Alan Cox wrote:
> > Hmmm. I take it from the tenor of your response that your input would be
> > for others on this mailing list (linux-usb) to NOT comply with Jeroen's
> > request?
>
> Each to his own viewpoint. I make no suggestions what others should do.
As someone with extensive experience in hardware and software
development, I feel impelled to toss in my observations:
I'd leave the politics (or religion) out of it. I doubt that a
collection of emails from readers of this mailing list would be
significant in corporate decisions. Pointers to web resources and
reports about the growth of the use of Linux can make a
difference. While Linux is a long way from matching the pentration of
Microsoft (and, I suspect, of companies like Apple), the estimated user
base is certainly large enough to be a serious commercial
consideration for any manufacturer.
Decisions such as whether or not to release engineering specs are
based on many things, and it is naive to think that it's a simple 'us
vs. them') argument. Among the considerations are:
1) if you manufacture something which you feel, rightly or wrongly,
has interesting engineering features, then you may, as a company,
wish to keep those details controlled. You don't want someone else
in the same business saying - 'Hey, what a good idea, we'll steal^W
adapt that for our next release'. And no, I don't wish to rehash
the argument that your competitors will be able to reverse engineer
drivers, put logic analysers on the hardware, etc.
2) There can well be arguments that controlled release of software
support ensures that a complex product will perform well with
respect to its competitors. A bad driver will leave people saying
that it's a bad product, which from the customer point of view, it
is. So, you want to ensure that anyone producing drivers for
widespread systems produce ones which take advantage of whatever
special features your product has and cope well with any
idiosyncracies which may be present.
3) which leads to - you want to support a limited set of developers
with sufficient resources that they make a successful software
product. Free disclosure leaves you to either close your support
and hope people understand the documentation, which may lead to
less than favourable products, or to attempt to support everyone
who gets their hands on the specs and fancies their ability to
develop from them. It's not easy to sort out unknown developers
into 'clued up and worth supporting' and 'timewaster'.
4) Some manufacturers may well have a product which is an initial
release of a product they are developing and expanding. With an
NDA, you can tell the devloper things like 'be sure to allow for
the following in your drivers, because we expect to have a new
model using this currently unused feature by Q3 2000'. And those
are indications of where you are going that companies have every
reason to want to keep very closed. Time to market with new
features is life and death in this industry and a head start to
your competitor on the direction you are going can make or break a
huge development
All that said in defense of NDAs, there are very good reasons for a
company to consider going open source. There is a demonstrable and
significant market in open source OSes, sufficient that even Microsoft
has put some effort into considering Linux and its possible impact on
their business.
And, it is possible to demonstrate the efficacy of the open source
development model in producing software which is demonstrably better
and faster to market than conventional closed development within
companies.
If anyone wants to convince Philips or any other company to adopt an
open policy, the above two points are the ones to use. Email from a
few hundred or even a few thousand users will be seen as the pleadings
of a small group of fanatics. Demonstration that these small groups
are producing software that is becoming a significant force in the
world at large and is producing better software than the traditional
shops is an effective persuader. Adding in pointers to others who have
gone this route - Corel, Word Perfect, Oracle, etc. helps to
re-inforce this argument.
Disabuse yourselves of the notion that everyone who wears suits and
works in an ISO9000 environment (or whatever that bloody nuisance is
numbered) is stupid or shortsighted. Ask your self why they take the
stance they do and what you can do influence that judgement.
--
Jim Segrave [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
