On 6/13/2010 12:41 AM, russell wrote:
Choochoo does have a point about usability and the home user. I started working
with Sun equipment in 1986, when Microsoft was just MSDOS and the early release
of Windows used to compete with GEM desktop. No in the PC reseller at the time
thought Windows would replace MSDOS it was too unstable.
We used to sell Sun 3/60s with Autocad (as it was developed at the time on Sun
hardware and ported to MSDOS). Many Universities had labs and networks of Sun
and DEC equipment. However, as Windows improved Sun equipment was replaced by
Windows because of cost. Those graduates entered the workplace bringing their
PC knowledge in to reduce costs. Apple at the time used to Graphics Design
market offering discounts to students, it is still a market you will find Macs
dominate.
Linux started in the education/home environment but it has entered the
corporate arena and a platform to replace Unix servers. Home users bring their
experience into work and can influence decisions. If the Linux installation is
reliable then further installations will occur.
Oracle needs to realise that the home market may not make money in the short
term but over the long term it does have an impact. It is the home users under
18 that do become the decision makers of tomorrow, they need to appreciate
exactly what Solaris/OpenSolaris has too offer. If Oracle want Solaris to
survive it needs OpenSolaris to really flourish in the home market or Solaris
will be replaced by Linux or Windows.
Well...
I don't see any System/32 or MVS or AS/400 systems running in (many)
homes, yet those markets are huge (in terms of dollars) and not
shrinking. There's the myth that you have to get everyone before they
turn 12 hooked on your product, or you've lost them forever.
Also, let's be clear - the major reason Linux took marketshare from UNIX
has nothing to do with the new generation being more familiar with
Linux. They weren't making the purchasing decisions. The reason is that
Linux ran on cheap x86 hardware, and the various UNIX-derivatives didn't
(even Solaris wasn't really usable on x86 until Solaris 10 was
released). Cheap hardware trumps expensive hardware, and IT departments
just worked-around the deficiencies of Linux vis-a-vis UNIX. Low-end
UNIX hardware simply couldn't compete on a price/performance standpoint
with x86 anymore.
Reality is that while the "home" user does drive absolute mindshare, the
bigger place that IT-to-be learns what it really useful is University.
And for the Enterprise market (which is what Oracle shoots for), what
you're aiming for in terms of mindshare is the Specifiers of technology
(i.e. upper IT/Dev) - they're the ones who are going to say to the
Business side: "if you want us to do Y for you, the best solution is
X". If you can get them to equate X = Solaris, well, there's a sale.
You're going for mindshare of the TECHNOLOGISTS, not the ordinary users.
Frankly, if you're working in an Enterprise where the BUSINESS unit is
specifying not their needs, but the actual solution to use, well, that
company is SCREWED. Leave now. It will only get worse. Or, be prepared
to engage in a long-running campaign to completely re-do their internal
decision-making process, because the one they're currently using is
bound to fail (or be horribly over-cost, invariably hurting the company,
and your sanity).
Of course, there is a market for giving the customer *exactly* what they
want. It's often lucrative. But, it's hideously painful to work in. If
you are a consultant or a consulting company (like IBM Global Services),
well, you swallow it and bank the huge check. But for those of us who
actually have a work for such a client, it's a never-ending saga of
pain, horror, and cringe-worthy decisions. People don't last long in
those positions. And, honestly, companies that work that way get bought
or fail. Because it's hideously inefficient.
So, yes, it is useful to have OpenSolaris be moderately usable on a
standard PC. However, let's be honest here, it's *never* going to have
enough development support to compete with Windows or Linux in terms of
a full-featured Desktop environment. All you're going to be able to
provide the Home User with is a very limited introduction (which,
however, may still be useful). It's actually really hard to introduce
the Home User to OpenSolaris's really useful technologies (and have them
understand why they are useful, vis-a-vis Linux/Windows). They're
simply not going to have the hardware and network setup to really get
the Full Shebang. *I* can set up a full SAN/cluster/VM/multi-path
network config with $1000 worth of eBay cast-offs. However, I *already*
know what Solaris does, and exactly what I should be looking for to do
so - your mindshare-gaining "Home User" isn't going to be able to figure
that out by themselves. So by pushing OpenSolaris for the Home User,
all you're going to end up with is "Yuk, this doesn't work anywhere near
as nice as Linux/Windows, it's crap, I'd never use this stuff..."
You'd make *much* more progress sponsoring University deployments, to
get all those CompSci undergrads hooked up to a Solaris server to do
their work (and, be able to tinker around on). Plus, huge percentages
of future IT workers come from undergrad "hangers-on" to the University
IT staffs. If those staffs are working with Solaris all the time, then
they rub off onto the undergrad groupies. And, really, what we're
looking for is Solaris groupies.
For Enterprise mindshare and training, the critical point is
Universities. Not Home Users.
Take a look at the *BSDs - University is where they get the vast
majority of their mindshare from, and newbie's cut their teeth under the
tutelage of the grey-beards. If Solaris can maintain sufficient
penetration there, well, it's a short lecture to a new freshmen hanging
around the computer cluster:
"So, you know that Linux stuff you're using, right? Here, kid, take a
look at my ${SUPER_DUPER_MACHINE} - it runs Solaris. Look at all the new
things it can do!"
"Gee, Mr. Grey-beard, that's COOL! I couldn't do that on my Linux box!
Where do I get one?"
"Well, right now, let me show you install OpenSolaris on your PC, and
then, let's get you an account on this ${Big_Iron}, and..."
*that* is how you get the techno-groupies hooked.
IMHO, making OpenSolaris work reasonably well on standard desktop PCs is
a good idea. Spending a lot of extra developer time on laptops (and
other funky hardware) or jazzing-up the UI really isn't worth the effort
- that time could be better used to improve the feature set that really
separates Solaris from Windows and Linux. Sure, take the low-hanging
fruit from the Linux world, but other than that...
--
Erik Trimble
Java System Support
Mailstop: usca22-123
Phone: x17195
Santa Clara, CA
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org