Re: You have to break eggs to make an omlette [was: LONG TERM PLAN]

[ru...@us.ibm.com]

Hans wrote:

>I'm actually very surprised that this is such a big issue. Here we have
>someone who's willing to offer a lot of time to help out in the best way
he
>knows by trying out a new architecture, on a separate branch not to
>interfere with the ongoing cleanup effort. And it get's "-1'ned". What's
wrong
>with this picture?

Let's look at how this will work in practice.  One or more people start
with essentially a clean slate and a roadmap.  Whenever possible, they
reuse as much as possible from the current implementation.  What this means
is that as much as humanly possible of the investment in time that some
people have made to date will not be lost.  So far so good.

Now think about what happens once the first line of code is copied over to
the new branch.  What do I do when I need to make a change in that line of
code?  At first, the problem won't be so bad, but think ahead a few weeks.
Now when I want to make a change, I need to make it in tomcat.current, and
then figure out if this code has been copied to tomcat.next, if so where,
and even if I find it I need to determine if the fix makes sense in the new
context.   Finally, tomcat.next may not be functioning adequately enough at
that point for me to test the change.

So what I am likely to do?  That's easy - I'm likely to sit out for the
duration.  I can't speak for others, but there is a distinct possibility
that some others may feel this way too.

So, let me speak not in terms of +1 or -1, but in terms of where I am
willing to invest my time.  I wish to invest my time in whichever path has
a largely running engine and looks like it has the best chance of
withstanding the test of time.  Whether it is a separate branch, separate
directory, or perhaps even temporarily broken in place - that's the place I
want to be.

- Sam Ruby

Re: You have to break eggs to make an omlette [was: LONG TERMPLAN]

[James Todd <jw...@pacbell.net>]

"Craig R. McClanahan" wrote:

> To say nothing of the fact that you are likely to focus on the day-to-day
> integration challenges and not pay sufficient attention to where you are trying
> to get.

well stated. a good part of the tomcat 2.2 stuff i had my mits on is
a result of meeting some pretty aggresive goals ... typical proto type
cycle.. with the initial goals (servlet 2.2 compliance for the most part,
j2ee integration and jakarta) knocked down i was *very* eager to
get in and re-design. hell, i was basically drooling ... with all sorts of
ideas i formulated over time and was eager to toy with ... till a career
change came upon the scene that is. "you get it right the third time"
(brookes) does seem to be accurate in my experience.

sometimes one needs to step to the side a bit (eg *.next) to see
the proper design/proto/play/bring back to the table in order to
amp up the momentum a notch ... for the better. perhaps communicating
about components/subsystems vs tomcat foo.bar and putting the
burden on the *next (or whatever it is called) path to track the
fixes are all that is required.

- james

Re: You have to break eggs to make an omlette [was: LONG TERMPLAN]

["Craig R. McClanahan" <cm...@mytownnet.com>]

rubys@us.ibm.com wrote:

>
> So what I am likely to do?  That's easy - I'm likely to sit out for the
> duration.  I can't speak for others, but there is a distinct possibility
> that some others may feel this way too.
>

Unfortunately, that's an argument that cuts both ways.  I'm certainly not
excited about expending the effort it would take to incrementally modify the
fundamental architecture with the baseline requirement that it always has to
work.

To say nothing of the fact that you are likely to focus on the day-to-day
integration challenges and not pay sufficient attention to where you are trying
to get.

>
> - Sam Ruby
>

Craig

PS:  Can we stop breaking eggs and making omelettes pretty soon?  Gotta watch
my cholesterol level :-).