Interestingly, more and more people have recently discovered the use of the event store as a reading model, leaving predictions and “correct” reading models until they are absolutely necessary.
We all know that working with projections increases complexity. At a minimum, you need to create new models, set the DAL for the reading model, and create forecasts for translating events into modified model models and associate these forecasts with the flow of events from your store. This requires more code, more moving parts, and some of them are not easy to verify. Read-side schema changes also require migration.
It seems that for many scenarios, reading all events (properly partitioned) might be enough to have your “read model”. It will take a long time until the system really grows, so you need to read tens of thousands of events to create a single user interface screen. But before you reach this point, you can just read the events. A file system can be used to store events, although tools like EventStore are free and fairly easy to use. Some indexing may be added.
This approach will allow you to significantly stabilize the domain, you will gain more knowledge about how the system works, configure events and be prepared to prepare for the “correct” model of reading into the system, but you may not need it.
Adam Dymitruk wrote a blog post about this , you might think that it is worth reading, even if you do not want to use this approach, Greg Young also gave the word EventStore as a read model in 2012.
source share