[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Gnu-arch-users] re: sad events
From: |
Thomas Lord |
Subject: |
[Gnu-arch-users] re: sad events |
Date: |
Sat, 08 Oct 2005 16:01:05 -0700 |
Michael Goikhman writes:
> The latest events are sad. It seems that neither Tom Lord nor
> Canonical are interested in developing Arch anymore. And noone yet
> volunteered as the new maintainer.
Let me expand on, perhaps even change my position about that.
Negatives, and then positives.
The negatives:
~ I am not interested in continuing as an official GNU Maintainer
I have too many political differences with RMS about the proper
handling of the GNU project and so I do not think I can represent
the project's values and policies well enough to be a maintainer.
*Of course* I agree about the importance of software freedoms and
will do my best to always chose GPL-compatible free software licenses
but those points of agreement are insufficient, in my opinion, to
also stand as a public interface to an official GNU project.
~ I am not interested in running a "fashionable" public free software
project.
I've taken no end of heat from vocal members of the Arch mailing
lists about such things as whose patches and bug reports I spend
time on and when. Often the implicit threat, and actual effect,
has been "Do our bidding or we trash your professional reputation."
The sad truth is that with the resources I've had, to keep such a
user community satisfied would mean making huge sacrifices to the
code quality of GNU Arch and to any work on making major leaps in
functionality.
There is room, certainly, to enable skilled and dedicated volunteers
and promising, eager-to-learn volunteers to help hack a public
project. There is value, certainly, in being able to receive bug
reports. But the wild-west gambler's-town style interaction I
encountered in this project doesn't fit the bill. A better, more
restrained approach is needed -- one which is less dependent upon
the way the winds blow on the mailing lists from week to week.
The In Between:
~ I am not interested in dying.
The same old sob story: I've no revenues. I'm in the red for october
(meaning that basic expenses like food and DSL and electricity through
the end of the month vastly exceed my remaining cash. I'm flat,
flat, flat broke in just a few days with nothing redeeming in sight).
I'm not interested in "working for fame" or any other such intangible.
I did that. I earned my own modest share of it. That and a $1.50
will buy me a cup of coffee (though I'll have to stiff the barista
on a tip).
The positives:
~ Especially with Canonical abandoning it, Arch can have a future.
Arch is no longer my heart's first true love -- that's for sure.
On the other hand, and especially in the direction of revc, I think
it can have a bright future.
I am willing to enthusiastically keep working on it under the right
circumstances, provided that I am able to go in directions that I
think make sense, especially directions that move far, far beyond
just "software source code revision control". Surely I have earned
some credibility with which to chose a course.
I think that porting some of the missing features from tla to revc,
which would be quite useful for source code management, is a good
place to start. This can be done without sacrificing the already
file-at-a-time CVS/SVN-like command set of revc or the nice
git-like performance characteristics.
I would certainly be willing to be technical adviser to a group
that wants to maintain tla as it stands (although they have to
pay more than lip service to my role in that capacity).
It is simplifying that Canonical is going in their own direction.
It would have been just swell to have help improving the UI in
this direction or that but, for whatever reason, they came with
far too much baggage.
~ All it takes is monetary investment
So, Arch got started (out of necessity) by my going into debt.
It got sustained by the stunning generosity of many people when
I was busking. It got sustained further while an angel investor
and I tried unsuccessfully to bootstrap a start-up business around
it. There seem to be a significant number of interested users.
It has certainly had its share of positive impact both directly
and by influencing several other systems.
Somehow, our collective mode of economic organization has failed to
actually reward *me* and allow *me* to, in good conscious and fully
living up to familial responsibilities, continue to work on it.
I reject the glib defamations of the Canonical crew and those they've
influenced. I don't even begin to claim perfection as businessperson
or entrepreneur but I'm certain I'm far from the outlier I've come to
be portrayed and widely regarded as thanks to their "efforts".
So I'm throwing up my arms a bit: Dear Lazyweb -- what the *hell* am
I supposed to do at this point to keep the effort going? Anyone got
any ideas? Any offers? Selfish and self-righteous replies from
Canonical, their groupies, and think-alikes will be called out as
exactly that.
~ A Vast Frontier Opens Up Before Us
Take revision control far enough and you wind up smack dab in the
middle of global file-systems generally. There is *so* much to
do that is worth doing in elaborating this framework for other kinds
of content management and with respect to new kinds of user
application that build upon it.
This weekend I've been working on raising the bar of wiki parsers.
I don't think I'll have time to finish it before I run out of cash.
Roughly, the idea is that Wikis, deep down, are based on multi-pass,
recursive parsing -- a rough parse of a large piece of text, followed
by a finer parse of each component. At each level of recursion,
the lexical and syntactic grammars change (a trivial example: wiki
mark-up that let's you make a stylized paragraph at one level --
but have a mathematical equation that can be fed to a compute-engine
and/or fancy formatter at the next level down). It ties in nicely
with revision control because wiki-style syntaxes tend to branch
and merge nicely, and applications which handle them tend to be
error-tolerant and provide user-friendly-recoverability for subtexts
containing syntax errors (such as after a slightly-off merge).
The rough idea there is to bump awiki to the next level as a precursor
for integrating it with revc. Many nifty applications flow from that.
-t
Emergency assistance welcome: address@hidden on paypal, same address for
inquiries into what directions I'd like to take my R&D efforts in
the future.
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- [Gnu-arch-users] re: sad events,
Thomas Lord <=