Friday, March 05, 2004
Conversation with Eric Raymond, Part 3
> No, I read your article very carefully. You
> attributed a normative belief that all software
> should be free to "open source" developers.
> If you had attributed that belief to "free software"
> developers, you would have been correct. As it is,
> you're wrong -- as RMS is likely
> to point out in terms less polite than mine.
My argument has long been that open source will lead
to free, whether or not that's what the developers
themselves think they are doing. That is the case
with GPL software (as noted in my last two responses),
and is certainly the
case with a BSD license. Granted, there are licenses
that involve releasing the source code while charging
for it, but look at how much noise was made over the
fact that Qt wasn't a "proper" GPL license. They now
release a GPL version, but that didn't used to be the
case. And either way, once you release the source
code, you have to run a lot faster to keep [ahead of] the
price erosion [which] results from giving away to all and
sundry the secret sauce that went into your product.
GPL LEADS to free, whether or not open source
developers, or Stallman, want to admit it in public.
The claim Stallman made that "Free software does not
mean "gratis" " was the basis of my "Response"
article. Stallman likes to play around with language,
but in this case, I think it's a way to spin the
apparently-misnamed "Free Software" into a liberty
sense and NOT a cost sense. I say that's
nonsense...that you can't escape the "free as in cost"
aspect of GPLed software.
> Richard Stallman will tell you himself that he is
> not an "open source" developer. In fact he'll
> insist on it. In his mind, what distinguishes
> "free software" developers from "open source"
> developers is the position you attribute to "open
> source" developers.
Perhaps this is the area that needs more clarity,
because he says things like "Free software does not
imply gratis." He probably thinks that somebody might
manage to make a buck from the sale of open source
software (and, again, SOME do, but you have to admit
that they are dwarfed, revenue-wise[, by proprietary software
companies]
> You appear to me to be more than a bit confused
> about the facts on the ground. Which is no crime or
> sin, but you should refrain from pomtificating about
> what "open source" developers believe until you
> have basics like the difference between the
> Stallmanite/FSF position and that of most open-source
> developers firmly in your grasp.
One man's pontification is another man's opposing
viewpoint.
I'll grant that open source programmers are certainly
operating under the assumption that open source can be
a stable business model for software development, but
I think the facts, such as RedHat's dance away from
Fedora towards Advanced Server as the foundation of
their business (which has LOTS of RedHat specific
stuff in it, including proprietary modules), is a sign
that many are starting to come to that realization.
> Those exceptions are sufficient to falsify the
> universal claim you made.
This isn't a courtroom, but a debate. I didn't say
that NOBODY makes money from open source software as
such. I think, though, that open source has been
around for long enough that there SHOULD be a lot more
pure open source software companies that make revenue
from software sales. There isn't. That should be a
sure sign that somebody is starting to realize that
the reason proprietary software appeared in the first
place was as a way to make revenue.
> It's a difficult business to be in. But Red Hat
> obviously manages it.
> Zope, Inc. is another example. You named a third
> yourself.
Yes, you've found a few examples (one of which I gave
you). On the other hand, I can name stacks of
proprietary software companies that makes LOTS of
money from software. I strongly question whether ANY
company will grow really big by basing their business
model on PURE open source. Like I said, I advocate
the hybrid model. I don't expect you to agree with me
on this.
> I'm just looking at the evidence. You seem to have
> the peculiar idea
> that revenue from software isn't "real" unless it's
> directly from sales --
> that consulting services and associated bundled
> goods don't count.
> Why define them out of the discussion?
Making software AS SUCH simply is not the same thing
as making software as a consulant. In the former
case, the market is composed almost entirely of
proprietary software companies, because that's how
they make money: by selling access to a secret. IN
the latter case, they make CUSTOM software for
companies, then pass it off to be owned by the client.
In the one case, someone is responsible for
maintaining JUST their own products, and dealing
EXCLUSIVELY with clients of those products. In the
latter case, they are responsible for whatever
software is OWNED by the client, either stuff they
make from scratch or stuff the client already owned
from past consulting relationships.
The reason I make the distinction is that the SOFTWARE
COMPANY model simply does not exist (or rather, to be
more precise, exists in a FAR more limited form) in a
world composed of 100% (or even the 95% you propose)
open source software. The open nature of open source
militates against it.
Granted, that open nature is wonderful stuff, and as I
discussed in the article, I benefit greatly from it.
That does not mean there isn't value in a business
model that provides the revenue to create a bunch of
client-oriented, self-funding software (as such)
companies.
As noted, self-funding proprietary companies conduct
the market research required to figure out how to make
new technology fit consumer needs. You even admitted
in your second response that proprietary sofwtare
companies are more "client-oriented" than open source
products (though you questioned whether their output
was "better"). The software markets works by
unleashing thousands of competing crucibles, each of
which attempt to best figure out what customers want.
The best end up becoming the biggest.
> The facts do not fit your construct.
I guess we'll have to see on that one.
> No, I read your article very carefully. You
> attributed a normative belief that all software
> should be free to "open source" developers.
> If you had attributed that belief to "free software"
> developers, you would have been correct. As it is,
> you're wrong -- as RMS is likely
> to point out in terms less polite than mine.
My argument has long been that open source will lead
to free, whether or not that's what the developers
themselves think they are doing. That is the case
with GPL software (as noted in my last two responses),
and is certainly the
case with a BSD license. Granted, there are licenses
that involve releasing the source code while charging
for it, but look at how much noise was made over the
fact that Qt wasn't a "proper" GPL license. They now
release a GPL version, but that didn't used to be the
case. And either way, once you release the source
code, you have to run a lot faster to keep [ahead of] the
price erosion [which] results from giving away to all and
sundry the secret sauce that went into your product.
GPL LEADS to free, whether or not open source
developers, or Stallman, want to admit it in public.
The claim Stallman made that "Free software does not
mean "gratis" " was the basis of my "Response"
article. Stallman likes to play around with language,
but in this case, I think it's a way to spin the
apparently-misnamed "Free Software" into a liberty
sense and NOT a cost sense. I say that's
nonsense...that you can't escape the "free as in cost"
aspect of GPLed software.
> Richard Stallman will tell you himself that he is
> not an "open source" developer. In fact he'll
> insist on it. In his mind, what distinguishes
> "free software" developers from "open source"
> developers is the position you attribute to "open
> source" developers.
Perhaps this is the area that needs more clarity,
because he says things like "Free software does not
imply gratis." He probably thinks that somebody might
manage to make a buck from the sale of open source
software (and, again, SOME do, but you have to admit
that they are dwarfed, revenue-wise[, by proprietary software
companies]
> You appear to me to be more than a bit confused
> about the facts on the ground. Which is no crime or
> sin, but you should refrain from pomtificating about
> what "open source" developers believe until you
> have basics like the difference between the
> Stallmanite/FSF position and that of most open-source
> developers firmly in your grasp.
One man's pontification is another man's opposing
viewpoint.
I'll grant that open source programmers are certainly
operating under the assumption that open source can be
a stable business model for software development, but
I think the facts, such as RedHat's dance away from
Fedora towards Advanced Server as the foundation of
their business (which has LOTS of RedHat specific
stuff in it, including proprietary modules), is a sign
that many are starting to come to that realization.
> Those exceptions are sufficient to falsify the
> universal claim you made.
This isn't a courtroom, but a debate. I didn't say
that NOBODY makes money from open source software as
such. I think, though, that open source has been
around for long enough that there SHOULD be a lot more
pure open source software companies that make revenue
from software sales. There isn't. That should be a
sure sign that somebody is starting to realize that
the reason proprietary software appeared in the first
place was as a way to make revenue.
> It's a difficult business to be in. But Red Hat
> obviously manages it.
> Zope, Inc. is another example. You named a third
> yourself.
Yes, you've found a few examples (one of which I gave
you). On the other hand, I can name stacks of
proprietary software companies that makes LOTS of
money from software. I strongly question whether ANY
company will grow really big by basing their business
model on PURE open source. Like I said, I advocate
the hybrid model. I don't expect you to agree with me
on this.
> I'm just looking at the evidence. You seem to have
> the peculiar idea
> that revenue from software isn't "real" unless it's
> directly from sales --
> that consulting services and associated bundled
> goods don't count.
> Why define them out of the discussion?
Making software AS SUCH simply is not the same thing
as making software as a consulant. In the former
case, the market is composed almost entirely of
proprietary software companies, because that's how
they make money: by selling access to a secret. IN
the latter case, they make CUSTOM software for
companies, then pass it off to be owned by the client.
In the one case, someone is responsible for
maintaining JUST their own products, and dealing
EXCLUSIVELY with clients of those products. In the
latter case, they are responsible for whatever
software is OWNED by the client, either stuff they
make from scratch or stuff the client already owned
from past consulting relationships.
The reason I make the distinction is that the SOFTWARE
COMPANY model simply does not exist (or rather, to be
more precise, exists in a FAR more limited form) in a
world composed of 100% (or even the 95% you propose)
open source software. The open nature of open source
militates against it.
Granted, that open nature is wonderful stuff, and as I
discussed in the article, I benefit greatly from it.
That does not mean there isn't value in a business
model that provides the revenue to create a bunch of
client-oriented, self-funding software (as such)
companies.
As noted, self-funding proprietary companies conduct
the market research required to figure out how to make
new technology fit consumer needs. You even admitted
in your second response that proprietary sofwtare
companies are more "client-oriented" than open source
products (though you questioned whether their output
was "better"). The software markets works by
unleashing thousands of competing crucibles, each of
which attempt to best figure out what customers want.
The best end up becoming the biggest.
> The facts do not fit your construct.
I guess we'll have to see on that one.
