Friday, March 05, 2004
Conversation with Eric Raymond, Part 1
What follows is my side of an email exchange I recently had with Eric Raymond. When I write fast, I often leave words out, and that is certainly the case with this email exchange. Therefore, where necessary, I've added the missing words, marking them with [square brackets].
I will post the full text of Eric Raymond's side of the discussion, provided he gives me permission to do so. A lot of it is interspersed in my reply, though, so you can get the gist of his argument from that.
As a side point, regarding my 50-50 comment (you'll know what I'm talking about later), I mean that 50% of software will be open source, and 50% will be proprietary WITHIN the same product. That's just an average. I do believe, though, that it will become more common for major software companies to release MORE of the source for their flagship products, though the amount that is public will average about 50%.
Proprietary software companies, however, will still run the show, and will be the source of most of the innovation in the computer industry.
-----------------------------------------------------------------
No, this isn't a group address. This is just my
catch-all for any email related to the articles I
write for ZDNet. Since it's public, it also attracts
lots of spam.
By the way, for a bit of context to what I write,
visit my web site (http://www.turtlenecksoftware.com).
I've written over 40 articles for ZDNet, which
doesn't put me anywhere near your league, but at least
it's a track record of my logic (or lack thereof) over
the years.
First off, I DON'T lump all open source programmers
into the "free (as in cost) software" movement. I DO,
lump Mr. Stallman into that camp, and would point you
to Stallman's recent article on ZDNet commemorating 20
years of free software
(http://zdnet.com.com/2100-1107_2-5135756.html). He
essentially attacks ANYBODY who chooses not to make
free software (as in fully open source), considering
their contributions unwelcome. In my response
(http://zdnet.com.com/2100-1107_2-5137217.html), I
also point out that open source software, particularly
under the GPL, inevitably leads to free as in cost
simply because the licensing model ensures that
everyone has the right to redistribute the code to
anybody and everybody. That makes the claim that
"open source does not imply gratis" [little more] than a thin veneer
designed to make the result more palatable to people
who make a living from software.
> It is therefore a bit annoying to see you writing as though we're all
> a bunch of fist-pumping Communards out to monkeywrench post-industrial
> capitalism.
I'm not a fist-pumping communard, and I write articles
for free AND release open source products. In other
words, I obviously don't view all open source
programmers in that light. I DO think, however, that
the open source community has given insufficient
attention to business requirements. Likewise, I think
there is an almost instinctive rejection of
proprietary software simply because it is proprietary,
while ignoring the REASON it is proprietary, which is
to generate revenue.
What I've always advocated since I first broached the
open source [topic] back in 2002 is a hybrid model, one where
the code for software is mostly available, but certain
key portions are kept private in order to ensure a
decent revenue stream. In fact, my first article on
ZDNet advocated that Microsoft open source the Windows
core, keeping the stuff that really differentiates
them proprietary in order to secure the required
revenue stream.
> No, what exercises most of us is not the belief that
> closed-source development is morally wrong but rather
> that it is, most of the time, *stupid* -- a bad deal
> for consumers, a huge time- and effort-waster
> for programmers, and a dead weight around the necks
> of IT managers everywhere.
On the other hand, a market drained of avenues for
profit is one that attracts less investment, and
provides fewer incentives for programmers to orient
themselves towards the needs of non-programmers. Why
has Linux had a difficult time making a
consumer-oriented operating system? I would suggest
that it has more to do with the fact that it relies on
volunteers. Volunteers tend to do the sorts of things
they are interested in. Proprietary software,
however, provides the incentives (a salary) for people
to concentrate on OTHER people's interests.
That isn't to say that Linux hasn't made great strides
in helping regular consumers - it has. But I would
suggest that has more to do with Linux being picked up
by companies like IBM, RedHat and even Lindows. They
provide the incentives needed to steer Linux down the
consumer-orientation path.
> everywhere. My personal prediction is that, at an
> equilbrium we won't see for some years yet, the
> software market will be about 95% open
> source and 5% proprietary.
My personal prediction would be closer to the 50-50
mark (give or take 10 either way). I still think
proprietary software companies are the bodies who best
tease out what regular consumers want, simply because
they are closest to their client base and have a
vested interest in serving them. Short of software
technology reaching a plateau, we will STILL need
those consumer-facing proprietary entities to keep
technology rolling forward. The experimentation which
goes on due to hundreds of thousands of
profit-oriented companies trying to figure out how to
make new technology fit consumer needs is powerful
stuff (that's probably going to be the subject of a
future article, after I've finished my move to
Dublin).
Of course, feel free to disagree (not that I have to
tell you that).
> ignores the very real problems of being locked in to a
> single vendor and at the wrong end of a monopoly relationship.
That is a valid point, and is part of the pragmatic
argument for open source. However, that doesn't hold
much water for regular consumers. They aren't
programmers, and will be locked in to whatever
platform proves most popular with the wider
development community. They're choices flow from the
whims of programmers.
That's the problem with software. Software needs
standards because it lacks any inherent
standardization. ANY standard you pick locks you in
to that platform. Lots of companies are already
locked in to a particular distribution of Linux.
David Berlind talks about that in some of his ZDNet
articles, specifically his most recent one on ZDNet
which I would highly recommend reading.
I'm not so sure that open source TRULY frees you from
lock-in. You just get another form of lock-in.
Granted, you have the source code and theoretically
could make your new environment look like it.
However, SAMBA and WINE haven't been stopped by lack
of access to source code so much as the fast pace at
which Microsoft evolves their platform.
Anyway, thanks for your response (and sorry if mine is
so long. Another problem with my articles is that
they are usually WAY too long...just like my emails).
John Carroll
What follows is my side of an email exchange I recently had with Eric Raymond. When I write fast, I often leave words out, and that is certainly the case with this email exchange. Therefore, where necessary, I've added the missing words, marking them with [square brackets].
I will post the full text of Eric Raymond's side of the discussion, provided he gives me permission to do so. A lot of it is interspersed in my reply, though, so you can get the gist of his argument from that.
As a side point, regarding my 50-50 comment (you'll know what I'm talking about later), I mean that 50% of software will be open source, and 50% will be proprietary WITHIN the same product. That's just an average. I do believe, though, that it will become more common for major software companies to release MORE of the source for their flagship products, though the amount that is public will average about 50%.
Proprietary software companies, however, will still run the show, and will be the source of most of the innovation in the computer industry.
-----------------------------------------------------------------
No, this isn't a group address. This is just my
catch-all for any email related to the articles I
write for ZDNet. Since it's public, it also attracts
lots of spam.
By the way, for a bit of context to what I write,
visit my web site (http://www.turtlenecksoftware.com).
I've written over 40 articles for ZDNet, which
doesn't put me anywhere near your league, but at least
it's a track record of my logic (or lack thereof) over
the years.
First off, I DON'T lump all open source programmers
into the "free (as in cost) software" movement. I DO,
lump Mr. Stallman into that camp, and would point you
to Stallman's recent article on ZDNet commemorating 20
years of free software
(http://zdnet.com.com/2100-1107_2-5135756.html). He
essentially attacks ANYBODY who chooses not to make
free software (as in fully open source), considering
their contributions unwelcome. In my response
(http://zdnet.com.com/2100-1107_2-5137217.html), I
also point out that open source software, particularly
under the GPL, inevitably leads to free as in cost
simply because the licensing model ensures that
everyone has the right to redistribute the code to
anybody and everybody. That makes the claim that
"open source does not imply gratis" [little more] than a thin veneer
designed to make the result more palatable to people
who make a living from software.
> It is therefore a bit annoying to see you writing as though we're all
> a bunch of fist-pumping Communards out to monkeywrench post-industrial
> capitalism.
I'm not a fist-pumping communard, and I write articles
for free AND release open source products. In other
words, I obviously don't view all open source
programmers in that light. I DO think, however, that
the open source community has given insufficient
attention to business requirements. Likewise, I think
there is an almost instinctive rejection of
proprietary software simply because it is proprietary,
while ignoring the REASON it is proprietary, which is
to generate revenue.
What I've always advocated since I first broached the
open source [topic] back in 2002 is a hybrid model, one where
the code for software is mostly available, but certain
key portions are kept private in order to ensure a
decent revenue stream. In fact, my first article on
ZDNet advocated that Microsoft open source the Windows
core, keeping the stuff that really differentiates
them proprietary in order to secure the required
revenue stream.
> No, what exercises most of us is not the belief that
> closed-source development is morally wrong but rather
> that it is, most of the time, *stupid* -- a bad deal
> for consumers, a huge time- and effort-waster
> for programmers, and a dead weight around the necks
> of IT managers everywhere.
On the other hand, a market drained of avenues for
profit is one that attracts less investment, and
provides fewer incentives for programmers to orient
themselves towards the needs of non-programmers. Why
has Linux had a difficult time making a
consumer-oriented operating system? I would suggest
that it has more to do with the fact that it relies on
volunteers. Volunteers tend to do the sorts of things
they are interested in. Proprietary software,
however, provides the incentives (a salary) for people
to concentrate on OTHER people's interests.
That isn't to say that Linux hasn't made great strides
in helping regular consumers - it has. But I would
suggest that has more to do with Linux being picked up
by companies like IBM, RedHat and even Lindows. They
provide the incentives needed to steer Linux down the
consumer-orientation path.
> everywhere. My personal prediction is that, at an
> equilbrium we won't see for some years yet, the
> software market will be about 95% open
> source and 5% proprietary.
My personal prediction would be closer to the 50-50
mark (give or take 10 either way). I still think
proprietary software companies are the bodies who best
tease out what regular consumers want, simply because
they are closest to their client base and have a
vested interest in serving them. Short of software
technology reaching a plateau, we will STILL need
those consumer-facing proprietary entities to keep
technology rolling forward. The experimentation which
goes on due to hundreds of thousands of
profit-oriented companies trying to figure out how to
make new technology fit consumer needs is powerful
stuff (that's probably going to be the subject of a
future article, after I've finished my move to
Dublin).
Of course, feel free to disagree (not that I have to
tell you that).
> ignores the very real problems of being locked in to a
> single vendor and at the wrong end of a monopoly relationship.
That is a valid point, and is part of the pragmatic
argument for open source. However, that doesn't hold
much water for regular consumers. They aren't
programmers, and will be locked in to whatever
platform proves most popular with the wider
development community. They're choices flow from the
whims of programmers.
That's the problem with software. Software needs
standards because it lacks any inherent
standardization. ANY standard you pick locks you in
to that platform. Lots of companies are already
locked in to a particular distribution of Linux.
David Berlind talks about that in some of his ZDNet
articles, specifically his most recent one on ZDNet
which I would highly recommend reading.
I'm not so sure that open source TRULY frees you from
lock-in. You just get another form of lock-in.
Granted, you have the source code and theoretically
could make your new environment look like it.
However, SAMBA and WINE haven't been stopped by lack
of access to source code so much as the fast pace at
which Microsoft evolves their platform.
Anyway, thanks for your response (and sorry if mine is
so long. Another problem with my articles is that
they are usually WAY too long...just like my emails).
John Carroll
