Hi Nuno, What I see is that there are two different things happening here: - Observation - Action So let's stick to the "real" world: For our 'observation'
Hi Greg, Third try and last. I think you will see precisely where we diverge perception. I'll use you cup example. Here is a recap. ... I agreed that the
Hi Mike, There is nothing innately wrong with transaction script. That being said, you should not assume that the "transaction script" in the repository is
I would argue that this is the wrong way. In this case your repository is becoming more of a transaction script (and you run the risk of having anemic
To be fair all of your requests will probably be retrieved in one round-trip select * from Requests req join Objectives obj on obj.RequestId=req.Id where
So this is where LINQ comes in. I'm not sure how well NHibernate supports this scenario but EF supports projections like this var address = (from order in
After watching videos from Greg Young and Udi Dahan I am finally starting to understand CQRS with Event Sourcing. At least I think so ;) I have some questions.
Let me walkthrough what we do now: How do you synch changes? <- The update event will will the in-memory datastore, for transactional data; and update the
Hi Greg, ... I understand that you disagree. I'll try better one more time. ... Maybe it is me that doesn't understand what the fuss is all about eventual
Imagine a "Phonecall" object. It's an inmutable object, that can answer you it's duration, its source and destination number, aaand that's it. If we assign
Hi Vaughn, Sorry for delaying, I usually don't get connected on weekends. Wife is time consuming :) ... I see the Context Map pattern as a clear solution to be
... Ah ... so what you are saying is: ... Assume data for different entities are spread out on different servers. Entity A is stored on Server X, and Entity B
... Err, sorry. Not in same memory as the application process. I was thinking about the individual shard servers having in-memory copies of their data.
To be clear, in many cases a network hit to load in a typical stateless way is not acceptable for various reasons (including latency). Mongo db with a set of
umm mongodb is not in memory in your process as was being discussed unless I missed some great new product named mongodb that came out. ... -- Les erreurs de
And for some inspiration: MongoDB sharding does all of this for you :-) Read "Life beyond distributed transactions":
"I'm not refuting the fact that the world act over inconsistent information. But the observer always act according to the with the infotmation observed." How
Hi Jorn, Yes, that is more or less what I did in my little sample. But in the end of the day in our context I just don't see the advantage of sending a command
How do you synch changes? Or if domain is too big for one server you have to add stickiness (likely a message router). In this case, how do you handle
yes for now our application are in single server for difference domain. however i am not sure what are the complexity specific for in memory datasource, would