Subject: Re: small worlds and better than ransom
From: "Stephen J. Turnbull" <stephen@xemacs.org>
Date: Mon, 08 Oct 2007 08:40:52 +0900

Thomas Lord writes:

 > You're wrong.   The way that Apache is hooked out the butt (has lots
 > and lots of hookness to it - like plugins, config foo, etc) --- it was
 > built, from the ground up as a switch.   LAMPishness was precisely
 > what they had in mind.   Anyone who says different is lying.

Can you document that?  It would be an interesting read.

But everything I've read so far is that (1) the LAMP stack was an
emergent phenomenon, nobody was aiming for it specifically but it
evolved simultaneously in several communities, and (2) the early
development was aimed at flexibility of the webserver per se, and it
was only ex post "LAMP" that people realized that generic flexibility
wasn't needed, rather a platform to support middleware flexibly was
the need.  It was only once the category of middleware was understood
that you could talk about LAMPishness.  No?

 > >  they just wanted a reliable webserver. 
 > 
 > Itself a category that intentionally aimed for LAMPishness.

Say what you like, there are plenty of uses for reliability other than
as part of a LAMP stack.  That's not why I wanted a reliable webserver
in the mid-90s, and was so pleased to switch over from NCSA to Apache
in 1998 or so.

Re: Canonicalization of Arch

 > They had, afaict from that time, they wanted to divert as many
 > debian patch submitters to being Ubuntu submitters as possible.
 > The same damn zero-sum thinking.

I'm sorry to break it to you, but all patch submitters are damned to
zero-sum-ness in the use of their time.  From a business point of
view, there are several reasons why you want patches coming to you,
rather than another organization.  Some are monopolistic, but others
are purely cost reduction (for yourself only and at expense of others,
which is selfish, but not monopolistic).

I seem to recall that you always thought it was a good idea to have
changes to the Arch protocol go through you, for example.  A monopoly
in the making?  Or did you have nobler motives that may have applied
to Canonical, too (in part, anyway)?

 > Arch seemed to me to be appreciated in light of its potential
 > to help change where people "submit" their patches.

You mean by Ubuntu?

I seem to recall seeing that the Canonical people carried over the
practice of posting archive URLs from the Arch culture, so that
integrators "pulled" patches rather than having random hackers "push"
them.  Attach an RSS feed to the patch submission channel, and anybody
can piggyback on that.  I see a commitment to rather selfish cost
reduction there, but use of Arch in that way is a clear indication
that if they're trying to be monopolistic, they're not trying very
hard.

Of course Debian has been fairly locked in to CVS, but it's not that
expensive in human time to integrate in a dVCS and push from there to
CVS.  So Debian developers could use the Ubuntu patch flow to a great
extent, I suppose, at a significant startup cost and nominal flow
cost.  Still, it would be quite possible, and need not kill Debian.

 > > I suppose the bazaar crew had a lot to do with the current state of
 > > the GNU Arch project, but it was more a matter of directing their
 > > efforts elsewhere, an elsewhere that directly competed with GNU Arch
 > > in a lot of ways.  Not trying to kill the GNU Arch project.
 > 
 > I strongly disagree but this isn't a courtroom so let's leave it at 
 > that, please?

OK.  Please note that I'm not interested in a verdict in this case,
but rather making the point that recruiting resources from
organizations making similar products is S.O.P. in business.  And in
free software, for that matter, but this isn't a courtroom....

 > > Surely you can't expect *me* to have anything but sympathy for an
 > > open-source-based commercial project with short-term business goals
 > > that led to a bitter fork because the upstream leader wasn't taking
 > > the upstream project in the direction of satisfying their perceived
 > > needs any time soon.

 > Actually, I can expect otherwise of you.  Keyword there is
 > "bitter."  That was needless and helped nobody.

I gather you don't understand the import of the Emacs fork as my
example here.

Bitterness is *always* needless, in some sense.  Even when I myself
participate in it.  :-(  If only we were programmed to obey the Three
Laws or even just the Golden Rule!  I have sympathy for *both* sides
as far as the "needless bitterness" goes, in fact, for all of humanity
which spend so much time, effort, and blood in the practice of
needless bitterness.