Dave Jones wrote:
> > Then there is the issue of architectures, at least in my book KS should
> > focus on the ones that are really live and not in maintenance mode.
> > x86_64, x86_32, PPC, ia64, ARM seems to be the driving ones these days,
> > m68k, Sparc32, and others, somewhat less so .....
>
> Again, I don't recall us spending any time at all discussing m68k, or
> sparc, whilst the others you mention were well represented.
Hi Dave,
I'm not too bothered about the subjects, but rather the issue that we
keep seeing this strict "only this small group, which defines the most
important people in the community" thing. Thats where I think the
current model fails, even if someone has done a lot of work all over
Linux for years, doesn't mean said people are the ones driving things
this year.
Personally I think Andrew's suggestion is really good, turning it more
towards the traditional conference means people who have something they
want to say are more likely to push for things. If one doesn't have
something to say, then going to the KS is probably not the right thing.
> One of the problems with this approach is sometimes we don't know about
> subjects that become important to us all until the last minute, and
> others that seem important now will become moot by the time the summit comes around.
Thats true, and there should certainly be space for new subjects coming
in on short notice. However, I would suggest that at least a significant
portion of the summit applies this requirement. Most of the more
important issues are architectural and it's often not something that
shows up last minute.
> So far though, there's been nothing proposed at all, so feel free
> to throw your hat in the ring, if nothing else, it'll kickstart
> the process.
Actually I'm in the process of investigating launching a mini summit
cabal, which I think would cover most of my current issues :)
Cheers,
Jes
On Tue, 2007-01-30 at 05:51 +0100, Jes Sorensen wrote:
> > So far though, there's been nothing proposed at all, so feel free
> > to throw your hat in the ring, if nothing else, it'll kickstart
> > the process.
>
> Actually I'm in the process of investigating launching a mini summit
> cabal, which I think would cover most of my current issues :)
Actually, perhaps we should track these more closely. At the moment we
have mini summits in
Networking
Wireless
Filesystems
Storage
Power Management
And probably several others I can't remember. Right at the moment, the
organisation and funding for all of these is completely ad-hoc, so if
mini summits are the way to go, it would certainly be better to move
them on to a more templated basis (so anyone wishing to organise one
would know whom to go to for these things).
James
On Tue, Jan 30, 2007 at 05:51:00AM +0100, Jes Sorensen wrote:
> I'm not too bothered about the subjects, but rather the issue that we
> keep seeing this strict "only this small group, which defines the most
> important people in the community" thing.
I don't think it's intentionally meant to come across that way.
Not being invited to kernel-summit shouldn't be interpreted as
"you're not good enough" in any way or form.
> Thats where I think the
> current model fails, even if someone has done a lot of work all over
> Linux for years, doesn't mean said people are the ones driving things
> this year.
Right. I see your point, and attendance shouldn't be solely down to
a "what have you done for me lately?" decision, which is why there are
additional criteria. That still doesn't make the process perfect,
but we're open to good solutions to solve the problem of trying to
pick 80 or so people out of the hundreds of developers that make
the first pass.
When Jon posted how the selection process worked last year a few people
(yourself included iirc) brought up concerns, but it seems no-one
has any real answers on how to improve things beyond the status quo.
It hasn't gotten easier by us shrinking in size slightly each year too.
This has both positive and negative points. Yes, more people are going
to get left out, but there's a point where so many voices in a room
just becomes uncontrollable, especially when it's a room full of
people with strong opinions. A number of people mentioned last year
that the level of interaction during the sessions seemed higher than
ever, with less people staring at laptops, and actually getting involved
in what was happening in the room. I strongly believe that the lower
head count was responsible for this.
The one solution (well, in part) to the lower headcount last year was
the addition of the mini-summits. If we had invited all the power management
guys, all the networking guys, all the wireless guys etc etc we would
probably have doubled in size. In future I wouldn't be surprised if
these specialised summits happen more often.
Perhaps one day even negating the need for kernel summit at all
(unless it becomes two days of wrap ups and cpu architect roadmaps),
well, maybe not, but hopefully it'll help at least partially address
the concerns of developers who didn't get to be at the kernel summit.
> Personally I think Andrew's suggestion is really good, turning it more
> towards the traditional conference means people who have something they
> want to say are more likely to push for things.
It may indeed have merit.
Dave
--
http://www.codemonkey.org.uk
Dave Jones wrote:
> On Tue, Jan 30, 2007 at 05:51:00AM +0100, Jes Sorensen wrote:
> > I'm not too bothered about the subjects, but rather the issue that we
> > keep seeing this strict "only this small group, which defines the most
> > important people in the community" thing.
>
> I don't think it's intentionally meant to come across that way.
> Not being invited to kernel-summit shouldn't be interpreted as
> "you're not good enough" in any way or form.
True, but unfortunately the KS has gotten itself a real bad reputation
for being a closed club of the same people meeting year after year.
If this is warranted or not is open to discussion, but at least thats
the general message I get when I talk to people and the subject of KS is
brought up.
> When Jon posted how the selection process worked last year a few people
> (yourself included iirc) brought up concerns, but it seems no-one
> has any real answers on how to improve things beyond the status quo.
In this case I think Andrew's suggestion of trying to twist it more
towards the traditional conference style would be worth investigating.
The other issue here is that at least historically it has felt a bit
like pounding sand when anyone trying to state that the summit wasn't
working too well as it has been operating the last couple of years.
> It hasn't gotten easier by us shrinking in size slightly each year too.
> This has both positive and negative points. Yes, more people are going
> to get left out, but there's a point where so many voices in a room
> just becomes uncontrollable, especially when it's a room full of
> people with strong opinions. A number of people mentioned last year
> that the level of interaction during the sessions seemed higher than
> ever, with less people staring at laptops, and actually getting involved
> in what was happening in the room. I strongly believe that the lower
> head count was responsible for this.
Well laptops are a problem, but I think some of this can be addressed
mostly at the on-site level. The other problem is often that people are
not interested or prepared for a given subject and therefore ignore it.
I think the requirement of having an abstract submitted in advance could
help here too.
> The one solution (well, in part) to the lower headcount last year was
> the addition of the mini-summits. If we had invited all the power management
> guys, all the networking guys, all the wireless guys etc etc we would
> probably have doubled in size. In future I wouldn't be surprised if
> these specialised summits happen more often.
Even the mini summits have the problem of being selective and some
projects are more likely to be included than others. For some projects
it's a lot more clear that there's a specific lead on it, whereas
others, such as file systems it's many very different projects in
parallel with different requirements.
That said, I think using the KS as more of an overall architecture
handling summit and leaving more specifics to the mini summits is a
good way to go.
Cheers,
Jes
On Mon, Jan 29, 2007 at 11:11:54PM -0600, James Bottomley wrote:
> Networking
> Wireless
> Filesystems
> Storage
> Power Management
>
> And probably several others I can't remember. Right at the moment, the
> organisation and funding for all of these is completely ad-hoc, so if
> mini summits are the way to go, it would certainly be better to move
> them on to a more templated basis (so anyone wishing to organise one
> would know whom to go to for these things).
One thing that might have made these mini-summits so successull might
have been the ad-hoc setup without much corporate or organizational
involvement. I'm looking forward to see if we can keep this spirit
despite the Usenix involvement for the next FS/Storage summit.
On Mon, Jan 29, 2007 at 11:11:54PM -0600, James Bottomley wrote:
> And probably several others I can't remember. Right at the moment, the
> organisation and funding for all of these is completely ad-hoc, so if
> mini summits are the way to go, it would certainly be better to move
> them on to a more templated basis (so anyone wishing to organise one
> would know whom to go to for these things).
Well, Usenix has offerred to provide logistical support for some
mini-summits if anyoen wants to take them up on it. Using some of the
sponsorship money from last year, we've proposed to make some hotel
conference rooms right before OLS available if anyone wants to do a
10-30 person mini-summit in Ottawa.
Is there any interest?
- Ted
On Tue, 2007-01-30 at 08:30 -0500, Theodore Tso wrote:
> On Mon, Jan 29, 2007 at 11:11:54PM -0600, James Bottomley wrote:
> > And probably several others I can't remember. Right at the moment, the
> > organisation and funding for all of these is completely ad-hoc, so if
> > mini summits are the way to go, it would certainly be better to move
> > them on to a more templated basis (so anyone wishing to organise one
> > would know whom to go to for these things).
>
> Well, Usenix has offerred to provide logistical support for some
> mini-summits if anyoen wants to take them up on it. Using some of the
> sponsorship money from last year, we've proposed to make some hotel
> conference rooms right before OLS available if anyone wants to do a
> 10-30 person mini-summit in Ottawa.
>
> Is there any interest?
Martin was looking into organising the VM summit thereabouts.
On Tue, Jan 30, 2007 at 08:30:00AM -0500, Theodore Tso wrote:
> On Mon, Jan 29, 2007 at 11:11:54PM -0600, James Bottomley wrote:
> > And probably several others I can't remember. Right at the moment, the
> > organisation and funding for all of these is completely ad-hoc, so if
> > mini summits are the way to go, it would certainly be better to move
> > them on to a more templated basis (so anyone wishing to organise one
> > would know whom to go to for these things).
>
> Well, Usenix has offerred to provide logistical support for some
> mini-summits if anyoen wants to take them up on it. Using some of the
> sponsorship money from last year, we've proposed to make some hotel
> conference rooms right before OLS available if anyone wants to do a
> 10-30 person mini-summit in Ottawa.
>
> Is there any interest?
>
I think there's enough of relevance for an embedded mini-summit this
year, particularly as it's not clear that there's going to be a power
management summit this year (though of course there are many other
topics to be looked at, too). Does that count?
On Tuesday 30 January 2007 08:30, Theodore Tso wrote:
> Well, Usenix has offerred to provide logistical support for some
> mini-summits if anyoen wants to take them up on it. Using some of the
> sponsorship money from last year, we've proposed to make some hotel
> conference rooms right before OLS available if anyone wants to do a
> 10-30 person mini-summit in Ottawa.
>
> Is there any interest?
Yes, suspect that a day attached to OLS may make a good power-management summit day.
-Len