This package provides the infrastructure for building your Meteor app with the CQRS (Command/Query Responsibility Separation) and Event Sourcing principles in mind. This enables you to build complex applications with a strong business logic that is easy to test and reason about.
meteor add space:event-sourcing
If you're new to concepts like CQRS, ES and DDD, read on to get a quick overview.
Storing all the changes (events) to the system, rather than just its current state.
You have. Almost all database systems use a log for storing all changes applied to the database (OPLOG anyone?). In a pinch, the current state of the database can be recreated from this transaction log. This is a kind of event store. Event sourcing just means following this idea to its conclusion and using such a log as the primary source of data.
- History: Having a true history of the system. Gives further benefits such as audit and traceability.
- Answers: You never know which questions about your system/users you will ask in one year.
- Time travel: Ability to put the system in any prior state. (I.e. what did the system look like last week?)
- Flexibility: By storing all events your can create arbitrary read-model projections at any time in the project. This enables you to present your data in any number of ways and optimize it for the client-side.
- Speed: Events are always appended, never updated or deleted which makes storing them blazing fast.
You can read more about Event Sourcing here:
Command/Query Responsibility Separation
We segregate the responsibility between commands (write requests) and queries (read requests). Where you had just one model to read/write data from/to your system, you now have two. Each one is optimized for its purpose, reading or writing.
The true strength of CQRS lies in the combination with Event Sourcing. Your event-store becomes the write model side of your system (optimized for business logic). The projected data-structures based on your event stream, become the read model (optimized for the requirements of the UI / client-side).
- You can optimize the reading / writing to your system separately.
- Your business logic becomes simpler because you don't need to think about UI concerns.
- You can create any number of different read-models (think: Mongo.Collections) based on your event stream. And rebuild them from scratch at any time in the project!
- Much faster data subscriptions/loading because the collections can be optimized for the UI.
- No more reactive JOINS or other nonsense that does not perform.
You can read more about CQRS here:
- Microsoft about CQRS
- CQRS FAQ
- Tutorial about CQRS & Event Sourcing (C# but still relevant)
- Martin Fowler about CQRS
Structure, practices and terminology for making design decisions in complex software.
The field for which a system is built.
Airport management, insurance sales, coffee shops, orbital flight, you name it. It's not unusual for an application to span several different domains. For example, an online retail system might be working in the domains of shipping (picking appropriate ways to deliver, depending on items and destination), pricing (including promotions and user-specific pricing by, say, location), and recommendations (calculating related products by purchase history).
"A useful approximation to the problem at hand." -- Gerry Sussman
An Employee
class is not a real employee. It models a real employee. We know
that the model does not capture everything about real employees, and that's not
the point of it. It's only meant to capture what we are interested in for the
current context.
Different domains may be interested in different ways to model the same thing. For example, the salary department and the human resources department may model employees in different ways.
It is a development approach that deeply values the domain model and connects it to the implementation. DDD was coined and initially developed by Eric Evans in his great book Domain-Driven Design: Tackling Complexity in the Heart of Software.
You can read more about DDD here:
- An Introduction to Domain Driven Design
- Domain Driven Design Quickly - Great Book!
- Common DDD mistakes
More and better documentation is coming soon …
meteor:event-sourcing
provides distributed Event Sourcing infrastructure,
with CQRS and DDD patterns in mind. It provides a base class for event-sourced Aggregates
and ProcessManagers, and pairs nicely with space:domain
to model
your business domain with Entities,
ValueObjects, and Domain Events.
It also provides a MongoDB based distributed EventStore which works a little bit different than "normal" event store implementations because MongoDB doesn't support transactions. To circumvent this, the concept of a commit is used, which bundles multiple events and commands together into one "transaction" commit. Here is a short blog article talking about the basic concept.
It heavily uses the space-messaging
package for Metoer EJSON
and
runtime-check
ed domain events and commands that are automatically serialized
into the MongoDB and restored for you. So you don't have to deal with
serialization concerns anywhere but within your value objects.
In lieu of a formal styleguide, take care to maintain the existing coding style. Add unit / integration tests for any new or changed functionality.
meteor test-packages ./
You can find the release history in the changelog
Thanks to CQRS FAQ (Creative Commons) for a lot of inspiration and copy.
Licensed under the MIT license.