Date: Thu, 03 Jun 1999 23:01:25 -0400 (EDT) From: Vince Vielhaber <vev@michvhf.com> To: Chuck Robey <chuckr@picnic.mat.net> Cc: Jaye Mathisen <mrcpu@internetcds.com>, Jordan K.Hubbard <jkh@zippy.cdrom.com>, Matthew Dillon <dillon@apollo.backplane.com>, freebsd-hackers@FreeBSD.ORG, Garance A Drosihn <drosih@rpi.edu>, Amancio Hasty <ahasty@mindspring.com> Subject: Re: Matt's Commit status (was Re: 3.2-stable, panic #12) Message-ID: <XFMail.990603230125.vev@michvhf.com> In-Reply-To: <Pine.BSF.4.10.9906032125480.82061-100000@picnic.mat.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On 04-Jun-99 Chuck Robey wrote:
> On Thu, 3 Jun 1999, Vince Vielhaber wrote:
>
>> > Just realize, IF you're loud enough, and succeed, the programmers will
>> > all desert you, and you'll have a nice place to argue, but no more
>> > software. Core here does an excellent job, with all the problems they
>> > face, most committers will agree to that pretty quickly, and they are
>> > the only ones with a vote. Look to reality for the reasons why.
>>
>> No need for blowing up, so relax. You may wanna grab that bourbon
>> bottle yourself and take a sip.
>
> Maybe you're right, but I'm tired to people talking about core like
> they're some evil overlord group. These guys are longtime FreeBSDers.
> They've been doing it a long time, and a large part of why they are
> still doing it is because of a feeling of responsibility. They're doing
> a great job, even if it's not 100% without error, and FreeBSD would go
> right down the tubes pretty quickly without coolheaded guidance, just
> because of the lack of control. You don't have to look hard at all to
> see several recent (and some ongoing) episodes of folks trying to go off
> on their own, and the only real restraint is the realization of what
> kind of limits core will allow.
>
> It's got a *great* track record. We can argue specific issues, but
> attacking core itself, that I'm going to jump up and shout about.
>
>> 2) Nothing was ever said about telling volunteers what to do or not do.
>
> You referred to a group other than core setting policy.
No, I did not say setting policy. A place for an appeal or an oversight.
I'd never advocate a group other than core for policy - if that's what
you got out of it then either you misread or I poorly worded something.
If it's the latter then I apologize.
>> 3) Nothing was mentioned about the technical abilities of an appeals
>> board or oversight group.
>
> You said that group would have no committers or core on it. Anyone
> who's shown skill and time has been pretty quickly asked to become a
> committer, so that pretty much means you're asking for either
> non-technical folks, or folks without the time to follow things close
> enough to have any real idea what's going on.
>
> Seeing as you did admit you haven't followed the issues closely enough
> yourself, you're one of the ones you figure would be on that board,
> right?
Actually no. I haven't followed the issues closely enough by choice. I'm
not sure I'd really want to get that deeply into the politics of it. Even
if asked I can't say that I'd consider doing it. The reason I'd want no
core or committers on it is strictly for political reasons. Core *appears*
to hold a heavy hand and it's very possible the committers and/or coders are
intimidated by them. Now with that scenario would you want either to be
on any such committee? I'm trying to look at this from a completely detached
viewpoint.
>
>> 4) Ever do or say something that someone told you that you're out of
>> line?
>
> Oh, yeah, sure, I make mistakes. Maybe wrong here, but I said above
> what keyed me off. It sure isn't bragging if I say here that I'm
> willing to bet I make more mistakes than you do, but I don't think I'm
> wrong here.
I dunno, I could probably give you a good run :)
>
>> 5) Let's go back to your volunteer thing. You have volunteers willing
>> to work for you - for free. It's no secret that in the past core
>> has been a problem for some (or maybe even many) of these volunteers.
>> They went away.
>
> As has been pointed out endless times, you have to know how to code and
> be willing to read FreeBSD code enough so that you can contribute well
> integrated code. If you can't code or don't have the time, no, you
> can't contribute that way. You can help us by putting in problem
> reports, but not by trying to gain control, even a small part. The
> coders control FreeBSD, because they love it. Those are really the only
> folks who get to directly contribute.
>
> Free contributors, out of control, are a mob.
Can't the same be said of a closed minded group whether they're core
or contributers?
>
>>
>> If there's noone they can go to, why should they (or anyone else) bother
>> to volunteer?
>
> The FreeBSD mailing lists are the most active, quick response groups on
> the net. Don't you feel silly claiming there's no one they can go to?
Not at all. Why should this ever have gotten to the lists? Don't you feel
silly that it had to be aired in public? Matt's complaints were expressed
just a few months ago in this same forum. "silly" is a weak term to describe
why it was handled the way it was, doncha think? Embarassing for *all* of
us (core, user, contributer, bystander, floorsweeper, whoever) is more like
it.
>> Earlier in this thread Matt was accused of running rough shod over everyone.
>> Looking over the history and hearing the complaints of some of those involved,
>> I admit not knowing both sides of the FULL story, that core may be partially
>> to blame. But who is there for the *VOLUNTEERS* to turn to? Core? They
>> may have caused the problem. Then once the emotions start to rise there's
>> absolutely NOONE for any of them to look to besides Jordan and/or David.
>> Like they don't already have enough to do.
>
> That last paragraph ... "core may be partially to blame". Did you read
> the entire thing, which went on in the -current and -committers list?
> If you didn't, then you don't get a chance to comment. If you want to
> jump on core, read the mailing list archives and get the arguments in
> order. You can't just slander folks that way, and expect it to be taken
> as just innocent sniping, because sniping isn't innocent.
I've slandered noone. Take another sip of the bourbon. There's more history
than this one incident. If you want to try to isolate it to that then I'm
not the one that shouldn't "get a chance to comment". Like I said, I've
slandered noone, and this isn't "innocent sniping". I'm not taking sides
in this - although it's obvious you already have.
> In fact, Matt was hearing a great deal of negative comment on his
> commits, from folks who originally did the code, and he was ignoring it,
> claiming it "slowed him down". This went on over weeks. The only way
> to stop him, demonstrably, was what was done, because all the mail
> arguments did was raise temperatures, not slow things down a whit.
Please reread both Matt's and John's comments. They're both summed up
in this very thread. Matt's are pretty much the same as they were a few
months ago, without going back and rereading John's I can't say if his are
but they're still here to read. It doesn't exactly coincide with what you
say above.
> Matt does *great* code, he just doesn't like to read and test as much as
> code away and react later to problem reports. Being that he was working
> on the virtual memory system, something which can very easily get
> totally beyond repair without careful testing, his refusal to slow down
> was more than could be tolerated. He may well write better code than
> nearly all of us, but he needed to do it slower. He sure as heck writes
> better code than I do.
>
>> It's obvious there's a problem with the status quo, but if the status quo
>> continues to run rough shod over any possible solution, status quo will
>> end up running out of volunteers and there will be yet another BSD to add
>> to the collection. Maybe the next one can be called ClosedBSD?
>
> Why, because you listen to a few days of the loudest complainers, and
> decide without further ado that they must be right?
Why do you ass-u-me that I've just arrived here? I may not have been
around here from the beginning but I've been around since at least the
early 2.0 or 2.1 releases (I can go dig up old CDs if it really matters
that much). I'd have been here sooner if I'd made a better choice of
SCSI cards. I also have subscriptions to both the releases and toolkit,
and I insist that any of my customers (and even friends) that I provide
tech support for also get a subscription to help the cause.
> There were good reasons behind both of the two recent problems, but
> you've admitted you haven't done the reading to comment on it. Why do
> you then feel qualified to say it's core's fault? Say I agreed with you
> (I obviously don't), what would your reasons be? If you can't *really*
> answer that, and not just add in suppositions, then you are being
> ethically dishonest.
Once again I did not say it's core's fault and if you insist on continuing
to insinuate that then I'd have to take back my previous suggestion of you
having another sip of bourbon! I'm taking NO SIDES! More than likely
there's enough blame and misunderstanding to go around - the one position I
will continue to take, tho, is that it *NEVER* should have gotten this far.
I'd hope that you'd also agree with at least that point. And please don't
question my ethics or honesty if you don't know me. To those who do know
me, you wouldn't be cast in a flattering light which is probably unfair to
you.
>> Now reread your own last paragraph. It goes at least two ways. How much
>> talent and support (technical, financial, advocational[1], etc.) are you
>> willing to lose before anyone's told they're outa line? For all that
>> matter, how much has already been lost? You used the phrase "crack pot".
>> Doesn't a cracked pot leak? Isn't that what's already happening? Talent
>> leaking out?
>
> The problem *was* folks who wanted to code far faster than the review
> process would allow. Core still feels that's wrong. If Matt were
> willing the *guarantee* that he wouldn't overrun reviewers, then he'd be
> back as a comitter tonight.
Humor me here. Are you saying that core couldn't keep up with Matt so
it's all his fault and if he were to promise to go to the bar more often
or get a second hobby everything would be ok?
> It is *not* worth it to bring on all possible contributors, if they're
> going to innocently cause more damage than they fix, even if they are
> doing it with the best of intentions.
Believe it or not I actually agree with you here. See, all hope is not
lost.
> Controlling contributors is the whole problem core's been working on,
> they fully realize the tradeoffs, do you really think they don't?
I would hope not. But let's face it, that's not the best choice of words
for that description, now is it? Or perhaps that's the problem? Core
couldn't *control* Matt? Sorry, that was a cheap shot and undeserved,
but you can see how easily it can be slid in by anyone else. See what I
mean? There's a big difference between controlling and working with (I
think the term 'peer' was used). Once again, please don't take offense,
'cuze there's none intended.
Anyway, it'd probably be best that if you want to continue the discussion
we did it out of the public spotlight, or at the very least not on hackers.
But feel free to include anyone you wish.
Vince.
--
==========================================================================
Vince Vielhaber -- KA8CSH email: vev@michvhf.com flame-mail: /dev/null
# include <std/disclaimers.h> TEAM-OS2
Online Campground Directory http://www.camping-usa.com
Online Giftshop Superstore http://www.cloudninegifts.com
==========================================================================
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?XFMail.990603230125.vev>