RPM roadmap

[pmatilai]

As some of you may have noticed, we quietly posted RPM roadmap on the RPM home page two weeks ago. The roadmap is not an April Fools' joke, only the timing is.

To my knowledge this is the first time a (public) roadmap for RPM has been published, so perhaps some explanations are in order. The roadmap is fairly terse and vague, on purpose. It's there to provide some insight as to where RPM is headed in the next few years. What we are announcing here is the general direction and select features/goals within that direction that we will be working towards in the given time-frame. It is not an exclusive list by any means, and I'm quite positive there will continue to be wonderful surprises from the community regardless of our plans.

RPM releases are made on time-based, annual cycle. An alpha version is released in Q2 and final in Q3, with beta(s) and release candidate(s) in between. This is how we've done it in the last few years anyway, so this is just making the de-facto official.

The obvious Big Thing in the roadmap is RPM v6 of course. Before you all get your ponies queued up, let me repeat the project description:

RPM v4 format is getting long in the tooth, this is showing in many ways such as use of 32bit integers which are running out and/or are dealt with complicated multi tag setups which external users are generally not even aware of.

This project is about mapping the needs and planning ahead of what we'd like to change for an evolutionary step to shake off some legacy baggage and clarify dark corners, not about reinventing a whole new package format.

The major goals for v6 and thus for the next years are:

• a proper format specification
• implementation of that specification
• ease of packaging, aka hands-free packaging

As stated in the project description, the format itself will be "only" a facelift firmly based off the current v4 format, but updated for modern needs in terms of storage sizes, crypto and so on. The details will be worked out and a specification drafted during the next year or two, as per the "RPM v6 scoping" in the roadmap.

Hands-free packaging is vague on purpose because it's just a general direction. There are some examples in the roadmap though, such as declarative user/group handling which is one of those obvious gaping holes in the current feature-set. That is the kind of thing we want to have available in v6 from the go. Also, mechanisms for language/tool ecosystems to build on for their specific packaging needs. There will be others, but as the roadmap hints between the lines, this will not be a multi-year branched development effort but a time-based release just like the others. It also means that the last v4 release is expected to handle v6 packages to a large degree.

As there will be questions, please ask them in the RPM roadmap and v6 questions topic. We'll update this main announcement with FAQ type info as necessary.