TNG-stable

[Karl Denninger]

On Tue, Sep 26, 2000 at 05:37:03PM +0100, Thom May wrote:
> > That is simply
> > NOT TRUE if the machines on your network are not Win95/98 clients, or if
> > you use things that require Exchange!
> 
> And that is simply bullshit. We have a very happy 
> system built on a Samba 2.0.7 server using NT,2K, and 9x. It works fine,
> within the limitations that the documentation states. It works
> entirely as advertised.

Oh please.

Look, I run 2.0.7 right now.  And I run Win2k.  It CANNOT join a Samba
domain.  It COULD join a TNG domain, but TNG was too unstable to seriously
consider for production use (and was CORRECTLY LABELLED as such)

> > Unfortunately, this myth persists.  It persists because the DEVELOPERS
> > want it to persist.
> It persists because it is *NOT* a myth.

Yes it is.

> > This whole hullaballo could be reduced to zero by simply saying, in plain
> > english and with no ballyho at all:
> > 
> > 	Samba provides SOME functionality for file and print service.  It
> > 	is NOT a Win2k replacement, it DOES NOT provide anywhere close
> > 	to a full set of MSRPC services, and we HAVE NO IDEA IF OR WHEN 
> > 	IT EVER WILL.
> Ask Microsoft when the *scheduled* release date for Win2k was. Ask them when
> the release date for Exchange 2k *is*. 

So what?

Look, TARGETS are TARGETS.  They can (and frequently are, in the software
world) MISSED.

But AT LEAST THERE ARE TARGETS!

> And I'd be willing to put money down that you would say the
> latter. And then come tearing up back onto the list at one
> second past the deadline, without pausing for breath, to bitch
> about the lack of product.

Wrong.

I'm a software and network engineer (yes, I do both coding and network
stuff) by trade.  

I've been in this business for close to TWENTY YEARS.  I have managed
LARGE projects for national roll-outs in which REAL products and services
depended on hitting targets.

I know the difference between a TARGET and a PROMISED RELEASE DATE.  I 
have had to explain those differences to Directors and Corporate Boards 
before, in V.E.R.Y. S.L.O.W. L.A.N.G.U.A.G.E. to insure that they
UNDERSTOOD when they were grilling me about it EXACTLY what I was 
promising and EXACTLY what I was *NOT* promising.

But unlike many people who were FIRED while I was not, I did not CONFUSE
people by allowing them to BELIEVE that I could deliver something in a
timeframe when I could not.  If it took me an hour of explaining all
of this, and all of the uncertainties involved, and why I had the word
TARGET in my presentation instead of RELEASE DATE, I took the time and
was very, very careful to make SURE they understood.

The problem here, and in many "open source" projects, is that the lack
of outside pressure (and thus accountability) leads you to do one of two
things:

1.	Promise the moon, either IMPLICITLY (by silence while people talk
	it up) or EXPLICITLY (by putting up release schedules you know
	damn well can't be met), and then fail to deliver.

2.	Promise NOTHING, then bitch when people say "but we need <X> and
	you're a scumbag for not delivering it 'on time'!"

BOTH approaches suck.

The essence of project management, no matter if you're being paid or not,
is to set REALISTIC expectations.  You should set them a bit BELOW what
you think you can deliver, and adjust them regularly so that they REMAIN
a bit below what you believe you can deliver at that point in time.

If you HONESTLY believe that you can't deliver Win2k PDC functionality
for another 18 months, SAY you won't have it for 24 months.  Explain
EXACTLY what that lack means, and what it prevents.  Allow people to
make INTELLIGENT decisions about whether this impacts their use of your
product or not.

> From my point of view as a systems administrator, and not a
> conslutant, I know which product I prefer day in day out. And
> I'm learning C in the hopes that one day I cna contribute to
> this and other products, with the idea of giving something back
> to the community.

I am a very capable network and "C" programmer, having both done some of
the actual coding and project management for some really big (several 
million lines of code) projects.  I've also done a large number of smaller
projects single-handed, and my experience with this stuff goes back to the 
days of Z-80 macro assembler!

However, I refuse to put my time (which is limited, as is everyone's) into
a project that lacks the kind of formal definition and structure that I
believe ANY large project MUST HAVE - to be successful in the long term.

> > 	Specifically by inclusion but not limitation:
> > 		It is not capable of being a PDC or BDC as defined by
> > 			the reference implementation, Microsoft Win2k.
> See above for comments on "reference"

Bullpocky.  If you're shooting at an undocumented target, then SAY SO and
NAME NAMES.  Nobody puts up with this crap when you do it ex-post-facto.

There have been allegations made here that Microsoft has INTENTIONALLY 
broken Samba interoperability.  I don't know enough to know if that's
true or false, but what I DO know is that allegations like that had better
be factually backed up and should be brought out in the light of day where
they can be examined, because that kind of conduct weighs VERY heavily
on corporate IS departments - or at least, it should.

> > 		It cannot host Exchange.
> Find anywhere on the samba.org website that mentions Exchange in
> anyway. I can't.
> > 		Win2k machines cannot join a Samba domain AT ALL.
> Which is a breakage on MS' part, not Samba's. 

Is it?  What published specification did they violate?  Show me.

> > 		MS administration tools DO NOT WORK.  AT ALL.
> But SAMBA admin tools do. 

Not relavent.

> Basically, the idea that you seemed to have missed is that this
> is not actually helping anyone. The way to help is to turn off
> your mail client, and start a decent text editor and a copy of
> GCC, learn C and contribute. Or write documentation. Or keep the
> website updated. Or *whatever*. but don't just bitch. 
> 
> I hope this is gonna be the last comment to the list. Karl, if
> you're gonna reply, just reply to me.
> 
> -Thom

I'm free to bitch if I want.

Others are free to listen or not as they deem fit.

--
-- 
Karl Denninger (karl at denninger.net) Internet Consultant & Kids Rights Activist
http://www.denninger.net	Cost-effective Consulting Solutions
http://childrens-justice.org	Working to protect children's rights