Hi Frank, all After the WG call on Wednesday I started to think about the implications of the 'Published Resources' term that I proposed.
I want to understand the relationship between 'publishing resources' to be made available via the change log, and 'publishing resources' as a general OSLC API construct. Are these two things always describing the same capability, and therefore the same set of resources? I believe so. Whilst I don't believe that all indexers will always want to index all available resources (often indexing a subset of resources might be desired), I do believe that we shouldn't second-guess the use-cases, and thus the set of published resources for the change log should indeed be the exact same set of resources 'published' via the generic OSLC API. (Note this doesn't imply that providers should not also be able to make subsets of resources available via additional Query Capabilities) In other words 'published resources' is a term with a single well defined meaning (the total set of resources available via the providers OSLC service) that has applicability beyond just the ChangeLog proposal. What I don't understand therefore is why we actually need a specific term for a capability here? (Resource Publishing Capability) Aren't we just saying that the provider needs to expose at least one Query Capability (oslc:queryBase)? On the naming issue around the tracking/logging capability my thoughts are: - tracking implies a partial assumption of client use case - logging describes the mechanism by which the information is delivered - what's really important is the 'data' or 'information' and how that changes over time (history) - some naming proposals talk about 'change' but are not explicit about the subject of the change (resources) - some naming proposals talk about 'tracking resources' but are not explicit about what is being tracked (changes) - hence I would suggest something along the lines of 'Resource Change Information Capability' 'Resource Change History Capability' If an abbreviated form is really necessary, then 'Resource History' would seem to be to be most descriptive of what the capability provides. Finally, regarding the collection of Core features required to support an index, to remain consistent with usage elsewhere (Reporting) I support 'Indexing Profile' as a preferred term to 'Indexing Requirements' Regards Benjamin Williams Rational Reporting Strategy Lead Senior Product Manager, Rational Reporting IBM Software, Rational Phone: 44-118 9793107 E-mail: [email protected] Find me on: IBM United Kingdom Limited Registered in England and Wales with number 741598 Registered office: PO Box 41, North Harbour, Portsmouth, Hants. PO6 3AU From: Frank Budinsky <[email protected]> To: Frank Budinsky <[email protected]> Cc: oslc-core <[email protected]>, [email protected] Date: 01-04-2011 21:56 Subject: Re: [oslc-core] Better name for OSLC indexing profile Sent by: [email protected] Thanks to everyone that provided feedback on this issue. I now think I have a better way to describe this. As mentioned below, to implement an "indexer" we will need 3 new capabilities of a service provider: 1. Expose its complete set of "public resources". By public, we mean all the resources that the service provider does not consider to be internal implementation artifacts (not to be confused with anything to do with access permission's). At the previous WG call, the word "published resources" was suggested as a good way to describe these resources. 2. Provide a change log that lists change events for the set of published resources in (1), as well as creation events for new resources that would have been included in (1). 3. Provide access to security information (ACLs) for the published resources, to allow a client to mirror the access rules of the underlying tools. If we were to name the capabilities, themselves, based on our discussions I would name them something like this: 1. Resource Publishing 2. Resource Tracking (or Logging) 3. Resource Security All three capabilities are intended to be optional OSLC features, but #2 does prereq #1. That is, you can't provide logging/tracking support without also publishing the resources being tracked. In that sense, you can easily think about #2 as being a profile, instead of a capability. Add to that the fact that #3 isn't part of the ChangeLog proposal (it's an independent capability, and intentionally being left as a separate OSLC proposal), some of the profile name suggestions were, not surprisingly, the same as these #1 or #2 capability names. What I have been trying to do is find an appropriate name for the combination of #1, #2, and #3, which would then be used to talk about the overall functionality and even come up with a NS that could be used for resources of all 3 capabilities. I'm now thinking that isn't really a good idea anyway, since there really are two completely independent proposals in this picture (ChangeLog and Security). Given that, I think we should simply use the term "Indexing Requirements" to refer to the three capabilities that are needed for an indexer, and use the logging/tracking capability (#2) to refer to the ChangeLog functionality. That is, we use the name: "Resource Tracking Capability" --- is "Resource Logging Capability" or "Change Logging Capability" a better/clearer name? which also prereqs the: "Resource Publishing Capability" Unless someone can see a problem with this, I'll update the ChangeLog proposal using this terminology. Since the next version is intended to be the official "convergence" draft, it would be nice to have agreement on this, but I suppose we're still allowed to change things like this before finalization if we want to. Thanks, Frank Frank Budinsky---03/31/2011 04:27:50 PM---Hi all, I'd like to finish this discussion soon, if we can. The discussion so far From: Frank Budinsky/Toronto/IBM@IBMCA To: Jim des Rivieres/Ottawa/IBM@IBMCA Cc: oslc-core <[email protected]> Date: 03/31/2011 04:27 PM Subject: Re: [oslc-core] Better name for OSLC indexing profile Sent by: [email protected] Hi all, I'd like to finish this discussion soon, if we can. The discussion so far has helped us articulate exactly what this "profile" is, even if we haven't been able to come up with a good name yet. Let me try to summarize and then offer a couple of new name suggestions. This profile requires a service provider to implement the following three capabilities: 1. Expose its complete set of "public resources". By public, we mean all the resources that the service provider does not consider to be internal implementation artifacts (not to be confused with anything to do with access permission's). At the previous WG call, the word "published resources" was suggested as a good way to describe these resources. 2. Provide a change log that lists change events for the set of published resources in (1), as well as creation events for new resources that would have been included in (1). 3. Provide access to security information (ACLs) for the published resources, to allow a client to mirror the access rules of the underlying tools. In a more concise form, we can describe this as: A service provider must publish its contents (resources), including change information (because the content is dynamic) and access control information. Jim used the word "Active", below, to capture the dynamic aspect. Sticking with that, I would suggest the following name for the profile: 1. Active Content Publishing 2. Content Publishing Personally, I like #2, "Content Publishing Profile". I think it's short enough and I don't think "Active" is not needed because it's obvious that a service provider's contents is changing/active. A service provider should just be asked to "publish its contents", which means enumerate the resources, including change information to allow clients to keep up to date with the changing publication. Please send me your votes or a better suggestion if you can think of one. Thanks, Frank. Jim des Rivieres---03/24/2011 01:51:07 PM---I don't have as much a problem as some with using the driving use case in the name. I believe in ca From: Jim des Rivieres/Ottawa/IBM To: Frank Budinsky/Toronto/IBM@IBMCA Cc: oslc-core <[email protected]> Date: 03/24/2011 01:51 PM Subject: Re: [oslc-core] Better name for OSLC indexing profile I don't have as much a problem as some with using the driving use case in the name. I believe in calling a spade a bloody shovel [OED], even if shovels can be used for other things. If I step way, way, back and imagine that the particular protocol we are specifying here catches like wildfire in the open web (think Atom or OAuth), the question I ask is what name would we like it to be known by. Something that the world can latch onto, and have a chance of keeping it straight in their minds amid myriad other Internal protocols. I believe the main reason that apps will be signing up to be providers is that they are sitting on some data resources that they would like to expose in an RDF format so that other apps would be able to build queryable indexes from. Any old app can expose resources with RDF representations without supporting this new protocol. So the RDF by itself is not sufficient. And this protocol says nothing about the indexes that may get built. Crucially, this protocol is about exposing information in RDF form in a way that support someone else to build and maintain an RDF-based index. And its not that interesting if the data is static - you'd have no need of a change log. So being a dynamic source also seems important (#1-#5 all touch on this aspect. So how about the "Active RDF Index Source" protocol. And that when we talk about a OSLC domain service provider, we could talk about what it would mean for them to support the "Active RDF Index Source Profile", which would entail implementing the provider role of the Active RDF Index Source protocol, and providing specified enties in OSLC service, service provider, and service provider catalogs that make its Active RDF Index Sources discoverable. Regards, Jim [OED] See entry in http://en.wikipedia.org/wiki/To_call_a_spade_a_spade Frank Budinsky---03/24/2011 11:34:47 AM---As discussed in the meeting yesterday, I've been using the term "Indexing Profile" (see http://open- From: Frank Budinsky/Toronto/IBM@IBMCA To: oslc-core <[email protected]> Date: 03/24/2011 11:34 AM Subject: [oslc-core] Better name for OSLC indexing profile Sent by: [email protected] As discussed in the meeting yesterday, I've been using the term "Indexing Profile" (see http://open-services.net/bin/view/Main/IndexingProposals for details) to describe the capabilities that MUST be provided by a service provider to support indexing (e.g., an enumeration of resources, a Change Log, etc). It was pointed out that, although indexing is the primary motivating use case, the profile we're talking about is really more general and should therefore have a more general name. What this profile really provides is the capabilities required by a client that wants to see content and track changes to content of a service provider. Some suggested names for this profile are: 1. Observer Profile 2. Notification Profile 3. Change Log Profile 4. Content Tracking Profile 5. Service Tracking Profile The first 3 seem not to cover the "enumeration of resources" part of the profile (and other things like security, which will also be part of the profile), so my vote so far would be for #4 or #5. Please let me know which of these name you prefer or if you have a better suggestion. Thanks, Frank. [email protected] wrote on 03/23/2011 11:39:32 AM: > [image removed] > > [oslc-core] Continuing Change Log and Baselines discussions > > Dave > > to: > > oslc-core > > 03/23/2011 11:41 AM > > Sent by: > > [email protected] > > We had a good discussion of Change Log issues today, thanks to Frank > for leading, but we did not finish. You can find my notes from the > meeting on the agenda page here: > > http://open-services.net/bin/view/Main/OslcCoreMeeting20110323 > > Since we did not finish the discussion, I'd like to schedule a > follow-up meeting next week: > > OSLC Core WG Meeting: Change Log follow-up Meeting > March 30, 2011 - 10AM US/ET > > I would also like to propose that we meet the week after that to > follow up on the Baselines discussion: > > OSLC Core WG Meeting: Baselines follow-up Meeting > April 6, 2011 - 10AM US/ET > > Nick: does April 6 work for you? > > Thanks, > Dave > > _______________________________________________ > Oslc-Core mailing list > [email protected] > http://open-services.net/mailman/listinfo/oslc-core_open-services.net _______________________________________________ Oslc-Core mailing list [email protected] http://open-services.net/mailman/listinfo/oslc-core_open-services.net _______________________________________________ Oslc-Core mailing list [email protected] http://open-services.net/mailman/listinfo/oslc-core_open-services.net [attachment "graycol.gif" deleted by Benjamin Williams/UK/IBM] [attachment "ecblank.gif" deleted by Benjamin Williams/UK/IBM] _______________________________________________ Oslc-Core mailing list [email protected] http://open-services.net/mailman/listinfo/oslc-core_open-services.net Unless stated otherwise above: IBM United Kingdom Limited - Registered in England and Wales with number 741598. Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
