Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [mylar-dev] confusion between task categories and queries

Mik Kersten wrote:

 We need to keep the UI parts and the infrastructure API parts
 separate in discussion of how to improve this support.  The UI needs
 to be simple and obvious, and the infrastructure needs to be flexible
 to support new kinds of connectors.  Starting with the UI:

 * The fact that queries can only show the tasks from one repository,
 and not mix local tasks, provides predictability and simplicity.  You
 always know that the things that will show up in a query will be the
 things that match the query parameters.

That is true, but on the other hand queries are often used in place of catgories only because there is no other way to get information about newly created tasks.

So, as we discussed it would be really handy to have another type of container in addition to query and category that would allow to mix tasks and query results. That container would allow to automatically make remote tasks "sticky" when user would open them from query hits, or otherwise drop hits by marking them as read/uninteresting. Then my use case with tasks that don't match query anymore dropped into Archive category should be resolved. Those jumping tasks break per-category time reporting.

Current workaround to this issue is to use query that would bring all the tasks (including closed ones), so they would be filtered out by "closed" filter and still stay under same query folder.

 * The user already does not need a query to see changes to tasks that
 have been added to a category.  All repository tasks with an offline
 handler get synchronized every time if they have changed on the
 server, no matter where they are.

 Not entirely. As it is been pointed out, you can't work with attachments.

 * We already do part of what you suggest: every repository task has a
 local task created for it, by virtue of AbstractRepositoryTask being
 a subclass of Task.  That's how the repository tasks get the Planning
 page with scheduling and notes information.

Right. I am just suggesting to make those tasks more rich and turn synchronization into the other way around.
 That would better support cases like Raphael were describing.

All it seems like we are missing something in repository integration. Right now you can have at most one remote repository per task. But what if I want to have more then one?

That bring us to the idea of task synchronizers. So, each task could have one or more synchronizers assigned to it. They will sync up this task with one or more repositories. This actually does not have to affect current UI experience at all and from the implementation/backend point of view Mylar already have most of the pieces. In a result, advanced users will have better control on task synchronization and allow more transparent integration with tasks managed simultaneously in Bugzilla/Jira and Outlook/MSProject/iCal.

 regards,
 Eugene




Back to the top