Re: [9fans] Publish and be damned.

[Lucio De Re <lucio@proxima.alt.za>]

On Sat, Apr 21, 2001 at 02:08:42PM +0100, Steve Kilbane wrote:
>
> Boyd's talking about each original author maintaining a package. Lucio's
> talking about an amorphous mass of people hacking on several packages,
> at whim. Boyd's starting point is simpler, because it's present-day.
> Lucio's reason for changing things is so that a package doesn't stagnate
> because the original author lacks time or inclination.
>
I sure as hell didn't express my views this well.  Thank you, Steve.

> Boyd suggests reading the code. That's fine if the code's good. If Lucio's
> suggestion takes off, the code will be less predictable, and having a
> change comment is more important, because the reason for the change may not
> be deducible from the change itself.
>
The granularity of changes, not only the comments, would be different.
I may work on a single issue for days, across a number of modules,
and only "check in" the result.  That in itself adds information,
if not documentation.  By the same token, I may make numerous
unrelated changes between dumps, which would require effort to
investigate separately.  And reversals (undo/redo) would also be
lost.

Within Bell Labs, I would assume that daily is a coincidentally
happy medium.  I could also believe that development there would
have adjusted to the daily rhythm, otherwise the interval would be
different.  But conditions outside Bell Labs are unlikely to be as
fortunate.

> But either way, I don't see any conflicts here. Russ's comments on avoiding
> conflicts with BL's interests aside, I don't see a problem with Lucio setting
> up the server. Those who like the idea will join in; those who don't, won't.
> Whether it'll produce anything better than what would happen otherwise,
> I don't know.
>
I hope I didn't miss an important comment from Russ, if there are
conflicts with Bell Labs, I would want to be amongst the first to
know.

My prediction is that success really lies with Bell Labs buy-in
(I'm very fond of that term, for my own philosophical reasons),
and some dedication by one or more project facilitators.  I would
assume that anything "published" on the CVS repository is as readily
available to Bell Labs as if it was published in any other form,
in terms of the licence.  Any other restriction or relaxation may
have to be discussed and agreed upon by a multitude of parties,
not a rosy option.

But volume of publication (shudder, what do I do when resources
start becoming tight?) may be the most significant factor.  If
there is enough out there to bring Plan 9 closer to the desires of
prospective users, then it is possible to reach critical mass.

Those of us who may think of Plan 9 as the True Religion may either
be estatic or horrified at the thought.  I, for one, am not sure
which camp I'm in, but evangelism at least is not up my alley.

++L