[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
It seems that, in demanding a "manager", he's trying to apply
the tool he knows to a situation where the tool isn't needed.
Apparently, he wants a leverage point that will grant control
to the wielder. That's a very strange thing for him to want,
from a FreBSD perspective, but a natural one to want, from a
The problem is that he wants to control a volunteer effort in
order to build what he thinks should be built, instead of what
they will build out of interest.
This is the same mistake Mozilla made, and the same mistake
Sun has made with Java (and then componeded it with a license
which makes it practically useless; the only thing it has
going for it is the old MIS manager's dream of being able to
get rid of all the tiresome programmers -- a dream which will
never bear fruit -- so that these days, if you can spell
"Java", you can get a realtively high paying position).
The statement "There is alot of friction between the BSD/OS
side and the OSD people. Mainly because of stupid things the
OSD wont accept and that should never have happened." is very
If there were a "manager" (what he really seems to want is a
person who is to FreeBSD what Linus is to Linux), there would
be someone that they could pressure to get FreeBSD to do what
they want, such as accepting changes from BSDI "for their own
I think that it's generally understood that the only person
who could credibly take on this role would be Kirk McKusick.
I think he also misses the point, and protrays one of the
greatest strengths of FreeBSD's as a weakness: people work on
what they feel needs to be worked on, and "owning" several
core team members and/or committers doesn't give you carte
blanche write access to the tree.
They need to learn from other commercial organizations which
have used FreeBSD in their own products -- they need to learn
from Walnut Creek CDROM, however much of that is left.
Sure, FreeBSD has poor productization, and while easy, the
install doesn't have that polished look of Windows (I think
that maybe BSDI had hoped that FreeBSD would become Windows
2000 desktop to BSDI's Windows 2000 server).
The FreeBSD organization also has limits to the complexity it
can handle. Big projects don't get done without a lot of
buy-in from the group, and complex projects can't be done by
large groups of hobbiests -- most of the complex things which
have happened in FreeBSD have had commercial funding of a
small team actually doing the work. There's truth to the old
addage that begin "too many cooks...".
Another interesting thing is that complexity is exactly where
large companies are attacking their Open Source competition.
There are many new standards that have been pushed through the
IETF at significantly high complexity levels, compared to what
went before them. Much of that complexity is, IMO, gratuitous
and aimed squarely at being too difficult for an Open Source
project to implement correctly. Microsoft and Novell are old
hands at this game, with their obscure wire protocols for file
sharing: Microsoft intentionally, and Novell by way of including
all of their historical APIs in their client developement kits.
Novell has actually been hoist on its own petard by doing this,
since programmers will use the oldest APIs they can to get the
job done, since they want to capture the largest potential
customer share from the installed Novell base. In turn, Novell
servers have to support all the historical APIs to keep such
programs running against new servers.
Netscape actually tried its hand at the complexity game, to try
and control the LDAP market. I actually gave very bad odds of
the OpenLDAP project ever implementing LDAPv3 because of this,
given their initial license, and given what I knew about the
FreeBSD and Linux capabilities in this area. I warned Kurt
Zelinga about this when he started, and actually distanced
myself from the project because of this. Happily, I've been
surprised, and Kurt's style is closer to the Apache/XFree model,
which doesn't have the drawing power of the FreeBSD model, but
which likewise attracts only highly skilled and self-disciplined
people. Put simply, FreeBSD and Linux have a lot of activity,
and OpenLDAP and Apache have a lot of action. Realize also,
that a single commercial employee working with the project as
part of their work can have much more impact, and therefore
control, than with FreeBSD or Linux (unless you buy Linus, as
Transmeta did, and even then, you don't get as much control as
you thought you were buying).
I think that's what BSDI thought they were getting when they
bought Walnut Creek CDROM; maybe now they are having some
"buyer's remorse" about not getting what they thought they
were buying. That's really a silly and short-sighted view, I
think. It ignores what they ended up getting, out of pique
for not getting what they thought they were paying to obtain,
and so they are squandering what value is there out of self
pity for what isn't.
Obviously, Bill's just one guy; his views may not be all that
representative of the views of BSDI at large, and so this
trieste might just be about Bill's understanding. On the
other hand, if BSDI is really losing FreeBSD folks from their
ranks, it's probably a view which is more than just one
Any opinions in this posting are my own and not those of my present
or previous employers.
To Unsubscribe: send mail to majordomo@xxxxxxxxxxx
with "unsubscribe freebsd-chat" in the body of the message