Date: Tue, 23 Mar 1999 13:36:08 -0600 (CST) From: Jonathan Lemon <jlemon@americantv.com> To: brett@lariat.org, freebsd-advocacy@freebsd.org Subject: Re: FreeBSD emulation for linux Message-ID: <199903231936.NAA10667@free.pcs> In-Reply-To: <local.mail.freebsd-advocacy/19990323054252$ca17@fish.pcs> References: <local.mail.freebsd-advocacy/4.2.0.32.19990322194937.03ee4600@localhost>
next in thread | previous in thread | raw e-mail | index | archive | help
Okay, I just have to stick my oar in here and add my two bucks.
A while back, I might have just beaten on Brett, but I've been
taking a couple of (gasp) Marketing classes at the University,
and they've been umm, interesting. Not to mention that they
contain many of the same themes that Brett's been pounding on, in
his own way.
In article <local.mail.freebsd-advocacy/19990323054252$ca17@fish.pcs>,
Brett Glass <brett@lariat.org> wrote:
>At 10:07 PM 3/22/99 -0700, Brett Taylor wrote:
>>You've suggested the following things that I remember recently:
>>
>> - that the Linux emulator has been bad for FreeBSD
>
>It has. Unequivocally. Oh, a few people have found it useful as
>a stopgap, but it is the ultimate reason for developers NEVER to
>target the platform and is thus horribly and irreparably
>destructive.
>
>Has no one here learned from OS/2? Read my lips: emulating a
>more popular platform is suicide. What part of that sentence
>don't you understand?
On one hand, Brett is correct; why should businesses target FreeBSD
when they can target Linux, and get both? This effectively records
FreeBSD emulation sales as Linux sales, and contributes toward the
Linux mindshare, which is what I perceive Brett is mainly concerned
about.
Step back a moment, and realize that businesses exist for one reason
only: to make money. (paying developer salaries and buying nice
toys is just an incidental benefit, not the primary goal of the company)
In order to make money, they need to sell their product, and at the
same time, reduce their overhead. Why would a company want to target
the FreeBSD market, when by targeting the Linux market, they also
incorporate FreeBSD? To a marketer, it's a fairly clear ROI (or bang
for the buck) decision.
Now, if we took our Linux emulation out of the picture, the decision above
*DOES NOT CHANGE*. FreeBSD still does not have the numbers to register
on a marketer's radar screen. All it would do is reduce our pool of
available software _and_ put up a larger barrier to gaining new users.
I view the latter as being more significant, in order for FreeBSD to
grow, we need more users, so we should be focus on making the transition
for new users as easy as possible. One way of doing this is to guarantee
software compatability (yes, it will still run your Linux software).
I agree that we need to attempt to raise the FreeBSD mindshare,
however, a FreeBSD emulator for Linux does not make a compelling
business case at all.
Marketers are already analyzing the _Linux_ community, and targeting
their software towards one distribution or another. Red Hat seems to
be currently "winning" the marketing competition in the Linux community.
If the other Linux distributions (Caldera, for example) can't convince
companies to make a binary that will run seamlessly on their Linux
variant, why should FreeBSD be able to?
Microsoft has the viewpoint of "Embrace, Extend, and Extinguish". For
FreeBSD, emulation is the portion of "embrace". Now, we need to move
on to "extend" - provide a _MORE_COMPELLING_ reason for people to use
FreeBSD. Note that I still get lines in my HTTP logfiles that say:
"Mozilla/4.0 (compatible; MSIE 4.01; Windows 95)
Microsoft started by "emulating" Netscape, providing compatability, and
then "extending" the browser so people used it due to it's new features.
My take on OS/2 is that it failed not because it provided emulation,
but because it did not provide enough added (new) benefits that people
wanted, fast enough.
Marketers tend to segment the market, and position a product for a
specific (usage x situation). FreeBSD appears to be currently targeted
at the High Volume, High Availability, Server Market, to be used by
Skilled Administrators. In this market, there is no compelling need
for Linux emulation, and the products are available as native FBSD
binaries (cf: HighWind software).
What Brett seems to be (indirectly) suggesting is that FreeBSD add
a new segment, (casual user x desktop), since this is where the Linux
emulation seems to be primarily used. While this might raise the
mindshare of FBSD, I don't see this as a good fit for FreeBSD's current
strengths, as it moves us away from the server market, and our existing
targeted user base.
This is not to say that we shouldn't address this segment at all, if
only as a "flanking" move, to prevent erosion of the current market
base. However, my opinion is that FreeBSD would be better served by
focusing on a slightly different, but still related segment:
(Unskilled Admins x High End Servers).
As an ancedote (and I'm sure I'm not alone here), I just installed
FBSD onto a customer's high-end Compaq machine, for duty as a web
server. The customer had ample Novell experience, but has very little
UNIX experience, but was willing to learn (and had a copy of Greg's
"The Complete FreeBSD"). In short, probably an ideal convert for us.
After I had finished the installation, the first question that he asked
was "Are you going to install X-Windows on the machine too?".
Lesson 1: The Customer Is Always Right.
Lesson 2: Listen to your Customer.
Whether they are technically correct or not, this underscores something
that should be looked at: "Ease of Use Out Of The Box". Something like
the "FreeBSD Desktop", working Gnome ports, and an automatically
installed "FreeBSD User Environment", will probably help us to win more
users.
>Not what I advocated. I noted that ports for 2.2.8-RELEASE were no
>longer being maintained less than two months after it shipped. What
>I advocated was that the ports tree support a release for at least
>6 months after it shipped. Otherwise, every release is automatically
>an "orphan." This is unprofessional AND bad PR-wise. WHATEVER has
>to be done to prevent that should be done. Period. Take your choice
>of technical approaches.
This brings up what I feel might be an organizational weakness with
the FreeBSD side of things: there appears to be no "company" that
pushes FreeBSD.
Not Walnut Creek; they are not in the FBSD business, they are in the
CDROM business. Not Yahoo; they are not in the OS business either.
Not FreeBSD, Inc.; this is a non-profit entity which doesn't seem to
be interested in being a business partner.
I imagine that one reason why RH got Linux onto IBM servers was that
they approached IBM with a business case. They probably demonstrated
that there was a market, offered channel support (advertising, offloading
technical support, free media and documentation, etc). In short, they
provided IBM with another potential market (Unix servers) while reducing
the risk to IBM by absorbing the support costs.
On the other hand, an advocate's recent request to IBM to "consider
bundling FreeBSD" was met with a polite request for a business case.
In my view, IBM doesn't really care what OS they sell, as long as
they can make money out of it. Present a good business case as to
why they should sell FBSD (complete with target market share and
expected sales volume), and they will probably do that. However, I
don't quite see the support infrastructure in place yet, so this is
an area where FBSD isn't currently competitive in.
>It is not even POSSIBLE unless an emulator exists.
>
>>Companies will say "sure there's this FreeBSD emulator for Linux, but they
>>have roughly 1/6th the number of installations _and_ can emulate Linux.
>
>The latter is the bigger problem. Again, Linux emulation has been a huge
>blow to the platform. At some point, after Linux emulates FreeBSD and FreeBSD
>wins native ports, FreeBSD's emulation of Linux should be deprecated.
But I fail to see how a "Linux emulation of FBSD" will provide a
business case. "Why should we write to this API, when it's not the
native market, has resistance in the Linux community, and is opposed
by our channel partners?"
I think a better approach would be to focus on the number of BSD
installations, until at some point the equation shifts - we reach
a certain critical mass at which marketers take notice; at that point,
perhaps, the Linux emulator will start falling into disfavor.
I would suggest that a better "bang for the buck", as well as possibly
getting more technical interest, _and_ have the possibilty of getting
more PR, would be a "FreeBSD upgrade" of a Linux system. Just imagine,
taking a FBSD kernel and a few support files, moving it to /vmlinux (or
whaetever) on an existing Linux system, and then rebooting. Presto!
The user now has a working system without changing anything else!
This would provide:
1) a fairly risk-free method of experimenting with FreeBSD,
2) a method of converting users (hey, this system works _faster_ now!)
3) coolness factor (necessary for early-adopters in the business cycle)
But would also risk:
1) tarnishing FBSD's reputation if it doesn't work, or doesn't do
better than Linux (hey, my device is not supported now!)
2) alieniating the Linux community
3) draining developer resources possibly needed elsewhere.
Gah, I think I've rambled on long enough, and I'm not sure where this
is going anymore.
Brett, my take on this is that some people would be more than
willing to support you, and help with advocacy, but the feeling
is that you need to identify or segment your target market a
little better, in order to provide some more focus.
--
Jonathan
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-advocacy" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199903231936.NAA10667>