OE, the TSC and the future

OE, the TSC and the future

[Richard Purdie]

I think its fair to say that OpenEmbedded has changed quite a bit over
the last few years. Prior to the Yocto Project, OE struggled to figure
out how to engage with the commercial side of the ecosystem and how to
scale and I think there have been many positive changes in the way
everything works, not least with the layers approach and the pull model
for changes.

Whilst the ecosystem has changed, the structures that make up OE such as
the TSC have not changed that much. They were setup to address problems
which in some cases no longer exist for example.

We're coming up to the next round of TSC elections, the board is aware
of this but have asked that we figure out the TSC's role going forward
before those elections.

There has been limited discussion about this on the members list
previously, it was met with a lot of silence but I do think the time is
right to think about things a bit.

I should note that whilst I did take an action from the TSC to start a
discussion and that the TSC does have some ideas in mind, I'm primarily
expressing personal views here, not those of the TSC. Other TSC members
are more than capable of expressing their own views!

In brief summary the TSC has been doing two main things, acting as a
task force and also being able to make a decision when needed. The
latter has not happened much at all, the main work was as a task force
on various issues, firstly engaging with the Yocto Project and figuring
that out, more recently dealing with infrastructure issues and generally
ensuring the health of OE.

I don't think the task force should be limited to the TSC members
although it can be lead by them (or delegated). As such I think the TSC
would like to see that element get opened up to a public IRC meeting,
maybe monthly at a set time when people get together and discuss those
topics. The TSC members could be responsible for giving that process and
meeting some kind of structure but it would be open to all.

The decision making element of the TSC would remain with them, likely
done by calling a special meeting as needed (we haven't needed many
official decisions).

At this point I'll open that to discussion. Any objections or other
proposals?

Speaking of TSC elections, I know I'm due for re-election first. I'm
also due to be away for a few weeks shortly so I'd like to make it known
that I would like to stand for re-election. I've hopefully done some
good for OE over what amounts to nearly a decade (scary when put like
that!) and I'm not quite done yet! :)

Cheers,

Richard

Re: OE, the TSC and the future

[Philip Balister]

On 06/18/2013 12:23 PM, Richard Purdie wrote:
> I think its fair to say that OpenEmbedded has changed quite a bit over
> the last few years. Prior to the Yocto Project, OE struggled to figure
> out how to engage with the commercial side of the ecosystem and how to
> scale and I think there have been many positive changes in the way
> everything works, not least with the layers approach and the pull model
> for changes.
> 
> Whilst the ecosystem has changed, the structures that make up OE such as
> the TSC have not changed that much. They were setup to address problems
> which in some cases no longer exist for example.
> 
> We're coming up to the next round of TSC elections, the board is aware
> of this but have asked that we figure out the TSC's role going forward
> before those elections.
> 
> There has been limited discussion about this on the members list
> previously, it was met with a lot of silence but I do think the time is
> right to think about things a bit.

In case people are not aware, this is a description if the TSC:

http://www.openembedded.org/wiki/TSC

> 
> I should note that whilst I did take an action from the TSC to start a
> discussion and that the TSC does have some ideas in mind, I'm primarily
> expressing personal views here, not those of the TSC. Other TSC members
> are more than capable of expressing their own views!
> 

OpenEmbedded still needs a group of people to provide technical
leadership that are appointed/elected by the technical people active in
the project. How the TSC provides this leadership is up to them.

I'd like to see the TSC interact more with the larger community and be
less focused on meetings. As the OE board chair, I am fine with having
us chase infrastructure issues and letting the TSC focus on improving
OpenEmbedded.

My challenge to the TSC is how can we improve the community of people
working on OpenEmbedded. Engaging with the Yocto Project has brought in
large numbers of new developers and I feel we have done a poor job
integrating them into the larger OpenEmbedded community.

I had hoped at one point we could use the Collab Summit for developer
meetings, like the old OEDEM (Open Embedded Developer European Meeting)
we had in the past. But the vibe there is all wrong and attendance is by
invite. Also the US West coast location is very painful for people from
Europe. (Mind you, Europe is painful for the US West Coast :)

To give you an idea to kick around, we need to create some form of event
that corporate types can convince their managers is valuable to attend,
while avoiding the typical conference mode of people presenting to rooms
of people. My goal for such an event is have developers and the user
community exchanging ideas with high bandwidth channels. As opposed to
IRC, EMail, and telecons we normally use. (We make good use of these
tools, they each have limitations though)

The word I have seen kicked around is "Unconference".


> In brief summary the TSC has been doing two main things, acting as a
> task force and also being able to make a decision when needed. The
> latter has not happened much at all, the main work was as a task force
> on various issues, firstly engaging with the Yocto Project and figuring
> that out, more recently dealing with infrastructure issues and generally
> ensuring the health of OE.
> 
> I don't think the task force should be limited to the TSC members
> although it can be lead by them (or delegated). As such I think the TSC
> would like to see that element get opened up to a public IRC meeting,
> maybe monthly at a set time when people get together and discuss those
> topics. The TSC members could be responsible for giving that process and
> meeting some kind of structure but it would be open to all.
> 
> The decision making element of the TSC would remain with them, likely
> done by calling a special meeting as needed (we haven't needed many
> official decisions).
> 
> At this point I'll open that to discussion. Any objections or other
> proposals?
> 

Like I said earlier, how the TSC achieves its mission is up to them.
Both of these suggestions looks sane. Having meetings for the sake of
having a meeting is not good.

The TSC is free to call on people as they see fit, however I would
expect any selection process to be very transparent and open to all
interested participants. Maybe a good first step would be document who
is doing what. It would be nice to have a web page on the wiki defining
layer maintainers, branch maintainers, active autobuilders etc. I know
all this is in readme;s etc, but having a central index of active people
would be really helpful to show people the extent of the OpenEmbedded
community.

Philip

> Speaking of TSC elections, I know I'm due for re-election first. I'm
> also due to be away for a few weeks shortly so I'd like to make it known
> that I would like to stand for re-election. I've hopefully done some
> good for OE over what amounts to nearly a decade (scary when put like
> that!) and I'm not quite done yet! :)
> 
> Cheers,
> 
> Richard
> 
> 
> _______________________________________________
> Openembedded-members mailing list
> Openembedded-members@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-members
> 

Re: OE, the TSC and the future

[Richard Purdie]

On Thu, 2013-06-27 at 10:48 -0400, Philip Balister wrote:
> On 06/18/2013 12:23 PM, Richard Purdie wrote:
> > I should note that whilst I did take an action from the TSC to start a
> > discussion and that the TSC does have some ideas in mind, I'm primarily
> > expressing personal views here, not those of the TSC. Other TSC members
> > are more than capable of expressing their own views!
> 
> OpenEmbedded still needs a group of people to provide technical
> leadership that are appointed/elected by the technical people active in
> the project. How the TSC provides this leadership is up to them.

Agreed, I still think there is a role for the TSC there, the time is
just right to think about how it works and whether everything makes
sense.

> I'd like to see the TSC interact more with the larger community and be
> less focused on meetings. As the OE board chair, I am fine with having
> us chase infrastructure issues and letting the TSC focus on improving
> OpenEmbedded.
> 
> My challenge to the TSC is how can we improve the community of people
> working on OpenEmbedded. Engaging with the Yocto Project has brought in
> large numbers of new developers and I feel we have done a poor job
> integrating them into the larger OpenEmbedded community.

Personally, I don't see a split between OE and YP communities. Anyone
submitting changes to OE-Core joins the OE-Core mailing list and becomes
part of the OE community.

I do wonder/worry whether we do get people involved into any of the
projects as much as we could/should though. There aren't that many new
faces taking ownership of areas of the codebase for example. We have
gone to a lot of effort to try and make many things "just work", perhaps
this is partly a result of that? It would be both a good and bad thing
in that case!

> I had hoped at one point we could use the Collab Summit for developer
> meetings, like the old OEDEM (Open Embedded Developer European Meeting)
> we had in the past. But the vibe there is all wrong and attendance is by
> invite. Also the US West coast location is very painful for people from
> Europe. (Mind you, Europe is painful for the US West Coast :)
> 
> To give you an idea to kick around, we need to create some form of event
> that corporate types can convince their managers is valuable to attend,
> while avoiding the typical conference mode of people presenting to rooms
> of people. My goal for such an event is have developers and the user
> community exchanging ideas with high bandwidth channels. As opposed to
> IRC, EMail, and telecons we normally use. (We make good use of these
> tools, they each have limitations though)
> 
> The word I have seen kicked around is "Unconference".

I think the key here is to find someone interested in holding this event
and that person making it happen. There are people around who could help
but I think it would need someone from the OE community to step up and
own it.

> Like I said earlier, how the TSC achieves its mission is up to them.
> Both of these suggestions looks sane. Having meetings for the sake of
> having a meeting is not good.

Agreed, I don't think the TSC has reached that point however there is a
need for wider involvement of others in the community. I can only take
silence in reply to my email as agreement although that makes it hard to
tell :/.

> The TSC is free to call on people as they see fit, however I would
> expect any selection process to be very transparent and open to all
> interested participants. Maybe a good first step would be document who
> is doing what. It would be nice to have a web page on the wiki defining
> layer maintainers, branch maintainers, active autobuilders etc. I know
> all this is in readme;s etc, but having a central index of active people
> would be really helpful to show people the extent of the OpenEmbedded
> community.

Who is going to do it? The information is all there, its perfectly
possible. I'd like to see new faces involved though, the existing TSC
members and board members do enough already. Much as I want to help with
many different things, I don't scale and the usual suspects aren't going
to either indefinitely :/.

Cheers,

Richard

Re: OE, the TSC and the future

[Mark Hatle]

On 6/27/13 10:38 AM, Richard Purdie wrote:
> On Thu, 2013-06-27 at 10:48 -0400, Philip Balister wrote:
>> On 06/18/2013 12:23 PM, Richard Purdie wrote:
>>> I should note that whilst I did take an action from the TSC to start a
>>> discussion and that the TSC does have some ideas in mind, I'm primarily
>>> expressing personal views here, not those of the TSC. Other TSC members
>>> are more than capable of expressing their own views!
>>
>> OpenEmbedded still needs a group of people to provide technical
>> leadership that are appointed/elected by the technical people active in
>> the project. How the TSC provides this leadership is up to them.
>
> Agreed, I still think there is a role for the TSC there, the time is
> just right to think about how it works and whether everything makes
> sense.
>
>> I'd like to see the TSC interact more with the larger community and be
>> less focused on meetings. As the OE board chair, I am fine with having
>> us chase infrastructure issues and letting the TSC focus on improving
>> OpenEmbedded.
>>
>> My challenge to the TSC is how can we improve the community of people
>> working on OpenEmbedded. Engaging with the Yocto Project has brought in
>> large numbers of new developers and I feel we have done a poor job
>> integrating them into the larger OpenEmbedded community.
>
> Personally, I don't see a split between OE and YP communities. Anyone
> submitting changes to OE-Core joins the OE-Core mailing list and becomes
> part of the OE community.
>
> I do wonder/worry whether we do get people involved into any of the
> projects as much as we could/should though. There aren't that many new
> faces taking ownership of areas of the codebase for example. We have
> gone to a lot of effort to try and make many things "just work", perhaps
> this is partly a result of that? It would be both a good and bad thing
> in that case!
>
>> I had hoped at one point we could use the Collab Summit for developer
>> meetings, like the old OEDEM (Open Embedded Developer European Meeting)
>> we had in the past. But the vibe there is all wrong and attendance is by
>> invite. Also the US West coast location is very painful for people from
>> Europe. (Mind you, Europe is painful for the US West Coast :)
>>
>> To give you an idea to kick around, we need to create some form of event
>> that corporate types can convince their managers is valuable to attend,
>> while avoiding the typical conference mode of people presenting to rooms
>> of people. My goal for such an event is have developers and the user
>> community exchanging ideas with high bandwidth channels. As opposed to
>> IRC, EMail, and telecons we normally use. (We make good use of these
>> tools, they each have limitations though)
>>
>> The word I have seen kicked around is "Unconference".
>
> I think the key here is to find someone interested in holding this event
> and that person making it happen. There are people around who could help
> but I think it would need someone from the OE community to step up and
> own it.
>
>> Like I said earlier, how the TSC achieves its mission is up to them.
>> Both of these suggestions looks sane. Having meetings for the sake of
>> having a meeting is not good.
>
> Agreed, I don't think the TSC has reached that point however there is a
> need for wider involvement of others in the community. I can only take
> silence in reply to my email as agreement although that makes it hard to
> tell :/.

In the last couple of years, I can only remember two or three meetings that we 
have nothing to discuss and we simply adjourned the meeting...

The key thing is that the roll of the TSC from technical body and necessary 
"fix-it" specialists is changing.  And we need to figure out how to better to 
get help when something is found to be in need of work.

Some of this will no doubt turn out to be janitorial type work.. and some of it 
may be new development.

>> The TSC is free to call on people as they see fit, however I would
>> expect any selection process to be very transparent and open to all
>> interested participants. Maybe a good first step would be document who
>> is doing what. It would be nice to have a web page on the wiki defining
>> layer maintainers, branch maintainers, active autobuilders etc. I know
>> all this is in readme;s etc, but having a central index of active people
>> would be really helpful to show people the extent of the OpenEmbedded
>> community.
>
> Who is going to do it? The information is all there, its perfectly
> possible. I'd like to see new faces involved though, the existing TSC
> members and board members do enough already. Much as I want to help with
> many different things, I don't scale and the usual suspects aren't going
> to either indefinitely :/.

I'd propose that we hold an open IRC meeting (similar to what was discussed in 
RP's original proposal) and attempt to gauge interest/self organize.  If a once 
a month open IRC meeting makes sense, great.. if people want to do a once a week 
sync-up, etc good as well.

Somehow we need to get more communication going between the members, 
contributors, TSC and general interested people.

> Cheers,
>
> Richard
>
>
> _______________________________________________
> tsc mailing list
> tsc@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/tsc
>

Re: OE, the TSC and the future

[Phil Blundell]

On Tue, 2013-06-18 at 17:23 +0100, Richard Purdie wrote:
> In brief summary the TSC has been doing two main things, acting as a
> task force and also being able to make a decision when needed. The
> latter has not happened much at all, the main work was as a task force
> on various issues, firstly engaging with the Yocto Project and figuring
> that out, more recently dealing with infrastructure issues and generally
> ensuring the health of OE.

It's certainly true that most of the current TSC members have been doing
a fine job on the "task force" front and I think we'd all be glad to see
them continue with that.  What's rather less obvious to me is whether
they actually need to be an elected body in order to do so; any group of
individuals who wish to form a task force are obviously free and welcome
to do so at any time.

As you note, the requirement for the TSC to actually make decisions has
been minimal/nonexistent of late and, again, it's not totally obvious
that having an elected body of experts on perpetual standby just in case
a decision might be needed is entirely necessary. 

In the many-layered world that we now inhabit, it seems reasonable to
let the individual layer maintainers (in consultation with their peers
when necessary) make the decisions that they think best for their own
trees.  Recent experience seems to suggest that, practically speaking,
this is already what's happening and when it comes to a contest of wills
the TSC have not shown any interest in overruling layer maintainers who
disagree with their stated position.

Plus, of course, we already have two bodies who are empowered by the OE
e.V. statutes to make decisions, namely the board and the GA.   Both of
these have wide discretion to do what they think best, and of course
they can convene a panel of expert advisers if they feel that any
particular issue needs specialist knowledge that they don't have.

So, all in all, I feel that we've come to the point where the TSC (as an
organisation) is no longer providing us with any particular benefits and
could be disbanded without causing any real hardship.  This would avoid
the administrative overhead of running elections for each of its seats,
which have recently been uncontested in any case, and would also avoid a
certain amount of potential ambiguity over where the TSC's jurisdiction
ends and the board's starts.

p.

Re: OE, the TSC and the future

[Sean Hudson]

On 7/5/13 6:46 PM, Phil Blundell wrote:
> On Tue, 2013-06-18 at 17:23 +0100, Richard Purdie wrote:
>> In brief summary the TSC has been doing two main things, acting as a
>> task force and also being able to make a decision when needed. The
>> latter has not happened much at all, the main work was as a task force
>> on various issues, firstly engaging with the Yocto Project and figuring
>> that out, more recently dealing with infrastructure issues and generally
>> ensuring the health of OE.
> It's certainly true that most of the current TSC members have been doing
> a fine job on the "task force" front and I think we'd all be glad to see
> them continue with that.  What's rather less obvious to me is whether
> they actually need to be an elected body in order to do so; any group of
> individuals who wish to form a task force are obviously free and welcome
> to do so at any time.
>
> As you note, the requirement for the TSC to actually make decisions has
> been minimal/nonexistent of late and, again, it's not totally obvious
> that having an elected body of experts on perpetual standby just in case
> a decision might be needed is entirely necessary.
>
> In the many-layered world that we now inhabit, it seems reasonable to
> let the individual layer maintainers (in consultation with their peers
> when necessary) make the decisions that they think best for their own
> trees.  Recent experience seems to suggest that, practically speaking,
> this is already what's happening and when it comes to a contest of wills
> the TSC have not shown any interest in overruling layer maintainers who
> disagree with their stated position.
>
> Plus, of course, we already have two bodies who are empowered by the OE
> e.V. statutes to make decisions, namely the board and the GA.   Both of
> these have wide discretion to do what they think best, and of course
> they can convene a panel of expert advisers if they feel that any
> particular issue needs specialist knowledge that they don't have.
>
> So, all in all, I feel that we've come to the point where the TSC (as an
> organisation) is no longer providing us with any particular benefits and
> could be disbanded without causing any real hardship.  This would avoid
> the administrative overhead of running elections for each of its seats,
> which have recently been uncontested in any case, and would also avoid a
> certain amount of potential ambiguity over where the TSC's jurisdiction
> ends and the board's starts.
>
I believe that Phil Blundell raises a valid question about the continuing
need for the TSC. However, it seems to me that the need for a group to
arbitrate on technical matters is valuable enough, by itself, to keep 
the TSC.
This applies even if that function is utilized infrequently.

In considering the future of the TSC, I offer the opinion that in the 
future,
tactical concerns shouldn't be the primary business for the TSC. Rather, I'd
like to see the TSC become more strategically focused. In particular, 
I'd like
to have the TSC produce a vision for OE in the 2-3 year time frame. As 
another
OE board member, I support Phil Balister's offer to have the 
non-technical matters
fall to the board to handle, with the caveat that the TSC becomes more 
focused on
the long term evolution and improvement of OE as a whole

Regards,
      Sean
  
Sean Hudson | Embedded Linux Architect
Mentor Embedded™
Nucleus® | Linux® | Android™ | Services | UI | Multi-OS

Re: OE, the TSC and the future

[Paul Eggleton]

On Friday 26 July 2013 00:06:32 Phil Blundell wrote:
> On Mon, 2013-07-08 at 13:51 -0500, Sean Hudson wrote:
> > it seems to me that the need for a group to
> > arbitrate on technical matters is valuable enough, by itself, to keep
> > the TSC.
> > This applies even if that function is utilized infrequently.
> 
> I remain slightly dubious about that, not least because the TSC's recent
> track record in terms of having its decisions implemented hasn't been
> entirely stellar.

As far as recent history goes, my recollection is we've only had one decision 
(on shell function indentation) that was backed out by maintainers. I think I 
speak not just for myself when I say those on the TSC that were opposed didn't 
feel it was big enough of an issue to try to battle for the original decision 
to be upheld. That doesn't mean that we aren't prepared to resolve technical 
disputes in future, as per the original TSC mandate, nor does it mean that 
there won't be a need for that in future.

> It's also been quite noticeable that, except for Richard (who started
> this thread), none of the current members of the TSC have offered any
> opinion on what the future role of that body ought to be or what value
> it brings.

I guess I hadn't responded yet because I felt like this thread was for others 
outside the TSC to express their thoughts first.
 
> > In considering the future of the TSC, I offer the opinion that in the
> > future,
> > tactical concerns shouldn't be the primary business for the TSC. Rather,
> > I'd like to see the TSC become more strategically focused. In particular,
> > I'd like
> > to have the TSC produce a vision for OE in the 2-3 year time frame
> 
> ... this is a good suggestion; I'm sure we'd benefit from a bit more of
> a strategic vision and perhaps this is indeed the direction that the TSC
> ought to evolve in.

FWIW, I agree. I don't think we should necessarily lose sight of the TSC as a 
body for resolving technical disputes however.

Cheers,
Paul

-- 

Paul Eggleton
Intel Open Source Technology Centre