Michael Hunger wrote:
You are a genius. Putting all strings together like this. Great thinking
and +1 from me of course.
Such a perfect development. I was thinking about the stuff Greg talks
about for a while now and found it very promising in almost any kind of
application.
Yeah, that's what I've been thinking too. He's been repeating the same
thing over and over, and it's starting to make a lot of sense. I think
the main problem with his "detractors" on the DDD list is that they need
to validate and rationalize why they are doing it the way they do with
RDBMS's, rather than actually thinking about what he's saying.
And as qi4j is our sandbox / playground we now can try out how this
works and see if we run into any issues.
What we mustn't forget in the process is performance and optimization
possibilities (at least think about them / point them out).
Agree.
Idempotence of operations / replay / projections of state ... Also the
whole SCM (darcs, git, svn) issue where we can learn tons of thing about
this kind of working or the use of logs of changes from the financial
industry (accounts) can surely help us a lot.
Yeah. As you know, in StreamFlow (GTD for Swedish government) there is a
ton of such requirements. I had a long architectural meeting yesterday
with the customer, and it was all centered around auditing, logging,
backups, security, data integrity, etc. I think what we are doing here
with EventSourcing and using BASE principles fit right into that line of
thinking.
What we also talked about is how this relates to versioning and
retrieving older versions of our entity state (WORLD). Or having certain
changes excluded from the projected version (i.e. discarded changes from
auditing) or whatever.
What there would perhaps be interesting is to get some more meta
information about the UseCase at prepare. The ES may ignore that OR use
it internally for auditing (user, time, context).
Agree. As I said on the DDD mailing list yesterday, it is important that
we can attach metainfo on both UoW and Usecase, and then use that for
various purposes.
What is also interesting then is to publish changes to Entities as an
atom feed. You have all the information there ...
Yeah, I think there will be lots of atom feeds coming out of this. For
the server, for one, but then also split up into individual entities
too. On the domain side, I need feeds for all changes done to objects
related to a particular customer. All sorts of funky stuff. But it all
becomes so much easier in a BASE world...
/Rickard
_______________________________________________
qi4j-dev mailing list
[email protected]
http://lists.ops4j.org/mailman/listinfo/qi4j-dev
