Open Source Open Forum

[Heater]

Not sure I follow all the happenings and ideas here but I gather there is a wish for:

1) Fostering a community of XMOS user, this forum for example.
2) Fostering a community of software developers, individual and/or corporate, who a willing to pool their efforts in creating a wealth of open source software for XMOS devices.

As such I would encourage a everyone interested in such goals to check out the open source software library of Parallax Inc. The so called "OBEX", which contains a wealth of reusable software objects for the Parallax Propeller micro-controller. http://obex.parallax.com/

There are striking similarities between the Propeller and XMOS devices:
1) Multiple independent cores/threads with deterministic timing.
2) The almost total lack of traditional peripheral hardware like UARTS, SPI etc etc.
3) The expectation that these peripherals are create in software.

These properties make it possible to fetch software objects for the Propeller from OBEX, an SD card driver say, and drop them into a core on the Prop with the total confidence that the execution of that driver will have no effect on the execution or timing of all the other software making up the application under construction.

My dream is that one day there would be a similarly vast software exchange for XMOS devices and that "plugging and playing" objects from that repository would be as simple as it is for the Parallax Propeller.

[davelacey]

I'm slightly dubious of an event in the way Jonathan has presented it here. My problem is that it seems like a lot of talk about "meta" issues where there is no current working model. I think I buy xmark's "critical mass" argument on this - there is no point having an event whose main purpose is discussing a lot of high level issues when there is not a diverse set (i.e. non-Xmos) people in the midst of working on the codebases. It may be that in 6 months or a years time we still have the same situation in which case we may have to rethink - but at the moment I don't think we have evidence that there are significant barriers to participation here and in the future we will have more contributors who can also contribute to meta-level discussions in the spirit of a meritocracy (which is the model a lot of open source communities strive for - I feel we should be no exception).

However, I also buy Jonathan's argument that promotion is good. My take on it is that promotion needs some more meaty content to it.

I'd rather that if there was an event it had a more technical focus (like the Xlinkers event was). So it mainly involves presentations about actual projects on XMOS devices, tutorial sessions, bofs on technical issues etc. I think can also be panel discussions on "meta" issues as well but I'm sceptical about this being the main focus. If it did have this more techie focus I also think the name might be better as something like "The XCore developers' forum" or some such.

I think having this slightly different focus would be more likely to encourage developers from companies using XMOS for high-volume commercial applications to attend (though I'd advise people not to get their hopes up too much here). In fact I think we want to encourage engineers in general (the ones that will actually do the core contributing) and attending an event with lots of technical content is much more appealing then attending one that has a lot of people talking about communities and future directions in the abstract.

Also, a three monthly period seems very frequent given the number of people involved and the effort it takes to attend. To take another example, the LLVM developers' meeting is yearly - and that is with a much bigger user base. I'd rather have something less frequent with lots of attendees than something frequent with dribs and drabs of people.

Dave

[jonathan]

Dave - think we're basically talking about the same thing, but I may have presented it wrong. I want a tech gathering, but to use the "new isn't open source cool etc" as a way of framing it. Giving it a theme... I don't mind the theme being something else at all! But it's easier to generate interest at all levels if you can clearly enunciate what the theme is.

I take your point regarding "meta-issues" but I also think you're underestimating the importance of getting this stuff right. When you go out and promote the community and/or the open source project(s), if your licenses are bungled/unclear/not-OSI compliant, you have no roadmap, no governance structure, you have a limited codebase and an almost non-existent community of contributors, you're going to find it very, very hard to be taken seriously.

Some kind of "event" to mark the start of something new is a way of acknowledging that you are small but you have an ambition to listen to ways in which you can grow. It avoids the bluster of content-free promotion and gives you something to talk about and invite people to.

In conclusion, I would love to see an event that gets together a load of people around the theme of something. Open source? XMOS tech users? Whatever - I don't mind, just pick a topic that can be announced...

This might sound a bit facetious, but these events will never happen if we spend time discussing the exact purpose and structure: these are meta issues in themselves. What is required is a commitment to bring the community of users - of all types - together (and soon), and the drive + resources to make it happen.

You can suck a huge amount of time + energy into getting to a point where you finally agree something, and then find there's no time or energy left to actually make it happen.