 8, 2014, at 9:13 AM, John Dickinson <me at not.mn> wrote:
>> 
>>> I'm not sure what to do with this. What are the scoring values? Where did the capabilities list come from (and how are the existing ones defined)? What do the different colors mean?
>> 
>> Scores are essentially binary. 1 is yes. 0 is no. .5 values represent a place where we were unsure or felt an item warranted additional input or discussion.
>> 
>> The capabilities were derived from Tempest tests. I did most of the sorting to group the 700+ tests into a smaller set of capabilities that would be easier to evaluate.  I am working on a brief description for each of these that will hopefully be available soon. Would love any input on where I might have gotten that wrong. We also recognize there are capabilities that may not be represented in Tempest. Eventually we want to be able to track those but it is not a short term priority.
>> 
>> The colors were used to filter out capabilities that looked to be a solid fit with core (green), were questionable or “on the bubble” (yellow) or clearly not a fit for core at this time (red).
>> 
>> We are particularly interested in filling in scores for capabilities that have not yet been evaluated. I think there was not enough familiarity with Swift in our sub group to accurately assess much of it, for instance. We would also love TC input on “.5” scores - particularly as it relates to the TC Future Direction column. 
>> 
>> Let me know if that helps.
>> 
>> Troy
>> 
>>> 
>>> --John
>>> 
>>> 
>>> 
>>> 
>>> On May 8, 2014, at 1:14 AM, Thierry Carrez <thierry at openstack.org> wrote:
>>> 
>>>> Adding the affected PTLs on the CC list.
>>>> Please follow-up to defcore list, Rob will moderate your post through.
>>>> 
>>>> Rob_Hirschfeld at Dell.com wrote:
>>>>> TC members,
>>>>> 
>>>>> I know many of you have been tracking the DefCore committee progress
>>>>> (background: http://robhirschfeld.com/tag/defcore/ ). 
>>>>> 
>>>>> We’ve reached that critical milestone where we have a partial first pass
>>>>> of the core capabilities list.  While we could continue to refine the
>>>>> list, the committee feels that it’s time for TC and community input
>>>>> before we make more adjustments.
>>>>> 
>>>>> The list  is NOT complete and needs your input on several areas.  Here’s
>>>>> the link >
>>>>> https://wiki.openstack.org/w/images/e/e3/DefCore_Capabilities_Scoring.pdf
>>>>> 
>>>>> Here’s the known issues
>>>>> 
>>>>> ·         There are a block of items that are not scored (colored red)
>>>>> against our criteria (see
>>>>> https://wiki.openstack.org/wiki/Governance/CoreCriteria for details).
>>>>> 
>>>>> ·         Also, many of the “TC Future Direction” items are not scored
>>>>> (they have 0.5)
>>>>> 
>>>>> ·         There are many other scores that need review and adjustment
>>>>> 
>>>>> ·         Some of the capabilities are highlighted in red because they
>>>>> need additional review.
>>>>> 
>>>>> 
>>>>> For simplicity, please direct comments back to the DefCore list instead
>>>>> of the TC list.  I’ll make sure that everyone is added to the list if
>>>>> you are not on it and get moderated.
>>>>> 
>>>>> Thanks,
>>>>> 
>>>>> Rob
>>>>> 
>>>>> *______________________________*
>>>>> 
>>>>> *Rob Hirschfeld*
>>>>> OpenStack Board, DefCore Co-Char & Dell Sr. Distinguished Cloud Solution
>>>>> Architect
>>>>> *cell*+1 512 909-7219 *blog* robhirschfeld.com, *twitter* @zehicle
>>>>> Please note, I am based in the CENTRAL (-6) time zone
>>>> 
>>>> -- 
>>>> Thierry Carrez (ttx)
>>> 
>>> _______________________________________________
>>> Defcore-committee mailing list
>>> Defcore-committee at lists.openstack.org
>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/defcore-committee
> 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 841 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://lists.openstack.org/pipermail/defcore-committee/attachments/20140508/26ba8e35/attachment-0001.pgp>