IMHO, the PC did not succeed because of its 'de minimis' standard. If anything, this caused many problems in the late '80s/early '90s due to the proliferation of many different and incompatible bus standards (EISA, MCA, VL-bus) and software issues like the need to provide different graphics drivers for different applications. Only the 'standardising' forces of Windows and PCI really propelled the growth of PC sales. Even now, with Linux starting to proliferate on different architectures, and powerful PDAs based on MIPS and ARM architectures, it is not entirely clear how the market will go in the future. I agree that Apple's restrictive policy was a shot in the foot, but the company is far from dead, and most of the enthusiastic Apple users I know wouldn't touch a PC with a 10 foot barge pole.
802.11 is a wireless LAN, occupying the bottom 2 layers of the traditional protocol stack, therefore existing networking applications can use 802.11 now. However, it does not really fall into the same category Bluetooth is put under, i.e. a PAN (personal area network). Bluetooth is essentially an infra-red replacer; just looking at the protocol/profiles gives this away (RFCOMM, OBEX). It has the fundamental advantage over IR of not requiring line of sight for transactions, therefore opening up possibilities which were not really available to IR. It is these possibilities which indeed are emerging and will undoubtedly contribute to the success of Bluetooth. The profiles are needed to provide some sort of standardisation to ensure interoperability is not a problem. Establishing a voice connection between two Bluetooth devices requires the same degree of standardisation which exists for, say, using FTP over 802.11. Interoperability is a key issue which must be present for Bluetooth to succeed. I agree that the implementation of TCP/IP over Bluetooth is a bit clumsy, but the pivotal point was chosen to be the serial link (i.e. RFCOMM), which is, after all, what the majority of people still use to connect to the Internet, therefore it seems appropriate enough to go along this route. And as Ronald Kyker has said, the MAC and physical layers are different, and do not map directly to IP packet structure unlike 802.11 datagrams.
ISDN was a quantum leap in technology for telcos who even now haven't got to grips with it at the edge, although seem to have adopted its 'big brother' ATM. ISDN is in fact more like 802.11 as a standard. The difference is that telcos did not have the end-to-end applications except of course voice, as that was primarily their business; only now are data and PSTN networks starting to merge properly, and the true vision of ISDN may well proliferate, albeit over DSL.
I would say Bluetooth is more like GSM, and look at the success that has been. True, it took a while to get there, but it's there - even in the USA.
So all this direct comparison between 802.11 and Bluetooth doesn't really make sense. Yes, Bluetooth could operate as a LAN with TCP/IP running over it. Yes, 802.11 could operate as a PAN, with applications to provide voice transfer, service discovery etc. But they are fundamentally different standards. And FWIW, let's have Bluetooth in the ISM band, and get the faster 802.11b going in the 5GHz band.
Note these are entirely my own views and do not necessarily reflect the opinions of Jennic Ltd.
--- --- --- --- --- --- --- --- ---
Replying to:
Though I used the word "Stalinist" for effect, the real issue is the
relative openness of the two approaches, which I emphasized.
IBM created a PC standard by creating a de minimis standard - essentially
the functionality of its BIOS and the processor/bus architecture
only. They left the market wide open for innovation at the apps level,
encouraging an ISV market, and at the hardware level, by allowing anyone
who could to create cards and peripherals. Compaq's entry effectively
blocked IBM's later attempt to try to squeeze their partners on margins and
by withholding strategic information.
Apple, on the other hand, took a much more controlling approach. Their
evangelists told app vendors what they could and couldn't do. Their
hardware designers discouraged add-on hardware (in many cases by competing
directly against their most innovative partners).
The Bluetooth vendors (esp. the cellphone centric ones that control the
standard) are determined not to repeat IBM's "mistake" in loosening control
of the industry. So they want to keep the ability to drive cellular bills
higher, block ability to merge Bluetooth networks smoothly into the
Internet fabric, etc. That's part of where their financial weight is
being applied. I.e., they are being like Apple.
Now we've all seen where Apple and IBM families of computing are
today. Apple is an annoying mosquito that Microsoft tolerates because it
prevents too much antitrust attention. A small cadre of devotees love
Apple for its design and its quirks (even its technical superiority in some
dimensions). It's the Jaguar or Corvette of the PC industry.
But innovation moves where innovative users are free to innovate, to
experiment. This is also the problem with the "walled gardens" of WAP
phones. The less advanced but open SMS capabilities seem to be much more
productive of user-centered innovations.
ISDN (the original "Intelligent Network" from the phone companies) was very
much like Bluetooth. Rather than being content to provide a digital path
between users and innovative services, the ISDN plan was focused on the
operators having complete control over what services the user saw. There
was some hardware innovation, but that was limited by the vision the ISDN
planners (get why I see Stalin's 5 year plans here?) had of what services
and apps users wanted or would buy. And a tradition of treating all
customers identically ("universal service") meant that diversity and
innovation was discounted. ISDN finally reached the market primarily in
the form of a low-priced way to get 128kb/s digital service to homes (which
had nothing to do with intelligent phones).
And the platform that attracts the highest rate of innovative *use* is the
one that wins. Innovative *use* is not glitzy visions or fancy plans -
it's customer driven, as they vote with their wallets among richly diverse
alternatives. And that means setting standards only at the lowest layers,
not the top-to-bottom "profiles" that define applications inflexibly.
Perhaps the biggest problem/symptom is the "TCP/IP" profile. Instead of
treating the Bluetooth network as the packet network it is, the profile
layers this on top of a simulated serial link. The idea seems to be to
"backhaul" the TCP connection to a single ISP somewhere. Contrast this
with the architecture of TCP and IP over Wi-Fi networks, which treats the
Wi-Fi network like a packet network. Since the TCP/IP world is rich with
*user* driven innovation, treating TCP/IP as a first-class citizen would
jumpstart Bluetooth. But that is the last thing that 2G, 2.5G, and 3G
operators want - because it might cut their access to "service revenue",
they'd be relegated to competing in commodity services like moving bits
fast. And the Ericsson crowd wants to keep its operator customers happy,
by letting them define when apps roll out and who can supply them. (Nokia,
I suspect, is much more interested in user innovation, which is why they
have invested so heavily in Wi-Fi and 802.11a).
I suspect that if Bluetooth does succeed it will be like ISDN did - the
radios will be exploited, but by renegade innovators who use it as a packet
network, violating the 1.1 standards. Microsoft may well be going in that
direction, if I read backchannel communications correctly. And "low power
802.11" is feasible using the same tricks that Bluetooth has used, but
without the burden of the "planners" who have glommed onto Bluetooth.
In 2000, 802.11 equipment was a business with over $1,000,000,000 in
revenues, and highly profitable. It's way ahead, and Bluetooth has missed
its window.