Luke Kenneth Casson Leighton
lkcl at samba-tng.org
Fri Jun 22 14:05:01 GMT 2001
> -----Original Message-----
> From: samba-admin at lists.samba.org [mailto:samba-admin at lists.samba.org]On
> Behalf Of Michael E Osborne
> Sent: Thursday, June 21, 2001 12:27 PM
> To: Luke Kenneth Casson Leighton
> Cc: samba at samba.org; samba-technical at samba.org; samba-ntdom at samba.org
> Subject: Re: web page
> I have been using Samba a long time and there was a time when I actually
> looked forward to your posts on this list.
yes, they were good times. i enjoyed what i was doing, i appreciated
the support and encouragement i got from everybody. it was fun,
exciting. i was fighting microsoft, but keeping them happy at
the same time [reports to secure at microsoft.com]. i was...
important, needed, respected and quietly appreciated.
> That time is definately over now.
well, things change. you know, like being told by
dr andrew tridgell for five months that any idea i
came up with was either unjustifiable, worthless,
that it was someone else's problem or responsibility.
i'll explain more, below.
did you know that when tim completed winbindd, which
i helped him design, andrew thanked tim for his work...
but failed to say thank you to me, in front of the
rest of the company, for the work done for THREE YEARS,
without which tim's work would not even have been possible?
simple thanks and gratitude go a long, long way in
project management - ESPECIALLY so in open source development.
> While I respect your technical abilities and appreciate the work that
> you have done on Samba, your posts now are simply sad.
yes. i know.
> I'm not sure what
> you intend these posts to accomplish. I can tell you how that they reflect
> *very* poorly on you not Samba or the people you are unhappy with.
okay. well would it help for me to explain _why_ i am making these
i will assume that the answer is yes, because of the people
who have contacted me [to say pretty much what you have said,
here], their reaction to my privately made explanations has been
ah, heck. here goes.
i started on the samba project back in june 1995. rewrote nmbd,
then in august 97, quit the company i was working for, and
decided to focus - in my _own time and money_, on samba.
with paul ashton, and later jean micouleau francois, we got
the first 'Welcome to the SAMBA Domain' message. it was great.
over the next few months, my weight dropped down to 63 kilos
(i'm 6ft1) which is 70% of what i am now. by january 98, i had
to use two hands to turn they key in the lock to get into my house,
and to ask the neighbour and friends to get the lids off jars.
[i damaged a couple of doors, you see, using them as a lever.]
august 98, i was working for ISS. i spent 18 months with them,
continuing to develop samba, with a view to licensing that
code to them so they could also use it in their products.
these were good times.
some time in '99, redhat went public. several people, including
the samba team, were offered shares in redhat. i didn't have
enough $ at the time, so could not take part.
needless to say, at a conference a couple of months later,
i was a bit annoyed when someone came up to me and said,
'oo, wasn't the redhat ipo good, i was able to pay off my
mortgage with it, i made $50,000 out of it, what did _you_
do?' and he was a bit embarrassed when i said, nothing.
november '99, valinux went public. they contacted several
people, including jeremy and andrew. jeremy was on holiday
at the time, and was working for them, so he got lots of
shares, but he had _assumed_ that i had also been contacted,
knowing the value of my work, he was very surprised.
i spoke to larry and chris de bono, but it was too late.
they promised to get back to me, to see if anything could
be done, but never did [and a year later, when i was introduced
by andrew to chris when we visited VA, he didn't acknowledge
that i was even there - just looked straight ahead, at right
angles to me, and ignored me.]
now, what were the 'warning' signs that, coding-wise, something
actually, from andrew, the warning signs were back in about...
october 97. i got word that andrew was jumping up and down
and had to be 'warned off' by john terpestra, to let me
discover this stuff on my own.
the warning signs from jeremy were actually back in 96.
does anybody remember samba-1.9.17-multiworkgroup.tar.gz?
the first time that a single host was demonstrated on the
microsoft campus of being capable of managing multiple domains
was samba 1.9.17-multiworkgroup, in august 96.
he he he :)
anyway. that code had the 'cascading WINS' capability that
chris hertel *only just recently added to samba 2*. and
it had it *four years ago*.
that code sat there, rotting. for several months - probably
about six - i was watching andrew and jeremy make *the
same mistakes* and *fix the same problems i already had*.
the archives are full of messages, 'i already fixed that
bug in the mwg code, two months ago'. '
'i already fixed that bug in the mwg code, _four_ months ago'.
until eventually, i gave up, and discovered DCE/RPC
[welcome to the SAMBA domain].
the warning signs, there, were when jeremy couldn't deal
with the rapid development pace i was keeping. i deliberately
sacrificed certain aspects, knowing that the code would
be thrown out or morphed later into better code.
to save time, i used static arrays, i mean we didn't even
_know_ what a sizeof() was, or ANYTHING! we were basically
missing an entire infrastructure and just network-reverse-
engineering not only a codebase that's 250,000 LOC in its
own right (the Dce/rpc IDL compiler _alone_ is 50,000, and
the TOG's NDR marshalling library is 20,000, and we only
_had_ about 20,000 LOC *in total* in the BRANCH_NTDOM)
.... which is where, in cvs, i ended up. with hindsight,
i realise that this was a mistake, to let this happen.
it would have been _far_ better to just continue in
main-line samba, and wow, big deal, the code works in
smbd absolutely fine, but the NT domain stuff? pffh :)
crashes like a Cable & Wireless CableTV play-thing.
... but well, so what? it's open source, right?
some bits are stable, some are in development. we _expect_
but no, jeremy wanted 'stability, stability, stability'.
in code that wasn't even pre-alpha quality and we
didn't even know what the protocol was, he got so
irked that he ran purify on it, cleaned it up and then
stamped his 'approval' on it - andrew even claimed
the code was _his_ - and refused to accept, repeatedly,
and adopt and run the same clean-up and assistance
process, again, on future versions of the same codebase.
with the result that HP sent in patches to the samrd
code that was 2 YEARS OUT OF DATE!
amazing. absolutely amazing.
my health got slightly better, at ISS. i had a chair
from ergointerfaces.com, it's utterly cool: the
keyboard is split and mounted on the arms on gimballs.
i was the only person to be able to _comfortably_ sit
at 'The Command Console' with my feet up on the desk,
six feet away from the monitor :) :)
good for the eyes, good for my arms, good for my ATTITUDE :)
so, anyway, november 99, i was irked, _all_ of the
Linux Companies that had recently IPO'd had offered
jeremy and andrew shares, and me, i had none.
i figured, hey, there's this new company, Linuxcare,
they're planning to go public, maybe i'll join them.
and with andrew's encouragement, i did.
he was really enthusiastic. when i arrived, the first
thing we did was go out canoeing on canberra's lake.
we went *through* the 100-ft fountain's spray, like
you're not supposed to. it was great, he really
looked after me, and i began the 'brain-dump' process
right then - i explained SURS to him - on the lake.
it became known as 'the canoe algorithm' :) :)
anyway, it became evident that there was something wrong,
at linuxcare. the warning signs were that a) i hadn't
received a computer to use b) i wasn't on a payroll.
i got the computer, after a while, but pay? _that_ took
8 weeks, and... okay, it cost linuxcare about USD $2,000
in bank charges, probably a lot more in other costs
such as lost time and international telephone calls,
and it cost _me_ my standing IMMEDIATELY - RIGHT FROM
THE FIRST WEEKS WHEN I STARTED - with the management
and some of my colleagues at linuxcare.
10,000 miles from anywhere, with no money, i stand to
lose my home in the u.k, and i can't do a DAMN THING
it's raining. i haven't slept. i'm worried sick.
four in the morning, i'm staying at a hotel, i call up
linuxcare's personnel manager, and i say, hi, can i
talk to you? she says sure, what's wrong? and...
i couldn't answer. that's never happened to me before.
i was _that_ upset, i couldn't answer a simple question.
anyway, she suggested i try get some sleep - 5 hours -
and i was able to answer her question. just knowing
that there was someone who could help was enough to
ease my mind, a bit.
anyway, alarm bells are going off everywhere. andrew's
starting to be affected by this. my salary from ISS
was what i was used to, and all of it accounted for.
compared to canberra salaries [a university lecturer at
ANU only gets $AUD 40,000, which is about $USD 20,000]
i was on _mad_ money. andrew, i think, got jealous.
i mean, why was i creating such a fuss when i was
being paid what he thought was stupid amounts of money?
so, amongst these psychological pressures - probably
some of the worst i've ever had to deal with in my life
so far - i'm still trying to explain to andrew what's
been going on in the TNG codebase for the last three
now, i've written a book about it. he's read it. and
it's _nowhere near_ everything. he even said to me,
'you know, i really don't get this stuff as well as
the thing is, i _know_ my weaknesses. lack of confidence
being one of them. ability to fight and stand up for
what i believe is the right way to do things, even
if it has limitations.
well, i'm learning.
anyway. in short, andrew became an enemy of every idea
i came up with. i mean evvverrryything.
over a period of five months, i eventually had to
stop speaking to him or to tell him anything.
because he wasn't interested any more in technical
solutions, he was more interested in doing it
'a different way from me'.
secrets.tdb. i spent three months developing an LSA.
i'd investigated LsaQuerySecret, and come up with an
interface to store $MACHINE.ACC and trust accounts.
it uses Unicode strings and NT times, which contain
andrew, of course, wanted unix conventions. char*
and time_t. now, of course, a domain name in anything
other than an 8-bit character set won't be recognised
by Samba 2.2 / Main, and now you know the reason why.
because mister dr andrew tridgell just _had_ to come
up with something that trashed three months development
work in favour of a quick hack of three days.
so, in the end, to save my sanity, i had to quit.
you've seen that: we now have TNG, and i haven't developed
a single _new_ line of code of Samba Source - including TNG -
for eighteen months, now.
FOUR YEARS WORK, and i'm now doing NOTHING???? how do you
think that makes me feel?
so now you know.
i own the copyright on 25 to 30% of the 350,000 line samba
codebase, and i have damaged nerves and tendons to show for
it, and not enough money to get that fixed, and not enough
money to retire from computing - which is my LIFE, dammit.
several companies and individuals are beholden, and in
a position to be blackmailed by me, over that code, and
not only _that_, but i am sufficiently... wound up,
hurt and insanely upset, enough to actually consider
doing just that.
i may just assign all copyright to the Apache Software
Foundation so that i don't actually succumb to the
temptation [that has implications, though: i have
conditional agreements with people regarding code copyright
/ ownership, it must be GPL or they receive ownership back
so please, when you talk of 'oo, he makes himself look
a bit of a sad idiot', perhaps you might wish to take
a look from my viewpoint, and see how things stand.
what would you do?
what would you do, andrew?
what would you do, jeremy?
> Just person offering a view
much appreciated. inspired me to offer this... reflective
and hopefully more rational view.
More information about the samba-technical