On Fri, 12 Aug 2005, Jim Grisanzio wrote:

> We have four community proposals on this list now:
>
>    * Solaris x86 drivers
>    * Chinese users
>    * Approachability
>    * Laptops
>
> I'm clear on the Approachability proposal, the feedback was good, so
> we're opening it. I'm not as clear on the Chinese users community, but

Jim - someone already said that if the proposed community got 3 CAB member
votes, then it was a go.  So far, I think that the Approachability proposal
only got my (single) CAB vote.  In fact, there were very few +1 votes
associated with it, despite a 2nd email from Stephen Hahn saying something
like "with all these +1 votes flying around can we please get some more
votes for Approachability".

I'm very much in favor of this community.  It turns out the JBeck told me
(while riding on the trolly/train in Portland) that Approachability is his
"day job".  I would be very pleased to see JBeck (a man of many talents)
and Stephen Hahn working in this area.  The benefits to the (Open)Solaris
community can only be enormous when you apply this level of talent.

So if you give this community a "go" then so be it!

> I'm ok with opening that community as well. I think the original thread
> got split into two lists and it's difficult to follow. It seems focused
> on education and marketing for China, and I didn't see anyone object.
>
> On Laptops and Solaris x86 drivers, I see some ways of combining those
> two (and potentially more) into one community. I'm thinking about the

The laptop community got more +1 votes than anything else that was ever
proposed to date.  There's a *huge* interest within the (Open)Solaris
community in running Solaris on a laptop.  With the current generation of
laptops running 64-bit processors and capable of 100Gb+ disks and 2Gb+ of
RAM, a better way to describe the interest in laptops, would be to describe
them as "mobile workstations" or "mobile development workstations".  Or
another way to look at the mobile workstation would be to describe it as
Suns last chance to build a developer workstation - since they _almost_
missed the entire market opportunity for (developer) desktop workstations
based on commodity hardware.!

Along with laptops is the *huge* interest in connecting "mobile
workstations" via wireless networks.  Currently that translates to using
wireless G with Caspers WiFi drivers (if you're lucky enough to have access
to them) - which make that combination very, very useful in practical
day-to-day situations.  Anything from demos to doing "real" work
on-the-road.  You only have to watch Keith Wesolowski and Bryan Cantrill
give mdb and DTrace demos on a couple of Acer Ferraris to see how
incredibly usefull and powerfull a "mobile workstation" is, in the real
world.

Now the whole "mobile workstations" market segment will take on a new, much
higher, much more visible market presence as we witness the arrival (RSN -
Real Soon Now) of:

a) dual core mobile CPUs  aka dual-cores for laptops
b) WiMax
c) solid state disk drives for laptops (Samsung)
d) 7,200 RPM disk drives (most are 5,400 RPM now)
e) more 100Gb+ mobile disk drive options
f) 4Gb+ USB solid state disks for:
  - booting an Operating System
  - backup of data/code/work
  - storing encrypted data that won't be compromised if the laptop is
stolen.
g) Virtualization: being able to run multiple Operating Systems on a mobile
platform.

I would suggest/recommend that since laptops require unique drivers, that
we not split drivers away from laptops.  Ques: What is it about laptop
drivers that make then unique?  Answers:

a) they must be written to conserve power when they are inactive
b) they will probably be many vendor specific quirks - as the laptops
manufacturers go crazy trying to distinguish their product from the
competition in a marketplace where most products are using commodity
motherboards and chipsets.  So, when a vendor "adds value", in terms of a
laptop, that will invariably require a customized driver or a driver
"feature" to exploit it.
c) Most laptop drivers will be sufficiently different that they won't apply
to desktops
d) desktop and laptop chipsets will probably diverge even further over
time.
e) many laptop drivers will need to be made power aware.  This is distinct
from a driver that needs to reduce power and heat generation while inactive
or while running more slowly during periods of reduced user activity.  The
driver will need to "know" what to do when the battery is about to die.  Or
how to behave differently (i.e. go fast) when the laptop is connected to
its A/C power supply.
f) many drivers will be laptop specific. For example, a battery monitor
driver or a CPU temp monitoring driver.
g) drivers will have to "play nice" with other system components whos
behavior changes dynamically in response to system utilization or available
power or system component temperature.

> user group community I formed, which now has about 10 individual groups
> and three more in the queue. So, one community with 10 groups, not 10
> individual communities. The effect is similar, but it's much easier to
> find user groups if you simply look in the one meta user group
> community. Each user group has its own mail list, and we use a meta
> ug-discuss list to communicate among groups. In theory, anyway. We're
> still getting going. A better example, probably, is the JDS and KDE
> experience. We had a proposal for a JDS community, but after some good
> discussion the suggestion was made to open one Desktop community with
> JDS and KDE. That seems to have worked out pretty well, too. Could we
> consider the same type of arrangement with Laptops and Solaris x86
> drivers? Perhaps there are even more elements that would fit within that
> combined community?
>
> To clarify the "community proposal" process we are doing here: when
> people started asking for new communities after the launch, I suggested
> that we have a simple process where a community member would post to the
> discuss list some information about the proposed community, it's goals,
> scope, participants, etc. We didn't have a formal governance process in
> place, so a quick public proposal of a new community seemed reasonable.
> It would serve to give the community notice that this was going on, it
> would encourage debate about the community so perhaps others would want
> to get involved, and it could also serve as a way to make new
> connections that would either expand or better focus the new community.
> If no one really objects, the community would be opened and hopefully
> the community leaders would consider some of the community feedback.
> That's pretty much it.
>
> Of course, when the governance is ratified that document will drive this
> process as well as clarify the definition of what a community is.

Agreed.  In the meantime - thanks for taking up the slack.

PS: Casper just started his (well deserved) vacation.  So we may not get to
have his valuable feedback on the laptop related drivers.

Regards,

Al Hopper  Logical Approach Inc, Plano, TX.  [EMAIL PROTECTED]
           Voice: 972.379.2133 Fax: 972.379.2134
OpenSolaris Community Advisory Board (CAB) Member - Apr 2005
_______________________________________________
opensolaris-discuss mailing list
[email protected]

Reply via email to