[cgl_discussion] PoC concall 10/8/03 Minutes

Rusty Lynch rusty at linux.co.intel.com
Tue Oct 14 17:32:15 PDT 2003


On Tue, 2003-10-14 at 17:09, Venkata Jagana wrote:
> >The timing came from Rusty who suggested end of November for dev
> >complete.
> 
> Not sure what's the basis for this target date? As I said earlier,
> based on our current plans, the kernel pieces would likely be
> available (target) by end of Dec and I believe (according to Xou
> Yixiong from Intel) the corresponding user pieces should be available
> by end of Dec too. Shirley from IBM and Xou are closley working
> together on this.
> 

Them mailing list seems to be holding email, making it hard to discuss
this.

I'm guessing that by the time you see this that you will already have my
previous message where I admit that I messed up the dates.  If not
then... yep, end of year is the target (all I do is walk over to
Yixiong's cube and ask him... you would think I could get that right).

> >We generally add 3-6 months of working with the community before the
> >software is generally available.
> 
> Sure, this model works for some projects but not for this kernel piece
> atleast -:)

We just need some kind of default buffer time.  Of course you can't
schedule an exact date for a feature to be included in any open source
project that you do not run, but we need something.

If we know for a fact the feature is problematic then we swag a bigger
buffer.  Do you think half a year is not realistic?  What would be
better... a year, a decade, a long weekend?   

I guess this is a more general problem of how do you make plans when
some of the steps in your plan are mostly out of your control.  The best
idea I can come up with for the kernel is to plan on working with LKML
using the process defined in Documentation/SubmittingPatches (.i.e send
patches, fix problems, send new patches, repeat...) and then from a
program management view add some rough guess one how long the process
should take.  Some times it just takes a day or two, other times it
never makes it in.

    --rustyl

> 
> There is a continued effort of making the kernel IPv6 mib support
> available for acceptance into the mainline as the individual MIB object
> group support is completed rather than waiting until the entire MIB
> development is complete. In fact, Shirley already submitted few kernel
> patches for acceptance and the review discussions are ongoing.
> 
> Thanks,
> Venkat
> 






More information about the cgl_discussion mailing list