D Improvement Proposals (DIPs)

List of old DIPs approved before this repo existed

Purpose

This repository stores and manages improvement proposals for the D programming language. Common examples include a change to existing language semantics, the addition of a new major features to the compiler or enforcement of new process as a standard. In general, any controversial change must be managed as a DIP and thus requires approval by the language authors and feedback from the D community.

Procedure

Submitting a new D Improvement proposal

Write a document for the new improvement proposal based on the template. All sections mentioned in the template are important - for example, a change implying breaking changes has almost no chance to be accepted if it doesn't describe a migration path to mitigate breakage.
Create a new pull request against this repository by adding a new document to the DIPs folder, using any spare ID (>= 1000). The DIP manager will provide feedback about what information needs to be added for the DIP to reach the required quality for further consideration.

The DIP document must be named "DIP.md". The exact DIP ID used at pull request stage is not important because during merging the DIP manager will replace it with the next currently available ID among the merged proposals, which will become the "real" ID for the DIP in future.

The pull request title should match the DIP title.
After any initial feedback has been addressed, the DIP manager will announce the new DIP in the official D newsgroup for community feedback. This will help evaluate stronger and weaker points of the proposal before it gets to the language author's attention.
Once a proposal includes all necessary details and the DIP manager considers it to be ready for evaluation by the language authors, the pull request gets merged with the DIP status being Draft. A DIP pull request should not be merged earlier than one month after the newsgroup announcement to ensure everyone has a chance to comment on it.

Getting a DIP approved

Once every few months the DIP manager has to pick one DIP from those that currently have Draft status. Proposals with more detailed descriptions and/or proof of concept implementations should have a higher priority.
The DIP is brought to the language authors for review. The DIP manager's responsibility is to gather and provide information about the proposal at their request. After each round of review the DIP manager must publish to the mailing list the outcome of the review along with a small summary.
Review should result in the DIP either being moved to Approved status or modified with a list of issues that need to be worked on before a final decision can be made. In case the DIP topic seems important but language authors decide it needs more research, a new topic on the Dlang-study mailing list may be initiated.

Collaborating on DIPs

Anyone can submit new pull requests with updates to merged DIP document as long as the original author is notified.
Discussion regarding the DIP's content is welcome in pull requests - everyone is welcome to participate in the review.
If there are many uncertainties about the proposal, consider first publishing document somewhere else and discussing it via the NG or e-mails. That will greatly reduce the number of back-and-forth changes in the DIP pull request later.

Advices for writing great DIPs

Writing a good informative DIP with solid chances to get approved is very hard work - it is often underestimated how much effort is required.

Focus on "why" as much as "what". At this stage of the D language development changes need to justify their existence with great value. Doubt about the necessity or value of a change will be a strong motivator for rejection.
Be very pedantic and formal in describing any required language/compiler changes. For example, make separate sections for grammar changes, semantic changes, expected compiler error messages and so on. A DIP should provide a detailed design documentation for a compiler developer to implement it.
Research any alternative approaches to solving the same problem and explain your choice. If there are any relevant success or failure stories in other programming languages, consider referring to them with explanation of how they might apply to D.
Showing example of real-world project that could benefit from an improvement can greatly improve its chances of being accepted. D is intended to be a practical language and a recurring questions when new changes are considered is "does it really make real projects better?". If that is not possible, try designing code examples with intention to answer the very same question.
Be prepared for a lot of work. There are always many ideas proposed but many fewer developers commited to pursuing the idea to the final stages of evaluation. The DIP system is not for submitting undeveloped ideas, it is a process for formal approval of language changes.

DIPs by the D language authors

Language changes initiated by language authors are also supposed to go through the DIP queue. By their very nature formal approval is not needed. Hence they are processed slightly different and an increased focus is put into bringing community attention and feedback.

At the time of writing this document only Walter Bright and Andrei Alexandrescu are meant as a language authors here.

The DIP manager responsibilities

The idea behind the role of the DIP manager is to have a person who will do some minimal initial research and quality control, saving time for language authors to focus on the actual decision. That implies gathering information, maintaining this repository and communicating to involved parties so that the process keeps moving forward. Essentially the DIP manager is supposed to act as a proxy between D users and the language authors to help handling the growing scale of DIP information reliably and effectively.

cym13/DIPs