Date: Fri, 17 May 2002 16:17:07 -0700 From: Terry Lambert <tlambert2@mindspring.com> To: Giorgos Keramidas <keramida@ceid.upatras.gr> Cc: Miguel Mendez <flynn@energyhq.homeip.net>, freebsd-chat@FreeBSD.ORG Subject: Re: The road ahead? Message-ID: <3CE58F73.1A7F50AF@mindspring.com> References: <20020516004909.A9808@daemon.tisys.org> <20020516151801.A47974@energyhq.homeip.net> <20020516172853.A7750@daemon.tisys.org> <3CE40759.7C584101@mindspring.com> <20020516220616.A51305@energyhq.homeip.net> <3CE43D08.1FDBF0A3@mindspring.com> <20020517163624.GB9697@hades.hell.gr>
next in thread | previous in thread | raw e-mail | index | archive | help
Giorgos Keramidas wrote:
> > > 2) Remove *all* man pages and online documentation
> >
> > B) #2 is defacto done; compare the info and man pages for gcc
> > or gld, some day
>
> But the info pages can be installed for both programs, as part of the
> system installation. What was your point again?
Humans obtain information by selecting a Schelling point, and
attempting to obtain information there.
Despite the attempts by the GNU crowds, the "info" format is
not the preemminent Schelling point for UNIX systems, man pages
are.
Therefore, any documentation which does not exist in the man pages
defacto does not exist.
If you want confirmation, I suggest you search for "Schelling" in
your favorite search engine.
While you are at it, ask yourself why you went to that search
engine, instead of any one of dozens of others: why does that
search engine itself represent a Schelling point to you? How
is it that you knew to go there, instead of somewhere else?
If you can answer these questions, you will be more than half
the way to realizing why, in the scheme of the Internet "bubble",
almost all of the "portal plays" have failed to result in the
traditional "brand loyalty" that marketers of traditional
products rely upon.
Or to put it another way: you will know why people's loyalty
won't stay bought. I'll give you a hint: there is a disproportion
between barriers to entry and barriers to exit on intangibles
which is not there with tangible consumables, which tend to keep
your "loyalty" by remaining there until they are "used up". Now
compare this with "my default home page".
> > > 3) Make the system horrendously obfuscated
> > C) #3 is defacto done
>
> As a documentation writer, I am terribly offended :P
>
> This also seems like you're suggesting that the FreeBSD developers do
> their best to obfuscate the programs and code they write. In a system
> like FreeBSD that tries to follow standards like SUSv3 or similar,
> this is not possible and not wanted. I think you're doing harm to
> FreeBSD in this. Harm that is not needed, and certainly not deserved
> by all the people who are trying to make FreeBSD an operating system
> and environment easier to us :-/
As a user, I'm terribly offended that any documentation at
all is required to use any software in the first place. ;P.
The standards which are conformed to were designed within
the same principles framework that resulted in software which
was hard to use without documentation in the first place.
The epitome of good technology is the device with but a single
button on it, which doesn't look fearfully dangerous to press,
and which is a "do what I wanted done" button.
A bad example from personal experience was the Whistle InterJet;
it had a lot of buttons on the front, which not only raised the
overall cost, but implied configuration and exposed complexity
(there was plenty of both, in fact).
People bought InterJets to connect their small businesses to the
Internet. They were less painful than the precursor, but they
could hardly be said to represent an epitome of good technology.
A much better paradigm would have been a single round green
button on the front, wich connected your small office to the
Internet.
As the engineers who developed and invented the InterJet, we
weren't allowed to build the product that people needed.
Instead, we had a lot of "input" from people who had already
decided that connecting a small office to the Internet was a
complex task, and that the product was there to serve a user
interface -- not a customer need -- and that the purpose of
the user interface was to manage complexity, using abstract
representational geometries that were somehow superiour,
having passed through a PhD who didn't think like an
"uneducated end user" in the first place.
So what would be the epitome of design for such a device?
1) If I press the green button, it connects my office to
the Internet.
2) If I hold the green button down long enough, the machine
shuts off.
In other words, the "big green button" is an ATX power switch.
There are a couple of intermediate technical obstacles, but they
are not completely unscalable. Most of them have to do with a
steadfast refusal to manage the "link.local" problem, the DHCP
server grab that Microsoft staged, and the lack of support for
service location in Microsoft Operating systems causing an
unnecessarily increased complication for hardware configuration.
But of course, actually *fixing* problems, instead of coming up
with some Rube Goldbergian scheme for working around them, does
not present a barrier to entry to competitors, does it? And for
some reason, we don't trust ourselves to step out in front in
the first place -- *and then be able to stay out in front of the
parade*.
There's actually a German company building an InterJet-like
device. You can buy one today, off the shelf, from Fry's
electronics, for around US$400. It has a keypad, like the
InterJet, and it has an LCD panel (at the bottom, rather than
the top).
Actually, I would call such devices in a small business "file
cabinet top" boxes: they fill the ecological niche in a small
business that in a home would be filled by set-top boxes. So
the location of the LCD screen on the bottom of the front panel
rather than the top is a good design decision -- as far as it
does.
But the German device embodies the same assumptions which led to IBM
eventually shutting down the IBM GSB Division (the former Whistle
Communications), when they didn't get out of it what they expected
to get out of it, when they bought it. It makes the assumptions
that an LCD panel is necessary at all, and that a keypad is necessary
at all, because, well, that's the assumption made in the design of
the InterJet, and they aren't really breaking new ground, so much as
they are copying a design.
To be fair, IBM's expectations were unreasonable to start with,
and those expectations were set by people wanting an exit for
money, rather than by people who wanted to convey a proper sense
of the product: they lacked faith, and so could not convey faith
to the IBM people. IBM thought they were buying a human factors
design which could succeed where their own designs had not, in a
market that IBM wanted to capitalize. If Whistle hadn't tried to
make the sale on human factors, the outcome could have been much
more positive. IBM's own "dirty little secrets" with regard to
business practices didn't help, and the overall result was a
cascade failure. No need to go into them now.
> > > 4) Become rich by selling FreeBSD manuals at $100 a pop :-)
> >
> > D) #4 Doesn't work in the presence of a monopolist with a
> > competing product who doesn't pull the same pig-trick
>
> So making a profit out of FreeBSD by using those same ``dirty tricks''
> mentioned above, can not be done. You're effectively stating that this
> is already done in B) and then later argue that it can't be done.
>
> Now I'm thoroughly confused.
FreeBSD is not a first mover. Such a strategy will work, if you
have first mover advantage. FreeBSD has no such advantage.
Note that a "first mover" is not necessarily the same as "first
to market". This is a common assumption, but it is wrong. The
Microsoft strategy, and, in general, the Japanese corporate
consumer electronics strategy are both incredibly successful,
and they are built on the foundation of being "first to be second"
for any given market. These companies rarely innovate, which is
why slashdot and other technical forums tend to find Microsoft's
protestations that they would be "unable to innovate in the future"
so laughable.
In fact, we can read Microsoft's successes and failures on the
basis of projects which do and do not fit this philosophy. Since
its beginning, Microsoft has failed in every "first to market" case,
and succeeded in *almost* every "first to be second" case. "Money",
"Word", "Excel", "Windows", "DOS", ...all but a few of their products,
in fact, follow this model closely. The notable exceptions (like the
Microsoft Mouse) are actually supporting products, more than they are
products in their own right: they are infrastructure. Novell did the
same thing when they bought a network card manufacturer, and dropped
the hardware prices low enough that network cards hit the point where
they became commodities, and it was possible for them to be used as
"paving material" in support of other products, which assumed LAN
connectivity for machines within a business: Novell's real target
market.
The FreeBSD approach fails to provide for a "Chasm Crossing Strategy",
as Geoffrey Moore (formerly of Regis McKenna, now of the Chasm Group)
would say. It has OK appeal to early adopters (classic technophiles
like it -- this is "geek appeal"), but it has yet to have any real
value in the mainstream.
While "Yahoo" is widely regarded as a success story for FreeBSD, it's
hardly the "reference account" that can be used to leverage "sales"
(for Free software, this is really a measure of market penetration
and mindshare, not revenue) in other market segments or categories.
Yahoo is not "GM" or "American Express" or "Delta Airlines", etc..
The issue is one of market segmentation, and then choice of the
correct segment in which to compete, such that you don't face the
barriers you will face if you attempt to shotgun all over the map.
FreeBSD has failed to service a market segment in which it has an
advantage like this -- a marget segment in which it is not up
against monopolistic forces. Instead, it shotguns, and tries to
be all things to all people, or, worse, declares that any market
in which it has failed to arrive at a success strategy is "not
our target market" (the reductio ad absurdum of that argument is
that, eventually, your "target" is pared down to one fourteen
year old "Barbie" owner in Des Moines, Iowa).
Right now, the initial barrier to entry is installation. Windows
comes pre-installed on much of today's computer hardware.
For that computer hardware which is being put to applications
where Windows is ill-suited, FreeBSD is not built to solve the
problems in that problem domain, out of the box.
In addition, the installation process is cumbersome, and has not
been abstracted sufficiently.
There is no single "big green button" for any particular segment
which FreeBSD "hopes to serve". And there is no direction for that
"hope".
Linux is prety much in the same boat; it has a public goal ("Destroy
Microsoft"), which a lot of jaded geeks can get behind, but it has
not arrived. It, too, is still on the "early adopter" side of the
"chasm" in the technology adoption lifecycle.
-- Terry
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3CE58F73.1A7F50AF>