It seems, from other postings on this list, that keitai
turnover is fast enough to make a JavaPhone market
pretty quickly, assuming that turnover alone is enough.
And that sounds pretty good.
Yes, as Jason points out, JavaPhone has an installed base of
exactly zero right now. But hey, so did the IBM PC six months
before product release. People developing feverishly for
the PC at that time nevertheless made out like bandits (not least
among these bandits: Microsoft), because the PC had a killer
market-making force behind it: IBM.
i-Mode+Internet is already a platform of sorts. And who
can doubt that DoCoMo, backed by NTT, has market-
maker weight, at least in Japan?
So: JavaPhone, here we come, right?
Not so fast.
Life isn't so simple these days. Sure, IBM rides the
Java bandwagon, but it no longer defines platforms
outside the mainframe business. And Microsoft *does*
define platforms, through their hardware testing labs.
Even that supposed Wintel co-duopolist, Andy Grove,
has said "I pray at the altar of the software gods".
Andy wasn't talking about Larry Ellison or Scott McNealy.
And that might matter here.
Pretend for a moment, though, that Microsoft doesn't matter.
I know that's hard, especially if you're young. But try.
So: JavaPhone, here we come, right?
Not so fast.
i-Mode+Java still might not make for a real de facto standard
platform. Technical factors can mess up the price/performance
picture. Then there's performance: performance of *what*,
specifically? This second point is key.
What really makes a chunk of electronics a platform
rather than a boat anchor or future computer-museum curio
is not price/performance. It's that gotta-have-it application,
delivered in a convenient, reasonably-priced package. That
was the spreadsheet for the 8-bit machines. Word processing
and small databases and low-end presentation graphics for
the 16-bitters. (Remember Lotus 1-2-3?) DTP for the
32-bit machines. (Getting niche-ier and more transient:
Video Toaster for the Amiga, midi for the Atari ST.)
The Internet as "platform"? The browser, piggy-backed
on the modem (a "platform" itself, first for BBSes, the
AOL), on boxes grown sinuous and burly on DTP steroids.
But, notably, on a "platform" we take for granted in the
U.S.: de-regulated telephone services. I'll get back to
this.
And then the browser itself became the first "virtual platform".
That very "virtuality" led to war, though. And through
the fog of war, things are less clear.
Java is a war baby. (Or rather, it became one - it was
in the oxygen tent before the Browser wars. Javascript
is the REAL browser war-baby among programming languages.)
And Java-as-alternative-virtual-platform still hasn't
yielded a killer app. Say what you will (and God knows, in
this industry, people will say anything), a platform is
not software; it's a class of hardware, typically starting
with some ROM chips. Platforms are *firmware*
defined, as hardware, and *application*-defined,
as marketable products.
So this is where I have questions. Let's look at the
the "convenient, reasonably-priced package" aspect.
Java didn't really make it for its original niche (consumer
electronics). It was definitely premature for that. When
Java was being prototyped, people writing for channel-
changers and washing machines were still tweezering 8051
controller code together. No doubt a lot of them still
are. (With all due respect to James Gosling: don't let a
programmer's-editor author and lisp hacker do your next
embedded systems language, OK?) You just couldn't
conveniently, cheaply package Java in handhelds. Virtual
machines throw a layer of software emulation onto the
system, and the system groans. It was a classic
Silicon Valley premature technology ejaculation,
surviving only because of Sun's stable market position.
So Java got retargeted to the browser market,
where it didn't do too well for similar reasons.
Oops, sorry honey. And see below, on applet bloat.
Finally, Java picked up some steam as an applications
language (probably thanks as much to Microsoft as to
Sun, really) and for server-side scripting, where its
bloat didn't matter. Sometimes, though, I wonder if
accurate accounting wouldn't show that Java's "successes"
have never really paid for themselves, but rather continue
to be subsidized out of the defense budgets of the
warring parties.
Only now is Java returning to its original mission. With
ROM footprints for KVM at a vaunted 128K or
thereabouts, I guess it's not going to hike the price of
phones very much, not these days. (In the early 90s, maybe
that would have been a problem, and anyway, without
internet access, how would people have gotten JVM
bytecodes into their mobile phones anyway?) 128K
is a good bare-bones ROM size for an entry-level
GUI-driven platform. The first Macintosh was in that
range. And Moore's law has relentless seems to have
pushed the needed processing power into palmable units.
Price for handsets is still an issue, though. Is DoCoMo talking
about KVM/i-mode as a special, costly feature, or as the
default in future models?
The latter case is instant platform win for Java - people will get
JavaPhones whether they know (or care) at all. Their hearts and
minds can follow later. If a JavaPhone costs a lot more, though....
let's cross our fingers and hope DoCoMo is reasonable.
So: JavaPhone, here we come, right?
Oh impatient youth.
It pays to look at *all* the past risk factors.
One thing that killed Java for client-side web apps was that applets
were quite bloated. They took forever to arrive before they started
doing whatever stupid net trick somebody thought of last week.
There was also a browser-user abreaction to "my machine
being taken over." That might have been more a matter of
not having the right OS, in the case of Windows at the timem
but it better not happen again.
Finally, beyond annoyance, there was the shrug. Where
was the killer app? Animation? Javascript did that
pretty well, without the sickening delays and crash-panic
butterflies in the stomach.
Let's look at each of these in turn, most important
first.
The user shrugged, because applets couldn't do much?
Yes, JVM security features meant that applets couldn't do
anything useful with stuff on your desktop. Presumably
this won't be an issue with mobile phones, with
"Application Service Provider" becoming an apparently-
viable business model. So the picture here is brighter
than it was for desktop-browser applets, though still
not certain. Let's get back to this in a minute.
The "gimme-my-computer-back" objection is probably not
a serious one for phones, since ANYthing you do on a tiny
screen naturally takes over the whole handset, at least in
user perceptions. People are used to keitais as phones, first,
and as i-Mode browsers second, rather than as a personal
computers. In any case, I expect embedded-systems folk to
do a better job at multitasking OSes than Microsoft was
doing with Windows ca 95.
That leaves the question of applet bloat and transmission
time. And here I must confess real uncertainty. I've looked
around the Web a little, but so far I don't have a solid
feeling for how quickly something like a game would
download to a JavaPhone in Japan. In my brief
flirtation with Java programming, I never understood why a
"Hello World" program didn't download practically instantly -
do you always have to drag along the whole Java class library
every time? Why? Would this be true for KVM-targeted
programs as well?
Let's assume that Java has cleaned up its act on this.
So: JavaPhone here we come!
Right?
You really aren't keeping track, are you?
Remember, it's not enough to be cheap and fast. Beauty is
in the eye of the beholder. Quality is what the customer says
it is. And what the customer wants is the killer app.
What can Java really give you? What will be its killer app on
a mobile phone?
If DoCoMo can make "Java = secure, economical, easy, mobile
on-line product ordering" an equation in Japanese consumer's
minds, you might really have something. Unfortunately, though,
you can already get that in most people's minds with a human
order-taking operator. It would be cheaper? Yes, but only if
there are layoffs. And this is Japan, and more specifically, NTT,
we're talking about here. Is huge, unionized, amakudari-ridden
NTT going to be one of Japan's biggest job killers? Hmm...
OK, then, a fad game could prime the market, especially
if it was networked adequately. In any case, though,
quick download times will matter, I suspect. And that's
a big question in my mind. Memory sticks might be a
workaround for this, though. Also, there's already a
whole electronics industry segment that does good pocket
games on somewhat larger screens. You'd really have
to play up the networking aspect to overcome JavaPhone's
screen-real-estate disadvantage vis-a-vis Bandai.
You'd also have to make sure people don't share games
too promiscuously. ROM has been the traditional
way to get paid in that market. OK, but then, what
do you need with Java?
Well, build it and they will come, and all that. Cross your
fingers hard. Let's say the killer app embryo is in the
bottle right now, awaiting only maturation and decanting.
I mean, who can predict these things? Who predicted
Tamagotchi? Not even Bandai!
So: JavaPhone here we come, right?
Um, maybe.
Remember way back there, where I asked you to
pretend something? OK, Dorothy, stop clicking
the ruby slipper's heels together and open your
eyes. You're still in Oz. This would be kind of
trippy and cool, except that Bill Gates is reading you
this Friday the 13th bed-time story. WWW might
soon stand for a re-animated Wicked Witch of
the West. He wants it. He has said so. He has
never said otherwise. Some people (Robert X.
Cringely not least) suspect that MS will win on
appeal. And a break-up might not even matter.
I mean, we're on our second supposed break-up
of NTT, and what difference has *that* made?
So, let's not forget politics: Java is a war baby. And that
war smolders on, in the courts and in the markets. The
DoCoMo/Java announcement, even those prototype units,
might be just NTT DoCoMo dangling a plum just out of
Microsoft's reach, holding out for a good price while
threatening - quite credibly - to feed the plum to somebody else.
The price for NTT only has to be good enough to exceed
the legal costs of DoCoMo's backpedaling on JavaPhone
promises, and NTT can milk the Japanese public to pay for
*that* for decades yet. Microsoft stockholders would probably
just swoon in ecstasy at the news, no matter what the temporary
hit would be to the bottom line. Microsoft a monopoly? It
doesn't hold a candle to NTT, "break-up" or no.
Microsoft doesn't own much copper, or fiber. Not for
lack of trying. In Japan, it can't get a grip. Telephony
is a platform where NTT holds the firmware. Firmly.
After all, whatever happened to the great Softbank-TEPCO-
Microsoft Greater Tokyo Internet Co-Prosperity Sphere?
Fiber to the home, over a decade after it was first promised
to the Japanese public?
Remember that deal? Well, execution of it is on indefinite hold
Why? "Technical problems."
Oh, I believe that. Yeah, yeah, right.
This is Japan. Here, you see, here it's not The Last Mile,
but The Last Meter, and who owns the crossing rights for
*that* moat? NTT. There's no done deal without them.
Even Microsoft has to pay the toll. And no doubt, Microsoft
will pay that toll, when the current distractions subside. And
some other tolls besides.
So yes, it's possible: an eleventh-hour pullback on the JavaPhone
promise, and then we all have to start learning VBi: Visual Basic
for i-Mode. [*]
I wouldn't know how to bet. If you don't already know Java,
though, learning it will probably stand you in good stead
*somewhere*, for a good while yet. Though it might have
to be in some country other than Japan, if you're committed
to wireless web development. And only on the server side.
So say I, and I've been right. Once in a while.
Michael Turner
www.idiom.com/~turner
leap@gol.com
[*] VBi could happen anyway - KVM, like JVM, doesn't
care what language the source was written in, so long as
the bytecodes are right. What I'm really talking about is
a non-JVM-derived execution architecture on mobile
phones.
----- Original Message -----
From: <jason.c.freedman@ac.com>
To: <keitai-l@appelsiini.net>
Sent: Monday, July 17, 2000 4:18 PM
Subject: (keitai-l) Re: i-mode vs net
>
> I don't know, I guess. Anyone know how fast the turnover really is in
> phones?
>
> Personally I try to keep my phone as long as possible, rather than
spending
> over 15000 yen regularly.
>
> My basic point was that there are 7Million iMode users and exactly none
> currently support Java. So you really have an installed base of 0.
>
> Jason
>
>
> keitai-l-bounce@appelsiini.net
> 07/14/2000 07:47 AM MST
> Please respond to keitai-l@appelsiini.net
>
> To: keitai-l@appelsiini.net
> cc:
> Subject: (keitai-l) Re: i-mode vs net
>
>
> Would it really be crazy? As the turnover in keitai's is quite high
already
> due to styling, I would think that Java support will be rather rapidly
> deployed....
>
> jason.c.freedman@ac.com wrote:
>
> > yes on all accounts. Although Java will be supported in the near future
> on
> > new iMode phones (sounds like you'd be crazy to use Java now when people
> > would need to upgrade hardware to support this feature (at a few hundred
> > dollars a pop)). Actually I think that this is a big issue with cell
> > phones. Unlike upgrades to browsers where MS or Netscape can just
> support
> > a new feature by downloading a new browser, with these phones you need
to
> > upgrade hardware.
> >
> > Changes the whole cost model for upgrades.
> >
> > Jason
> >
> > keitai-l-bounce@appelsiini.net
> > 07/14/2000 01:29 PM CET
> > Please respond to keitai-l@appelsiini.net
> >
> > To: <keitai-l@appelsiini.net>
> > cc:
> > Subject: (keitai-l) Re: i-mode vs net
> >
> > David
> >
> > If I'm right myself, I don't think i-mode supports javascript bacause
the
> > system is based on HTML 1.0 at the moment and also the memory allocation
> > per
> > page of data is only 5k, which puts a stretch on what content, smart or
> > otherwise, which can be incorporated.
> >
> > Somebody correct me if I'm wrong, but thats what I know anyway!
> >
> > Matt
> > ----- Original Message -----
> > From: "M. David" <davidm1@hotmail.com>
> > To: <keitai-l@appelsiini.net>
> > Sent: Friday, July 14, 2000 1:17 PM
> > Subject: (keitai-l) i-mode vs net
> >
> > > I didn't see it on the i-mode pages, so just to check, i-mode does not
> > > support
> > > java script for now right?
> > > This might be a good way to serve stuff to net browsers (MS, Netscape)
> > vs.
> > > i-mode.
> > >
> > > David
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> ________________________________________________________________________
> > > Get Your Private, Free E-mail from MSN Hotmail at
> http://www.hotmail.com
> > >
> > >
> > >
>
> --
> Eric Hildum
> Director, CTC Business Unit
> Itochu Technology Inc.
> 3100 Patrick Henry Drive
> Santa Clara, CA 95054-1850 USA
> Tel: +1-408-653-2818
> Fax: +1-408-727-4619
>
>
> -- Binary/unsupported file stripped by Listar --
> -- Type: text/x-vcard
> -- File: Eric_Hildum.vcf
> -- Desc: Card for Eric Hildum
>
>
>
>
>
>
>
Received on Tue Jul 18 10:15:02 2000