Jazz interoperation with non-Jazz repositories
Are there known limitations with using Jazz proxies for a loosely
coupled software lifecycle tool that uses its own external repository? If so, what particular types of jazz-specific operations are limited? Also, are there any limitations specific to process operations (advisors, participants, events, actions, etc.) when proxies are being used? Thanks -Dave- |
5 answers
Hi Dave,
I take it you've been browsing the Interop component APIs? The Interop component is a generic mechanism which can be used to implement interoperation with external repositories, not something that a Jazz user can make direct use of. We're using it to implement interoperation with Rational ClearQuest. The basic idea is that objects in an external repository can be associated with analogous objects in a Jazz repository. Properties of the external object can be imported into Jazz and then applied to properties of the associated Jazz object, and vice versa. The ExternalProxy type represents the linkage between a Jazz object and an external object, and also the current state of the synchronization between the two objects. "Synchronization rules" must be defined to declare how properties are mapped between Jazz and external objects. At this point (in the currently released builds), there is essentially no support for interoperating with any particular type of Jazz object; only the generic framework is there. As I said, we're using it to create an interoperation mechanism between ClearQuest records and Jazz work items, but none of that is included in any current builds. So the answer to your question about limitations is essentially "totally limited", in that there really isn't enough there yet for someone outside the Jazz development team to use. As far as process operations go, any modifications to Jazz objects that are on behalf of interoperation will be subject to the same process rules as if a user tried to do the same modifications. The Interop component is acting as an "agent" for the external repository user. Changes from an external repository may or may not be successfully applied to a Jazz object, depending on process rules. Hope this helps, John |
Hi Michael,
We are using ClearQuest. We have modeled our process in ClearQuest. It's certainly our intention to make it as generic and flexible as possible, but I'd hesitate to say that it will truly support *any* schema, since ClearQuest schema design is pretty wide open, while Jazz items have fairly fixed properties. You can dream up all kinds of fields for CQ records, not all of which will be directly mappable to Jazz work item properties. Of course, you could only map the subset that is mappable, or also cause the non-mappable fields to be stored as custom attributes on work items. The other caveat is that we'll support a number of straightforward transformations between CQ and Jazz property values (e.g. straight copy, or mapping one enumeration to another), but for anything complicated or unusual, you'd have to write Java code to implement your own mapping algorithms. John |
Hi John,
As I said, we're using it to create an We are using ClearQuest. We have modeled our process in ClearQuest. I don't know the details of our ClearQuest customization, but I know it was heavily customized. I would expect that we would have to do some work to map our ClearQuest entries to Jazz work items. Or is the ClearQuest integration generic enough to work with any ClearQuest schema? Michael |
Thanks for the info John!
Some new questions regarding team process. 1) Does/will the jazz framework and it's components provide a set of predefined process advisors/participants and event process handlers that can be enabled/disabled/tweaked so that an administrator can change process flow? Is there a list of somewhere? I didn't see these when navigating the "process state" and "process implementation" views in the Jazz .5 binary image 2) Are there any plans for a more graphical representation of overall process flow when customizing the process definition? I've played with the content-assist feature though it still results in xml input. Thanks -Dave- jrvasta@us.ibm.com wrote: Hi Dave, |
Dave,
The process specification in project areas and the process customization in team areas list only those process enabled operations, e.g., scm client deliver, for which a specific behavior (permissions, advisors, participants) is defined. In order to define process specific behavior you have to add the process enabled operation to the process specification/customization and define its permissions, advisors, and participants. The list of predefined advisors and participants is currently only discoverable by using content assist inside the advisors/participants XML element. The same holds true for event handlers, only that you have to list the events for which you want to have custom behavior. We will provide better discoverability means. The concrete timeline is not yet fixed. Customizing the process in team areas and changing the process specification of a project area are process enabled operations themselves. This means that process defines which roles have the permission to execute this operations. Thus, it depends on whether these operations are limited to an Admin role or are available for others as well. We will work on a better presentation of process specific information such as where does the process configuration for a given operation comes from when this operation is executed in the context of a given team area by somebody playing a given set of roles. This will allow for a better understanding of what the actually active process for a given team area/role set combination looks like and how it gets composed at runtime. The concrete timeline for this work is not yet fixed. Cheers, Kai "Dave Colasurdo" <davecola@earthlink.net> wrote in message news:f2fn3t$kpt$1@localhost.localdomain... Thanks for the info John! |
Your answer
Dashboards and work items are no longer publicly available, so some links may be invalid. We now provide similar information through other means. Learn more here.