Homepage GitHub

V1.0 release checklist

The old combinatorial explosion problem in PyDSDL was one of the potential release blockers because it may have required amendments to the DSDL specification. The problem was fixed recently so this blocker is removed (no spec changes were required).

Earlier we discussed that no stable release should take place until v1 is actually deployed in production. We already know that v1 is being leveraged in PX4 and by some of its adopters. Although the numbers are slim, the positive trend is perceptible. Here are the search results from GitHub of a UAVCAN v1-specific keyword (v0 is NOT included in these results, also, the UAVCAN organization itself is excluded): https://github.com/search?q=canardTxPush+-user%3Auavcan&type=Code

There are also independent third-party implementations which I mentioned in the 2020 recap and roadmap.

I am concerned that the delay of the official stable release might be driving the wrong message to current and prospective adopters, slowing down the adoption of v1. Considering that there currently are no known technical obstacles, should we not go ahead and publish the stable version now, or at least define a more specific set of objectives that have to be attained to make it possible?

Given the beta world we live in I don’t think that label is a huge blocker.

Shouldn’t the consortium be involved in making release determinations now?

The decisions are not going to make themselves. I don’t presume there will be any objections since the Consortium itself, being formed just recently, was never involved in the design of UAVCAN.

The reason I raised this question is this message we received not that long ago:

We are thinking of becoming a member, but can we get the newer and latest specification than the specification version 1.0 beta if we become a member?