Tuesday, March 09, 2004
Event More Conversations with Eric Raymond, Part 8
My apologies if this is very boring to non-IT types. I probably should put all this somewhere else.
If this conversation continues, I might move it to a link off my home page.
------------------------------------------------
Okay, last set of responses for today. If I wait much
longer, I have no ride home.
> Give it time. The serious commercialization only
> started around 1996.
Though if I'm right, that will NEVER happen. Okay,
SOME will manage to make a buck selling pure open
source software. Like I said, though (and you keep
dodging), there are plenty of examples of open source
software companies running to the hybrid model, RedHat
being the BEST example.
> I've been a Unix fan since 1975; the early days of
> quasi-openness were
> my formative years as a programmer. In 1984 I
> thought AT&T's move to
> make Unix into a "real product" was a wonderful
> idea. I spent most of
> the next ten years slowly and reluctantly learning
> that it was a
> disasterously *bad* idea -- that the pressure of
> commercialization
> destroyed most of the virtues that had made Unix
> interesting and
> useful. This experience was an important part of
> what put me on the
> road to open source.
The problems faced by Unix are NOT faced by users of,
say, the Macintosh. You want a unified software
platform that you can build products for. I would say
that that is a LOT harder when the source code is
available for all the world to see. You leave
yourself WIDE open to fragmentation. GPL helps
matters somewhat, but it won't stop RedHat (that
example again) from spiking their implementation with
proprietary add-ons that people start to depend on.
> > Either way, you have to admit that when you have
> > access to the souce code, you have a LOT more
> leverage
> > to create an incompatible variant of the product.
>
> Yes, but it's rare for anyone to actually do this.
> The bad effects
> of having multiple proprietary implementations and
> no standards are
> much worse.
Oh really? I've been a consultant for MANY years, and
it is EXCEEDINGLY common for me to find a client that
has decided to "fork" their version of an open source
product. Likewise, Linux distributions are
sufficiently different that you can't take a product
that runs on distro A and assume it will run on distro
B.
So, I think that it is VERY common that people end up
making incompatible variants of an "open source"
product. In fact, your traumatic experience of Unix
was a good example.
My apologies if this is very boring to non-IT types. I probably should put all this somewhere else.
If this conversation continues, I might move it to a link off my home page.
------------------------------------------------
Okay, last set of responses for today. If I wait much
longer, I have no ride home.
> Give it time. The serious commercialization only
> started around 1996.
Though if I'm right, that will NEVER happen. Okay,
SOME will manage to make a buck selling pure open
source software. Like I said, though (and you keep
dodging), there are plenty of examples of open source
software companies running to the hybrid model, RedHat
being the BEST example.
> I've been a Unix fan since 1975; the early days of
> quasi-openness were
> my formative years as a programmer. In 1984 I
> thought AT&T's move to
> make Unix into a "real product" was a wonderful
> idea. I spent most of
> the next ten years slowly and reluctantly learning
> that it was a
> disasterously *bad* idea -- that the pressure of
> commercialization
> destroyed most of the virtues that had made Unix
> interesting and
> useful. This experience was an important part of
> what put me on the
> road to open source.
The problems faced by Unix are NOT faced by users of,
say, the Macintosh. You want a unified software
platform that you can build products for. I would say
that that is a LOT harder when the source code is
available for all the world to see. You leave
yourself WIDE open to fragmentation. GPL helps
matters somewhat, but it won't stop RedHat (that
example again) from spiking their implementation with
proprietary add-ons that people start to depend on.
> > Either way, you have to admit that when you have
> > access to the souce code, you have a LOT more
> leverage
> > to create an incompatible variant of the product.
>
> Yes, but it's rare for anyone to actually do this.
> The bad effects
> of having multiple proprietary implementations and
> no standards are
> much worse.
Oh really? I've been a consultant for MANY years, and
it is EXCEEDINGLY common for me to find a client that
has decided to "fork" their version of an open source
product. Likewise, Linux distributions are
sufficiently different that you can't take a product
that runs on distro A and assume it will run on distro
B.
So, I think that it is VERY common that people end up
making incompatible variants of an "open source"
product. In fact, your traumatic experience of Unix
was a good example.
