"Sébastien Demathieu" <sebastien.demathieu@xxxxxxxxxxxxxxx> wrote in message
news:25e51269685119cd90907e244ee66341$1@xxxxxxxxxxxxxxxxxx
Hi Tas,
> Sébastien,
> Its pretty hard to find a part of MDDi which isn't relevant to our
project
> however there are some areas which are particularly interesting.
That's great! Would you or your company be involved in the MDDi project?
Anyone interested in contributing to this project is very welcome.
> The model
> repository is probably the most important right now and is the one we
will
> be addressing in the next phase. The current repository we have uses a
sql
> database wrapped in a java api. It's really too primitive, proprietary
and
> has become too inflexible. We will be completely redesigning it.
At this point, we do not plan to provide a repository. We will focus on
integration of existing repositories (EMF-based or others) using the
ModelBus.
>Here is are
> some of the initial features we are looking to add add/refine.
>
> * Transactional Version Control
> * Pluggable Persistence (JDBC, XML, Version Control Systems etc)
> * Multi-user support enabling different users to concurrently work on
> difference parts of a model
> * Management of inter-model dependencies and automated re-validation of
> related models
> * Allow customisable meta-data to be become associated with models. E.g
> (completed/incomplete, abstraction level, markup type)
Some of the features you are looking for seem to be in the scope of MDDi.
The pluggable persistence features can be handled by different model
repositories (based on JDBC, XML or others) connected to the ModelBus.
In any case, the ModelBus will support concurent accesses between the
different tools.
The management of inter-model dependencies is definitely an interesting
idea.
I do not think it is supported by the ModelBus. However, a tool that would
provide such features would be really in the scope of the project.
> It would be helpful to understand what the current plans are for
addressing
> the repository aspect of MDDi. I'm imagining the focus be on the
defining
> the interface between repositories and the model bus. Right?
You're right, a great deal of work will be done around the interfaces
between
the tools (e.g. the repositories) and the ModelBus. Tools will be
connected to the bus thanks to adapters. Some adapters will be initially
available, other will be developed later, if required. An integration kit
for the ModelBus will be released as well.
> If I understand
> the current proposal correctly the model bus architecture would
discourage
> direct collaboration with a model repository and instead models would be
> stored and retrieved via the model bus. I am especially interested in
how
> this might work taking into account functionality such as version
control
> and concurrent access.
You're right, one of the advantages of the ModelBus is that it will
individually interface each tool, so that it can cooperate without any
direct knowledge of other tools specifics (operating mode, connectivity,
exchange format).
Regards,
Sébastien
> Tas.
> "Sébastien Demathieu" <sebastien.demathieu@xxxxxxxxxxxxxxx> wrote in
message
> news:3fdf4393545290a2d44d6b1b2a06fc40$1@xxxxxxxxxxxxxxxxxx
>> Hi Tas,
>>
>> Your project seems quite similar to what we are trying to do with MDDi.
>> The project proposal is being completed and your feedback would be
>> valuable.
>> Please let us know if you have any comment about specific points of the
>> proposal.
>>
>> As the MDDi project covers a wide range of activities, I was also
>> wondering if
>> you have any specific interest in the project.
>>
>> Thanks.
>>
>> Sébastien
>>
>> Tas Frangoullides wrote:
>>
>> > Hi,
>>
>>
>>
>> > Just thought I'd let you know that I'm particularly interested in
this
>> > project and might have something to contribute (and receive) from it.
>>
>>
>>
>> > The company I work for has a product which uses models to drive
business
>> > process monitoring. I have a keen interest in model driven approaches
> and I
>> > have recently started a project to migrate the product into more open
>> > architecture by steadily replacing its propriety meta-models, model
>> > repository and modelling tools with standards based and open source
>> > components where possible. The ultimate goal was to define a model
> driven
>> > development framework based on eclipse, emf and uml2, which would be
>> > published to the open source community. This Eclipse project seems to
> have
>> > similar goals and I was delighted when I saw the proposal. This makes
> our
>> > task a lot easier and it makes sense for us to put our efforts into
an
>> > existing open source project rather than create our own.
>>
>>
>>
>> > I look forward to following the project and hope to be of value to
it.
>>
>>
>>
>> > Tas.
>>
>>