On Saturday, February 16, 2002, at 02:25 AM, Dawid Weiss wrote:
Kimbro,
I looked at your proposal and two things are missing, which we found very
important for our current project (and had to develop ourselves to tell the
truth).
1. Hierarchical locking
A hierarchical locking manager (locks on collections/ subcollections). This
should perhaps be implemented as part of the transactional service (because
we don't want all users to be pending for a long transaction locking the
entire collection and subcollections.
What's the use case for this? I would think our goal should be a more granular locking scheme, locking entire portions of the tree seems like a bad idea. But, then I don't know what you're actually using it for.
2. Versioning
Optional, transparent versioning of resources is of extreme use in many
projects. XHive has it, if you want to take a look at an example solution.
I agree, this could be important and has been brought up several times in the past. It's also possible to use it to help with implementing transactions.
I don't claim these are the most important features in the world, but they
the fact that many commercial vendors include them in their product should
be convincing enough to at least include them in the 'todo' list.
Kimbro Staken XML Database Software, Consulting and Writing http://www.xmldatabases.org/
