Reply-To: "Geraldo S."
From: "Geraldo S."
References: <email@example.com> <firstname.lastname@example.org>
Subject: Re: Open Source Consumer Products
Date: Wed, 20 Nov 2002 18:07:24 +0100
X-Newsreader: Microsoft Outlook Express 6.00.2800.1106
NNTP-Posting-Date: 20 Nov 2002 17:08:28 GMT
"Lewin A.R.W. Edwards" wrote in message
> > What do you guys think, would it be wise for a company to produce 'Open
> > Source' consumer products (i.e. the design files and source codes, both
> > firmware, drivers and application programs)?
> I haven't yet seen a message that addresses the (to me) big problem:
> Cost (to the purchaser).
> Consumer electronics are virtually by definition high-volume products.
> In high volumes, it's sensible to use ASICs and custom parts for
> everything. That's why you can buy a brand new 4-head stereo VCR for
> under $70. I can give you engineering diagrams of every component down
> to the last gear, sourcecode for every byte of firmware, and
> schematics, and you still won't be able to build that VCR unless you
> come to me and either buy the five different ASICs I designed (plus a
> myriad proprietary mechanical components), or borrow a silicon fab and
> make a set for yourself, or spend months developing an emulator for
> those ASICs. The emulator will likely cost twenty times as much as the
> ASIC. And getting stereolithographic prototypes of the mechanical
> parts will cost you something like $10,000.
> ... on the other hand, you could just buy one, with a warranty, for
> $69.95 at Wal-Mart, and scavenge these parts. Oh, wait, what was the
> point of this exercise again?
What's the rationale behind this story then? A VCR is a pretty one-sided
story. Most CD players, tuners, receivers, amplifiers and TV's are built
using off the shelf parts. Anyway, except for the mechanism, most things in
a VCR could probably be built using conventional parts. And when you
purchase those parts in large quantities, I'm sure the price would be
comparable, maybe slightly more.
> Open-source only works well when the components required to use the
> free IP are available off the shelf. This is also why the ready
> availability of fairly detailed construction information for nuclear
> weapons is not a serious problem; you can get the information, but you
> can't use it.
> Open-source also only really makes sense when you can expect to be
> able to do something useful by modifying the design. What can you
> really do to improve your Walkman, that you can't do by plugging
> something else into the headphone jack?
Anyway, the morale of my story is that most consumer electronics will in the
future be based on generic parts such as MCU's and DSP's with very few
custom parts (programmable logic such as PLD's could still be used but these
are also COTS parts). And more and more appliciances will use SoC
(system-on-a-chip) parts where most of the functionality will be wrought by
Referring to your Walkman example. I could design a ARM based MP3 player
which, by reprogramming it, could suddenly play Ogg Vorbis files or even
MP3Pro, AAC and MP4 files. I see real value in this. People could add
software which alters the standard playback behavior by adding playlists or
something (would probably depend on the device having a flexible user