1.7 planning: Collecting Projects to Estimate

Sam Hartman hartmans at MIT.EDU
Mon Jan 22 14:14:10 EST 2007

Hi, folks.  MIT is starting the planning process for the 1.7 release
and we would appreciate your input.  The process is going to be
divided into three phases.  The first phase, which will last two weeks
starting today will focus on identifying potential projects of use to
the developer community and our customers.  During the second phase
MIT will pick some of these projects to explore in more detail.  We'll
produce enough of a design to validate our estimate of how much of a
resource commitment the project would be.
During the third phase we'll pick which projects MIT will dedicate resources to.
Naturally we encourage  other parties to work on enhancements to Kerberos and to contribute patches back to MIT.

The last year has seen a lot of contributions back to MIT.  That's
been great, but it's also had a price: we've been so busy auditing
code, getting it to work with our coding style and generally getting
it integrated that we haven't had much of a chance to work on the code
ourselves.  We're going to approach this problem in three ways.
First, we're going to flesh out some style guidelines for external
contributions.  Second, we will continue to build partnerships so that
people we've worked with before can do some or all of their own
integration work.  However we're also going to have to limit the
resources we dedicate to integrating external code.

So, here's how the next two weeks are going to work.  We encourage
people to provide brief descriptions of projects that they believe
should be included in 1.7.  The project needs to be scoped so that it
can be completed within 1 to 1.5 years given all the other things
going on.  The project description needs to be sufficiently detailed
that we could take a potential design and ask whether implementing
that design would meet the project description.  However the project
description should be brief.  "Clean up the build system," isn't a
good description because it's not detailed enough that we would all
agree on whether we were successful.  "Implement IPV6 for kadmind" is
about right.

We specifically do not want discussions of the technical details of
projects at this point nor even discussions of how to design the
project/whether it is implementable.  Clarifying questions are fine.
Comments that a project would be useful to a particular user community
would be greatly appreciated.  Specific offers to dedicate resources
to a project would be most welcome.

In order to give you an idea of what we're looking for, I will seed
the discussion with a set of projects we came up with internally.


More information about the krbdev mailing list