[9fans] re: license terms?

[Douglas Fraser dwfraser@lucent.com]

Anssi Porttikivi wrote:
>
> I guess the question is, what is the business model going to be. If you
> ("the company")

trust me, 'I' am not 'the company'.
We all try to work on things that we feel are compelling.
Nobody wants to spend life in a hole moving dirt around. However, there
are other people in the management and marketing chains that decide
where products and programs go. We present a product map, a schedule,
and a budget; then a management team says go or stop. The efforts of
any indivdual enginneer can make a difference in the development of a
product, but has little to no impact on its distribution and marketing.

> believe you can't make money selling an operating system,
> then don't. But I guess you want to sell something! The license will have to
> reflect that decision. I understand perfectly, that the management doesn't
> like the attitude that you will just give the source code away and pray that
> the magic of the free software movement brings back glory and money.

More than doesn't like it. If you give something away without prior
approval, you have just lost your job. And when you go for another job
and you say "I lost my last job because I 'open-sourced' a product I
worked on, against the wishes of my employer" you will not get a new job.
It may be a sad thing to say that, but it is the truth. And as I said before,
at my little level in the scheme of things, paying the mortgage, clothing and
feeding my family, those ALL come first. All the fame and glory of the
world is worth less than dirt if it doesn't take care of my kids.

> You could try selling the OS. There's must be more to the world of OS' than
> Microsoft, proprietary Unix clones, free Unix clones and obscure niche OS'
> by unknown companies. Personally I believe there is a market for a
> well-branded, technologically advanced ("smaller, faster, more versatile,
> more reliable than both Windows and Linux") general purpose OS marketed by a
> well known big company.

I agree. How the marketing side of Inferno got botched, I have no idea.
I wasn't there. I waited for a long time for a reference design to be
released on widely available VME based hardware. To no avail. I do
deeply embedded things for a living. No display or keyboard. No sound.
I thought Inferno had some very interesting possibilities for distributed
applications, which my old area was heavily involved in. As it is, they
use vxWorks. I would have loved to seen a distribution model that followed
that of most RTOS vendors. I, for one, do not care to see the source code.
If it works straight out of the box on my hardware, great. I have enough
work to do without porting the OS to my hardware.

> All it needs is a full application suite to get it started and some
> marketing. OS is like a language, nobody likes it unless it has a living
> culture of users (=applications) around it.

Or unless it provides a powerful abstraction that enables the development
of complex systems that are robust, easy to architect, easy to build.
For example, the 'living culture' of ObjecTime is very small, but when
properly put to use, it is an amazing tool.

> I think the best license choice would be to distribute a shareware version
> for a registration fee. Then you could sell various "professional" editions
> with a higher price through various channels. Most of the source could be
> made open, but not necessarily all has to. One interesting source code
> licensing mode is the one used by Troll Tech which makes the Qt toolkit
> which is used in Linux KDE environment: you can use the code freely in free
> products, but it costs to use it in commercial software.

I agree with most of this statement. The core system should be a common
abstraction that can be linked as a library to the drivers that provide
the hardware layer. The source for that hardware layer should be open.
That provides a template for development that promotes the use of the
system on as wide a base of hardware as possible, at the same time, keeping
the core of the system intact such that you do not end up with a million
subtle variations on the theme.

I will happily say that my view is colored by my joyful experience of
developing embedded applications. I like my job, I just wish I had some
better tools to do it with. The old models are getting a bit tired.

Doug