2 posts tagged with "refactor"

During the last quarter, we've been on a slow but sturdy path toward the monolith version of Edlib. As we consolidate and simplify the code, we've also been able to work on improving the test coverage in Content Author in particular. From the implementation of Codecov at the beginning of April, we've managed to improve from 54 % to 60 % coverage, which is quite a formidable change. Once we merge in the new monolith core of Edlib, we'll be able to present a more correct picture of the total coverage of the PHP code and the project in total. We are also moving away from using Node in the core to more PHP. This does not limit the use of Node or any other language when creating content types, as the new architecture will be content type agnostic, meaning they will be more detached and bound to the core in a more generic way. This is done to make Edlib more attractive as a development platform and keep complexity down.

A new design for the new Edlib Explorer is in the making, and we are doing the necessary research to ensure the UX and UI are done right and solve real problems for real users. We are also making sure Edlib works better on small devices through an improved responsive design.

We have also recently gone through the getting started instructions. By adding some missing documentation, fixing some failing processes, and, last but not least, making the instructions more precise, we aim to improve the process.

For the next quarter, our main focus will be finishing up the work with the Edlib monolith and implementing the new design and making sure it's all compatible with Content Author. We'll also be doing a fair share of bugfixing and you can read the other planned activities in our updated roadmap.

If you would like to know more about the project, or have any questions, feel free to contact us!

All the best

The Edlib team

Since the last blog post, we've been continuously working on improving the Edlib "plant", and in reference to the previous blog post, we've concluded that we need to replant it. This means we are now rewriting and consolidating most of the code to make further development more straightforward and secure. The need for a better overview and control has been growing simultaneously as the code base has grown. Part of the problem has been the use of many separate services, and now the decision has been made to move towards a monolithic Edlib. Besides being a big job, it will also provide us with an excellent opportunity to document and sort out flukes we've been unable to touch earlier. This enables us to reconsider the essence of Edlib - what to keep and what to throw away.

For instance, this means assessing and swapping out problematic or heavy services, such as Open Search, with the more lightweight Meilisearch. We'll also be looking into what systems for operations are sufficient and beneficial for Edlib. It is no secret that K8s is a resource-demanding system and is most likely overkill for the current needs. We are swapping K8s out with the serverless Lambda with Laravel Vapor on top of it.
As the roadmap indicates, this work will occur in the first half of 2023.

Sadly, Edstep - one of the platforms using Edlib, had to be discontinued since the last blog post, but all the learning content used within Edstep lives on in Edlib and can be accessed through Gamilab. Perhaps Edstep will rise like a phoenix from the ashes one day.

If you have any questions, feel free to contact us!

All the best

The Edlib team