← Back to team overview

multi-touch-dev team mailing list archive

Re: Strategy/Direction for Multi-touch

 

On Mon, Mar 15, 2010 at 05:29:11PM -0700, Rick Spencer wrote:
> Please keep in mind that we ship Beta 1 this week! We are at the end of
> the product cycle and need to stay very focused to keep a project like
> this from ballooning in scope and then following over. The more focused
> the scope, the more likely the success.

Right.  Now that we are so late in the cycle any impact from adding
these drivers to lucid will have to be carefully examined.  We should
note that the drivers only hit the mainline tree in v2.6.34-rc1 and as
such the stability of the code base is as yet unknown.  We cirtainly do
not want be to updating them weekly through release.

> On Mon, 2010-03-15 at 18:16 -0600, Duncan M. McGreggor wrote:
> > Note: moving conversation to mail list.
> > 
> > Rick and I chatted about scope on IRC, and he provided some excellent  
> > perspective. To restate the goal of the Lucid work:
> > 
> > Provide MT support in the kernel so that developers can start building  
> > exciting tactile apps for Ubuntu ASAP.
> And focused just on Ntrig. Other hardware may be enabled, but ntrig
> should be the focus of testing and other work.
> 
> > 
> > As such, here's what I'd like to propose:
> > 
> > Phase 1:
> >   * Lucid-targeted
> >   * use the current blueprint
> >   * focus on getting kernel work completed for Lucid
> >   * we want to do the kernel bits right
> >   * this may mean taking option 1) below for the user-space bits, but  
> > putting this work in a PPA, not in Lucid
> There should be no expectation that anything *but* kernel support will
> be delivered in lucid. A PPA delivered before, during, or after lucid
> ships is acceptable for these purposes. Of course I assume there will be
> a lot of work to package bits so that the kernel bit can be tested.

The current plan for the kernel bits first delivers them in a PPA for
testing.  Those are then tested in combination with Bryces bits.  If
they test well and the review of the patches has them low risk for inclusion
it then aims to pull them into the main Lucid kernel before release,
otherwise they remain in PPA until stable.

The first kernel builds with these patches are now available in one
of kernel PPAs.  Bryce, I think we were going to have a common place for
all 'multitouch' packages which these can be copied to once thats
identifified.

If anyone actually has a way to test these kernel bits other than boot
them then I would appreciate any feedback:

   https://edge.launchpad.net/~apw/+archive/orange/+packages

> > Phase 2:
> >   * Lucid+1-targeted
> >   * create a new blueprint
> >   * move non-Lucid tasks from Phase 1 blueprint to Phase 2 blueprint
> >   * take the usable user-space bits from Phase 1 and follow the  
> > development plan of option 2) below.
> > 
> > The understanding in this approach would be that we will be throwing  
> > away some of the work done now in order to get the MT kernel code  
> > ready for Lucid.
> 
> What work would be thrown away?

-apw



Follow ups

References