Feature Freeze for OCCI Core & Model
Folks, since the OCCI core and model seems to have stabilized now (after hard and tempting discussions), I would suggest that we do a _formal feature freeze_ to enforce some stability for the other parts of the OCCI specifications, *effective immediately*. The Feature Freeze is effective immediately. If there are any serious objections against it, please let Thijs know until Monday, October 18, 2010, 12:00pm UTC. Best, Thijs & Alexander
There are a few items I would like to clear out before a complete feature freeze of Core. - Have there been a decision on whether to solve the Link "source" attribute problem in HTTP rendering or in Core? If decided to solve this in Core changes are necessary. I vote for adding the Link "source" attribute workaround to HTTP rendering doc _only_. - I believe it is mostly decided to not use the http://scheme.ogf.org/occi/core#tag Category for user-defined Categories. If so, no changes needed to Core. But if not, changes would be necessary. I vote for skipping the "tag" Category altogether since it is unnecessary. User-defined Categories can be implemented with Core as it is today. - Collections are not explicitly described in Core. However since the use of Categories automatically implies collections I would like to add a section to Core to further clarify this. I can do the change this weekend and it will _not_ change Core, just clarify things. - I would like to further clarify the two use-cases of Categories by introducing the name "type Category" for taxonomy categories and "tag Category" for folksonomy categories. Again, this will not change anything, just clarify things. regards, Ralf On Fri, 15 Oct 2010 16:51:22 +0200, <alexander.papaspyrou@tu-dortmund.de> wrote:
Folks,
since the OCCI core and model seems to have stabilized now (after hard and tempting discussions), I would suggest that we do a
_formal feature freeze_
to enforce some stability for the other parts of the OCCI specifications, *effective immediately*.
The Feature Freeze is effective immediately. If there are any serious objections against it, please let Thijs know
until Monday, October 18, 2010, 12:00pm UTC.
Best, Thijs & Alexander _______________________________________________ occi-wg mailing list occi-wg@ogf.org http://www.ogf.org/mailman/listinfo/occi-wg
Am 15.10.2010 um 17:26 schrieb Ralf Nyren:
There are a few items I would like to clear out before a complete feature freeze of Core.
- Have there been a decision on whether to solve the Link "source" attribute problem in HTTP rendering or in Core? If decided to solve this in Core changes are necessary. I vote for adding the Link "source" attribute workaround to HTTP rendering doc _only_.
+1
- I believe it is mostly decided to not use the http://scheme.ogf.org/occi/core#tag Category for user-defined Categories. If so, no changes needed to Core. But if not, changes would be necessary. I vote for skipping the "tag" Category altogether since it is unnecessary. User-defined Categories can be implemented with Core as it is today.
Personally, I don't care about that -- we could or could not make tags normative, it's really just a question of taste. If someone wants it, it would be a no-brainer for me to add it.
- Collections are not explicitly described in Core. However since the use of Categories automatically implies collections I would like to add a section to Core to further clarify this. I can do the change this weekend and it will _not_ change Core, just clarify things.
Good point. Would be great if you take care of this -- I'd handle this issue as well as not breaking the freeze.
- I would like to further clarify the two use-cases of Categories by introducing the name "type Category" for taxonomy categories and "tag Category" for folksonomy categories. Again, this will not change anything, just clarify things.
Sounds reasonable. Maybe then I didn't get your second bullet, though... Regards, Alexander
On Fri, 15 Oct 2010 16:51:22 +0200, <alexander.papaspyrou@tu-dortmund.de> wrote:
Folks,
since the OCCI core and model seems to have stabilized now (after hard and tempting discussions), I would suggest that we do a
_formal feature freeze_
to enforce some stability for the other parts of the OCCI specifications, *effective immediately*.
The Feature Freeze is effective immediately. If there are any serious objections against it, please let Thijs know
until Monday, October 18, 2010, 12:00pm UTC.
Best, Thijs & Alexander _______________________________________________ occi-wg mailing list occi-wg@ogf.org http://www.ogf.org/mailman/listinfo/occi-wg
Am 15.10.2010 um 17:26 schrieb Ralf Nyren:
There are a few items I would like to clear out before a complete feature freeze of Core.
- Have there been a decision on whether to solve the Link "source" attribute problem in HTTP rendering or in Core? If decided to solve this in Core changes are necessary. I vote for adding the Link "source" attribute workaround to HTTP rendering doc _only_.
+1
- I believe it is mostly decided to not use the http://scheme.ogf.org/occi/core#tag Category for user-defined Categories. If so, no changes needed to Core. But if not, changes would be necessary. I vote for skipping the "tag" Category altogether since it is unnecessary. User-defined Categories can be implemented with Core as it is today.
Personally, I don't care about that -- we could or could not make tags normative, it's really just a question of taste. If someone wants it, it would be a no-brainer for me to add it.
- Collections are not explicitly described in Core. However since the use of Categories automatically implies collections I would like to add a section to Core to further clarify this. I can do the change this weekend and it will _not_ change Core, just clarify things.
Good point. Would be great if you take care of this -- I'd handle this issue as well as not breaking the freeze.
- I would like to further clarify the two use-cases of Categories by introducing the name "type Category" for taxonomy categories and "tag Category" for folksonomy categories. Again, this will not change anything, just clarify things.
Sounds reasonable. Maybe then I didn't get your second bullet, though... Regards, Alexander
On Fri, 15 Oct 2010 16:51:22 +0200, <alexander.papaspyrou@tu-dortmund.de> wrote:
Folks,
since the OCCI core and model seems to have stabilized now (after hard and tempting discussions), I would suggest that we do a
_formal feature freeze_
to enforce some stability for the other parts of the OCCI specifications, *effective immediately*.
The Feature Freeze is effective immediately. If there are any serious objections against it, please let Thijs know
until Monday, October 18, 2010, 12:00pm UTC.
Best, Thijs & Alexander _______________________________________________ occi-wg mailing list occi-wg@ogf.org http://www.ogf.org/mailman/listinfo/occi-wg
Hi, I think I'm ok with a freeze for most features. Not happy with the definition of of collection/Category. By definition a Category is a collection with one member. We should clarify that in the editorial. I do have one concern about credentials and... the impact that will have on Resources...There are 3 companies releasing a product that incorporates certificates to manage keys for encrypted storage using Oasis KMIP. Using wrapped certificates, groups of partners can share keys while maintaining privacy and independent key life cycles. Adding credentials to console, networking and external Resources models, we can find one Resource may support multiple credentials. It impacts provisioning and consumption processes. During provisioning, an occi definition may add several credentials at on time. So we will need to support groups (via collections) of credentials on provisioning. I'm having a tough time envisioning the core model without credentials support. I think credentials need to be sibling to Link, Resource and Action. Credential's UML attributes should be "type" and "certificate". "Type" expresses the credential format and should correspond to ANSI X9 TR-31 [X9 TR-31] . The UML "credential" is byte storage for the credential. Since Credential is a subclass of Kind, collections and Links apply. I'm not sure if the Category's "title" attribute would be mandatory, I'm leaning towards the [0] more than the [1]. The credential scheme does not address provider HTTP logon credentials. If we did, we would need to determine which certificate we would use to pass in user/password. cheers, gary On 10/15/2010 8:51 AM, alexander.papaspyrou@tu-dortmund.de wrote:
Folks,
since the OCCI core and model seems to have stabilized now (after hard and tempting discussions), I would suggest that we do a
_formal feature freeze_
to enforce some stability for the other parts of the OCCI specifications, *effective immediately*.
The Feature Freeze is effective immediately. If there are any serious objections against it, please let Thijs know
until Monday, October 18, 2010, 12:00pm UTC.
Best, Thijs& Alexander _______________________________________________ occi-wg mailing list occi-wg@ogf.org http://www.ogf.org/mailman/listinfo/occi-wg
participants (3)
-
alexander.papaspyrou＠tu-dortmund.de -
Gary Mazz -
Ralf Nyren