Subject: Re: Of Mips and Devs [Was: Monthly Gentoo Council Reminder for January]
Newsgroups: gmane.linux.gentoo.devel
Date: 2008-01-10 03:10:39 GMT (1 year, 25 weeks, 1 day, 7 hours and 20 minutes ago)
Well, I guess it's something that's been needing to be faced for some time now,
as difficult as it is to do. Regardless of the accusations and
counter-accusations flying around in this thread, I'll just go ahead and state
the fact that yes, we are a "slacker arch".
Why? Because there's just no time anymore these days and no one left really of
the original team. And a lot of that really is my fault. Tuxus may have laid
the first keel of our ship, but I was the one who, so long ago, made her
seaworthy and crewed her. But now, she's largely a ghost ship -- adrift in the
seas, and a hazard to the other ships.
And for that, I apologize to all of you.
My availability lately has been diminished as I've started wearing more "hats"
at work, and I find myself with barely five hours a night of free time to
meander about in whatever entertains me in my free time. Occasionally, I pull
up the window that connects to my Octane and keyword a bunch of things, and
knock out some bugs. But I'll be honest, that's been a rare thing these days.
So many of you by now are probably thinking "Ah hell, he's jumping ship on us".
No, not yet. I have to fix what I created before I even begin to ponder those
thoughts. To just up and leave with the mess I've allowed to occur would be
unfair, and frankly, just not something that's in my character.
So how can this ship be righted? I did a quick scan of most of the mails in the
thread to get an idea of some of the existing opinions (while trying to pass
over the arguments), and here's what I found that needed to be addressed.
1. It's been suggested that mips drop all stable keywords ('mips') leaving
unstable keywords as-is ('~mips').
Contrary to whatever damage and/or impact this may create, I think this is a
good idea for us. I've always ran ~arch on my Octane, and with a few
exceptions, have found ~arch to be pretty usable. Does something sometimes
break? Sure, but when one looks at all the wacky crap that exists in our arch,
sometimes you're left wondering how it all manages to work anyways.
Besides, we've always been a more experimental arch to begin with. Usually
we've been the first to try some new hair-brained idea (like automated netboot
builds via catalyst or running the most bleeding edge glibc), so this would just
be another item in our tumultuous history to take a swing at.
That said, however, I don't think it would be appropriate to commit a patch to
portage that wipes out all our stable keywords in one go. I think it would be
more appropriate to phase such a change in gently, because as far as I know, no
one else has really done this. The other archs typically maintain a
stable/unstable set of keywords in the tree. So I think this should be managed
by the profiles. I've been needing to do some profile cleanup anyways, so I can
probably fiddle with a 2008.0-dev profile set to only do ~arch, and then see how
that goes.
Thus as 2007.0 and 2007.1-dev die off, so does our stable keywords. This frees
up other package maintainers to not have to worry about one of our "pesky"
keywords holding things back, and should give us more freedom to move at our own
pace, relative to those who have free time and those who don't.
2. Many have wandered if we as a team are still alive.
And the answer to that really is a resounding "No". Individually, me and
Redhatter are probably the only ones who still do anything (and Redhatter,
thanks for all the work you've done keeping things alive). The rest had other
priorities come up in their lives that ultimately required them to resign or
fade into some un-indexed inode someplace. And it's my fault for not replacing
those lost team members with new folks.
I've got a guy in mentoring right now, but even that's been really slow as both
of us have found time to be a scarce thing. But I'd like to get some kind of a
"team" back together, and hopefully get them on the right track to run things so
that I can step off the platform as "Lead" and take more of a backseat role,
which I feel is something better suited for me anyways. It's not like I've lost
faith in Gentoo or found I have zero time, it's just that I don't have enough
time to operate effectively in the role of a Team Lead anymore. Right now, I
fit in better as "That crazy old guy who lives in a cave", or something equivalent.
But I can't do that till I get things back in shape, and so, I'll need help from
the rest of you guys. I need people to step up who want a chance to play with
an arch that's about as insane as a pikachu slamming PCP. People who don't mind
running machines that are heavy, sometimes noisy, usually slow, and weren't
exactly designed with "energy conservation" in mind. People who want to help
bring us closer to the rest of Gentoo, rather than off in our own little world,
as we are oft found to be.
I'll fill those interested in all the dirty information they need to hunt down a
machine we support (but to make it easy, just acquire an Octane), so e-mail me
and ask whatever questions you need to know.
3. Should Gentoo even continue to support mips? Do people even *use* mips?
My opinion is that yes, we should continue to support mips. Ultimately, I'll
leave that decision to the higher authority, but I think if a new team can be
assembled, and I can be allowed to step aside to more of an advisory role, that
mips can function normally again. And maybe, regain some of the respect we've
lost over the years for various reasons.
As for whether we even have users, I can say affirmatively that we do. Not many
for sure, like say Sparc or PPC, but we do. Part of the problem with this is
our area of focus. The Mips Team really only focuses on SGI Hardware, because
this hardware is readily available on eBay, and usually at good prices. Mips as
an architecture spans a swamp-load more of various devices. Everything from a
PSP to your cable modem is usually run by some variant of a mips processor.
However, I made the decision long ago to only focus on the workstation hardware
because I wanted Gentoo to be the "User's Distro" on these machines. I didn't
want us to run off and support these obscure development boards that cost an arm
and a leg, and are only available to very specific individuals who just happen
to know the right people. Lord knows the SGI machines alone keep things
interesting as far as support matrices go.
But largely, Linux/MIPS leaves people with two choices for a distribution:
Debian or Gentoo. And while I give props to the Debian people for keeping the
mips binary world alive, I don't think it'd be right for us to pull out and
reduce those choices to one. After all, Gentoo is all about letting the users
have choices, right?
So there you all have it. My thoughts, my opinions, my apologies. In the end,
I'll go along with whatever the rest of the distribution wants to do to rectify
things. After all, most of it stems from my own inactivity, by and large, and
that hasn't made us a lot of friends around here. So it's time to fix that, and
put an end to all this pointless, utterly stupid bickering that drives away some
of the best talent we have.
Besides, BSG returns in two months. I will probably become more scarce than
George Carlin at Catholic Mass when that happens, permanently affixed to my
television trying to grok whatever crazy stuff Moore throws out in this final
season.
So, thoughts?
--Kumba
--
"Such is oft the course of deeds that move the wheels of the world: small hands
do them because they must, while the eyes of the great are elsewhere." --Elrond
--
gentoo-dev <at> lists.gentoo.org mailing list