Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 10 Mar 2004 23:34:07 -0800 (PST)
From:      Mike Hoskins <mike@adept.org>
To:        Paul Robinson <p.robinson@mmu.ac.uk>
Cc:        'Willie Viljoen' <will@unfoldings.net>
Subject:   RE: Desktop FreeBSD
Message-ID:  <20040310222605.J90761@snafu.adept.org>
In-Reply-To: <000201c4068b$8250a300$6f01a8c0@MITERDOMAIN>
References:  <000201c4068b$8250a300$6f01a8c0@MITERDOMAIN>

Next in thread | Previous in thread | Raw E-Mail | Index | Archive | Help
On Wed, 10 Mar 2004, Paul Robinson wrote:
> If you don't think workgroup servers are servers, then please, go and
> find another project to troll on. A large amount of work is being done
> in both the Linux and BSD communities to help us deploy our preferred
> OSes into an environment in which currently, Microsoft dominates and
> excels in.

i don't think anyone said "workgroup servers are[n't] servers."

also, dominate != excel.  e.g. i can access mail via IMAP on an exchange
server with a reasonable number of mailboxes (<500) and a reasonably sized
IS (<100MB/user)...  i can do the same thing on UNIX running courier IMAP
and it is much faster, measurably so in fact, even with more users (total
and simultaneous).  from a technical/performance standpoint, the UNIX
server is the better bet.  it also costs a lot less, given staff that
knows how to run through a UNIX install and setup courier.  (lots of free
docs on that, which make a lot more sense than some of the m$
"whitepapers"!)

the point is, yeah, sure...  we can get better, and i'm all for doing
everything we can to take more market share away from m$.  however, you
seem to imply microsoft is in certain markets because they "excel" in
those areas -- not entirely true.  they are in many markets because of
advertising and business initiative.  that's the way a lot of big
businesses work, and it's not just that way with m$ products.

it's not exactly that simple, which is why i used the 'entirely' quatifier
above, but it is often the case.

> There is a phrase about England you know, that states "it is a nation of
> shopkeepers". What this elucidates is that 95% of businesses employ less
> than 5 people. The low-end workgroup server market, in the UK at least,
> is MASSIVE. If Unix does not dominate in this space within 10 years,
> Unix is most likely to die out eventually.

UNIX isn't dying out, it's changing shape.  that's evolution!  if joe
shopkeeper can get a functional mail/av/firewall/gateway appliance, a
fileserver, and a handful of workstattions which offer better performance
and higher uptimes for a fraction of the cost of 'beffier' servers +
windows licensing fees, that's the bottom line and they'll bite.  in fact,
a lot of consultant friends are making a living setting these things up
and customers (including non-profit offices with <20 employees) are
grinning ear to ear.

> > It's sad that people are willing to pay large amounts of
> > money to Microsoft
> > for their inferior technology, just because it has a
> > reasonably nice looking
> > GUI, and at the same time, want the nicest, glitteriest, easiest,
> > "everything-est" GUI on UNIX, but always want it for free.
> Oh please. Really, you're not doing the project any favours by spouting
> this fascist party line.

actually, he's partially correct.  the GUI is actually ugly, imo.
however, the technology is often inferior.  just ask your local IT guys.
when i happen to walk across the cube farm and check in on ours, i don't
see happy faces.  print servers, file servers, etc. aren't just chugging
along happily like my UNIX boxes supporting millions of subscribers on <1k
servers.  instead, they're often croaking under reasonable (office of <250
people) loads.  my point: the /technology/ may not be inferior, but the
implementation often is.

as i stated above...  'it's not exactly that simple' -- one thing m$ does
well is delivering a look and feel/feature set business people like.  the
business people have the money, so m$ survives.  this is the case for joe
user as well.  even if they routinely get bsod's, they'll keep going back
for more...  because at least the add printer wizard looks the same every
time (or actually exists, since they'd hate to edit text files).

so there are lots of reasons people use m$...  most of which have nothing
to do with 'better' technology, and most certinaly do not have to do with
good implementations of that technology.  i don't mind if you say
"sometimes business people just want to use windows" but part of me does
mind when you try to say "people use windows because it's better at a lot
of things".  this is freebsd-advocacy, after all.  each platform has
strenghts and weaknesses, and the reality is people often choose a given
platform (including open source) without knowing the strengths or
weaknesses!  obviously there are other factors at work, but time levels
the playing field.

> - As of Windows 2000, with Active Directory they have a set of security,
> deployment and user management tools that quite frankly, makes anything
> we have right now look a bit... well... simplistic AND over-complicated
> at the same time! How did we manage that?

actually, no.  deploy openldap and you can get what AD offers with less
frustration.  from a directory perspective only mind you, but it can be
tied into our OS with work.  sadly, linux is actually better at this right
now.  (i think i mentioned ~700 servers.  a lot of those are linux
authenticating off of openldap.  sadly, to do the same with *BSD would be
more work...  and also largely impossible since they're running java
apps.  on what open source OS does IBM develop/test their JDK?  not *BSD.)

the problem with m$, and things like AD which was supposed to represent
their willingness to support sensible standards, is that they bastardize
the technology just enough to make it a nightmare for the implementation
folks.  (that kills your ROI argument, but i'm getting ahead of myself.)
"it's LDAP...but nowhere near the RFC".  kind of like sockets, COM, DCOM
or whatever else they're calling the trojan horse this week.

> - As of Windows 2003 Server, they have something that can actually
> compete with Unix in both the high-end and the low-end server spaces.

they've been competing since NT 3.5, sadly.  at least on the low-end.
high-end has taken off a lot since 2000.  however, i've seen lots of cases
where windows was chosen for all the wrong reasons.  not for performance,
not for features, not for cost of deployment, not for "TCO", etc.  still,
all that matters is it was deployed.  again, the reasoning is complex and
goes to show that having billions of dollars to pour into market research
pays off.  why did VHS win over beta?  the experts still use beta.  it was
largely due to an influx of affordable VHS players...  now guess which
servers are more affordable when spec'd for a given workload? (hint: not
those running windoze.  also not those running solaris.)

> IS Management is not about the best technology, or the cheapest. It's
> the one that produces the biggest and quickest ROI. For 99% of companies
> out there, MS represents the best ROI right now. It does so, because you
> don't need to be a computer expert to setup a server, just have some
> common sense. The same can be said about Unix, but you have to go and
> read a bit more. And you know what, if it means somebody has to read ONE
> PAGE more of documentation, for 99% of the workgroup market, that's one
> page too much.

actually, you're 1/2 on and 1/2 off.  m$ servers are not for the faint of
heart, at least if they're doing much and need to stay up.  it's not
eaiser to keep a windows server functioning than it is a UNIX box...  and
it's not easier to setup either, if you're doing anything useful.  if you
pull in AD, m$ DNS, Exchange, etc...  you'd better have a "computer
expert" to set things up and monitor your servers -- or the CEO is going
to be upset when he can't get his email.  the same can happen on either
platform.  i've watched countless MCSE's pull their hair out over cryptic
error codes that mean little to me as a UNIX guy at heart.

> Would you like me to build you a cross to hang yourself on, you tortured
> soul?

do you work for m$, or are you a UNIX user?  i know the answer, but your
love of sarcasm keeps me guessing.

> That argument really needs to die. You either work on something because
> you're being paid for it, or you love it. But when somebody points out
> it doesn't work well in a particular scenario, it might be best if you
> don't start acting with indignation.

i'm just trying to point out that many things costing more money and
giving little to no benefit are deployed.  people are starting to see the
light, however, and it will continue to improve.  open source is on the
rise, particularly outisde the US.  thankfully, many paranoid governments
don't trust the capitalist glutton we call m$, and are opting for open
source for their technology initatives.  even in the US, i've been seeing
more and more things that would have been unquestionably m$ go 50/50 (so
IT guys can get their feet wet without taking too much risk).  none of
this points to "UNIX is dying in 10 years" -- so please do more research.

> You're the one on -advocacy. You're the one telling them that you are
> providing is better and complaining when they say "sorry guys, but samba
> just played with our heads till 4am, so we deployed 2003 server
> instead".

samba can play with your heads, active directory can play with your
heads...  you either have staff that likes technology and figures out how
to make it work, or you hire one of the 300 other people in line for the
same job who will gladly make it work.  regardless of platform, things
break.  people make technology work, like most other things in life.
unlike your observations, i've been seeing a natural growth toward open
source (*BSD included) in traditionally non-UNIX camps (IT).  that makes
me happy, and i wish you could share in that vision.  i don't think m$
will (or should, free choice and all) go away, but i do know UNIX isn't
dying (including *BSD).

the best thing to do (which i thought we'd been doing for years, and i'm
sure we'll continue to do so) is to compete with yourself.  we've got
people that love what they're doing working on the project.  we don't have
billions of dollars.  time is on our side, and we're getting better day by
day.  so is the competition, but they're charing a lot more for inferior
implementations...  because they do it for the money.  i believe that's
why places i've seen are transistioning away from m$ (and closed source
UNIX) in favor of open source projects like Linux and *BSD.

with all that said...  i think everyone's partially right.  the choice of
OS/platform/whatever is highly personal (no matter how technical), and
will likely never change.  that's why measuring 'useability' is so hard
for software folks, on all platforms.  i also agree it's not a good idea
to have blinders on -- that doesn't help anyone.  i will, however,
adamently disagree if anyone says m$ is always (or often) chosen because
their technology is superior.

-m

--
 "Information Warfare? Given the state of the industry, what we need is
  Information Welfare."  --Richard A Steenbergen



Want to link to this message? Use this URL: <http://docs.FreeBSD.org/cgi/mid.cgi?20040310222605.J90761>