Skip to content

Source Code Contributions

Jump to bottom
Christopher Ross-Gill edited this page Aug 1, 2016 · 1 revision

We invite developers to contribute to the R3 source code base.

Table of Contents

Invitation

For R3 to improve, you are invited to contribute to its source code, documentation, tests, examples, and applications.

We will make it easy for developers to obtain source code online and to upload new submissions. We don't want to hinder the flow of contributions with a complicated process.

All code contributions will be freely published and distributed. Everyone can use this codebase for their projects and products, without restriction.

As a developer, your submission must agree to this condition. And, like other languages, we will be creating a "legal entity" to protect the source code base.

Standard Releases

For source code that is part of Rebol 3 itself, there will be a standard version that is integrated. This is the official source that will be supported.

Not everything that is contributed will be part of the standard. But, that's ok. You can always download other code yourself, and even include it within your own applications. That's for you to decide.

As for the official standard, we need to support it over a long period of time, so it must be something that can be supported over that time. In addition, we want to keep the standard system lean and mean. This is our no-bloat principle.

The Main Problems

As with any open source project, it is important to recognize a few things that we will need to manage carefully:

  1. A developer may come and go. Someone who adds a feature this month, may not be around next month to fix it or otherwise support it. That can lead to some big problems over time.
  2. A developer may not learn the model. Some developers add features or concepts without understanding how the system really works. They may even duplicate features that already exist, or add a slightly different variation as a new element, rather than changing the prior one.
  3. A developer's code may be too complex. A lot of developers do not spend time simplifying their code. Sometimes they just hack on it until it seems to work, then submit it.

How to Avoid Them

What can be done about these problems? I think a lot of this is related to educating developers, and some of it will be how we decide to manage the code. In general, we need to make sure:

  1. Code is understandable by more than just the one contributor. Sometimes this is related to code structure, and other times it is related to code commenting or documentation.
  2. That the design model is documented well enough for developers to understand it. Now, not every developer likes to read such documentation, but all we can say is that if their code is too different from our model, their contribution will never become part of the standard design.
  3. Code that is too complex will not be accepted. If a developer does not have the time to polish the code, then he (or we) can find someone else to help work on that.
I think all of this can be summarized by this simple statement:
We will not accept code that we believe to be unsupportable.
Clone this wiki locally