Re: Worse is better

Date view Thread view Subject view Author view

From: Ian Andrew Bell (
Date: Wed Mar 15 2000 - 21:03:29 PST

The pattern is extremely simple, if you ask me. In each category,
Marketing / Business Development is the key differentiator. In the
well-known case of VHS/Beta, the VHS camp (led by Panasonic? or was it
Philips?) won over content people early by involving numerous manufacturers
and utilizing their many distribution channels. BETA was, and remains,
entirely within the domain of SONY. SONY bought the Columbia properties so
that they could drive consumer electronics standards using their content
library (like MiniDisc).

Intel vs. 680x0 was similar in that the 8088 series were all channelled via
multiple manufacturers running the same body of content (DOS/Windows),
rather than individual manufacturers running proprietary content (MacOS,

The same thing is currently happening in wireless networking (Bluetooth),
home networking (HomePNA), and others. It's a lesson learned over and over
again, and in my opinion it's one of the fundamental rules I apply in
evaluating technology and standards:

In order to do it right you need one technology proprietor/caretaker, but
in order to grow the business around the technology you must allow others
to participate in the channel.

By that theory IBM actually made a wise decision in enabling cloning,
because although their share of the market dropped from 90% to less than
10%, the overall size of the PC market was enabled to grow a few thousandfold.

Arguably, if they had not, we'd all be relatively equally distributed
across multiple, non-interoperable computing platforms: Atari, MacOS,
Amiga, Coleco, Nintendo, IBM, etc. That in itself wouldn't necessarily be
a bad thing but if development and economics can't be focused on a single
area and a single marketplace, I doubt it's possible that Moore's Law would
have held true.

I also doubt that consumer computing products such as games, etc. would
have evolved as quickly if the talent and the marketplace were so
distributed. And, in a great irony, the consumerization of the internet
would have been slowed by the need for Internet Service Providers, Web
Browser makers, etc. to support more than 2 (in most cases, ONE) platforms.

I guess the lesson then is to create a technology and "push it away",
allowing the industry and the marketplace to self-organize around its
evolution. I can't really think of an example where this hasn't prevailed.


At 10:09 PM 15/03/00 -0600, Jeff Bone wrote:

>For those that don't get the subject: reference, cf:
>For some years now, I've been informally keeping a kind of mental list of
>technically-superior "losering" technologies and their successful competitors.
>I guess I've always hoped to discern patterns in all of this, so that I can
>avoid making those same kinds of mistakes. (Note: I certainly don't think
>there's ONE failure pattern, but perhaps there are metapatterns of failure...)
>Anyone care to contribute to this list, or correct it?
>The format here is better but loser / worse but winner / category. In no
>particular order, a partial list:
>* Beta / VHS / consumer video formats
>* Colecovision / Nintendo 8-bit / second-generation game consoles
>* Inmos' Transputer / SPARC / first widely-deployed RISC architectures
>* Objective C / C++ / workday OOP languages
>* 68XXX / Intel / consumer PC chip architectures
>* Linda / message passing (i.e. CORBA) / distributed computing paradigms
>* NeWS / X11 / distributed window systems
>* ATV / HDTV / advanced television formats
>* Scheme / Pascal / pedagogic programming languages
>* Smalltalk / Java / portable VM-based OOP languages
>* Macintosh / Windows / ditto operating systems
>* NeXT / Windows / ditto
>* peer-to-peer buddy lists / client-server buddy lists / buddy lists
>* Newton / Palm / early runs at PDA fence
>* BeOS / Linux / new-gen OSes (don't start with me! ;-)
>* capabilities / public key crypto / comm. security paradigms
>* Plan 9 / Linux / new gen OSes (again, don't start with me! ;-)
>* Inferno / Java / portable "virtual operating systems" (in some sense)
>* UnterMUD / ad-hoc solutions / distributed MUDs
>* MOO / Mush / MUDs
>* QNX / VxWorks / realtime, embeddable OSes
>* anything else / SMTP / asynchronous message transport
>* anything else / HTTP / synchronous request-reponse
>* anything else / Perl / network "glue" programming
>* CMDA-TDMA-etc / GSM / domestic cell standards
>* Wankel rotary engine / standard combustion engine / motors
>etc. etc.
>Contributions much appreciated.

Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Wed Mar 15 2000 - 21:04:28 PST