Subject: Re: Opportunity lost? Challenge declined!? (LONG. COMPREHENSIVE)
From: Ian Lance Taylor <>
Date: 10 May 2001 10:18:39 -0700

Adam Theo <> writes:

> * First, on the matter of the software (here-on called 'work') I am
> wanting to publish (An alarm clock program written in Perl/Tk and
> using XML. I know, what a little program to bring me to this list
> asking about licensing.): On this matter I have wanted to release it
> with legal "back-up" (At the time thinking using software licenses.)
> that would allow me, as Mr. Blankley put it, to be compensated for my
> investment in development while still allowing the*source* of my work
> to be widely viewable by users and third-party developers alike. I
> want to allow users (or third-party reviewers) the open opportunity to
> completely evaluate the work before buying and using it, so they can
> rest assured it does not contain bugs, trapdoors, or will send
> personal information back to my servers. I also wish to allow
> third-party developers the freedom to "work-off" of my work, to
> develop patches of bugs or security holes, add-ons or plug-ins of new
> or improved features, or allow their own works to use features in my
> work, all without having to jump through legal rings of fire or pay
> royalty fees. I would not allow these users or developers freedom to
> re-distribute my work without my permission, so therefore any works of
> theirs that makes use of mine would have to be independant and reliant
> on whether the end user has my work installed as well. This may seem
> foolish, obvious, or impossible in many ways, so I apologize. I am
> just stating all of this for completeness.

Your goal is to be compensated for your investment in development.  I
suspect that a program like this is not going to cost very much--
people aren't going to make a significant investment in an alarm clock
program.  So to regain your investment, you need to sell a fairly
large number of copies at a small price.

How are people going to find out about the program?  At a small sale
price, the cost to acquire a customer has to be very very low.  You
can't afford to run an advertising campaign.  You need information
about the product to spread by word of mouth.  This suggests a
shareware model, but my understanding is that shareware typically does
not make money.

Perhaps you can make a subset of the program which is small but
useful, and release that as free software.  Then the small version of
the program can automatically encourage people to purchase the
full-featured version.  If you do this, I believe that the
full-featured version can be free software too; I predict that if
people pay to get the full-featured version, they will choose to not
distribute it outside their organization even if they could.  That is
what we found to be true at Cygnus.  Even if they do distribute the
full featured version, they're sure as heck not going to run their own
advertising campaign, so most people will still come to you.

Maybe it doesn't make sense to make a subset of the program.  But
frankly, if you are the only person who can distribute the program,
proprietary or not, I don't see how you will make money off of it.
How are people going to find out about it?

I'm focusing on the details here because in business, as in
programming, the details matter.  There is no one-size fits all
solution to repaying an investment.  If there were, we would all be
rich, or none of us would be.

> * Four, on the matter of Intellectual Property (specifically the
> rights of the work's author): If an author creates code, and knowing
> the full issues of the action, decides to release it as open source
> (or free software, acknowledging the differences between the two.),
> not only do I see that as okay, but even greater as a noble and
> selfless act that benefits his peers and all of mankind alike. My
> concern is for the authors who desire assured compensation (Not
> speaking in context of supply-demand where people could find his work
> un-desirable and not wish to buy it. This is in context of assured
> compensation for every copy of his work others do use.). For those
> authors who for one reason or another want to keep their work
> proprietary, I feel they should have that right (and of course face
> any consequences of not having a desirable product). I feel the free
> market, with it's consumer choice and supply-demand forces at work
> should decide what business models float or sink, not the will of a
> few making others feel their only choice is to follow. I know that
> sounds harsh, but it is only to explain that I feel the consumers, end
> users, should be able to choose how they aquire software, even if it
> is from authors holding on to proprietary source. I do not feel
> authors should be given the choice of only OSS when it comes to
> releasing their works. I fear the consequences, which is in my Fifth,
> and final ("hoo-ray", right?) point.

I think most people agree that everybody should have the right to keep
their work proprietary.  It's possible that RMS in his wilder moments
has argued against this.  But normally he just says that though people
have the right to release proprietary code, he has the right to not
purchase it.

So, I think that either I have misunderstood your point, or that you
are arguing against a straw man position which nobody actually holds.

> * Five, on the matter of consequences of an all-OSS/FS world: I do
> believe many software developers would be seriously hurt if they could
> not make a living from their livlihood. I do not agree with the
> argument that this would just result in the "weaker" (using a
> Darwinian outlook) programmers being forced out of the field. I
> believe it would hurt the "strong" and the "weak" alike. I then do not
> agree with the follow-up argument that this is simply my own beloved
> "free-market forces" at work, that it is natural for formerly
> money-rich industries to sink, and other, new ones, rise. It is not
> the free market forces at work because it is not the result of the
> consumer's decision. It is the result of authors feeling they are
> between a rock and a hard place when it comes to profiting off their
> work.

Again, I don't think anybody is seriously arguing for an all free
software world that is not the result of the consumer's decision.

(I will parenthetically note that the proprietary software world is
not the result of the consumer's decision either; it is the result of
the government's grant of monopoly rights to the creators of
intellectual property (perhaps that should be intellectual
``property'').  Take the courts out of the picture, and there is no
proprietary software.  I'm not saying that the current state of
affairs is wrong.  But I don't think it is somehow the natural one.  I
don't know what the natural state of affairs is.  I don't even know if
that is a coherent question.)

But, that said, I believe the proper way to consider an all free
software world is to ask about the benefits to society as a whole.  It
is not to ask about the benefits to software developers.  It is
possible that society as a whole will do better if there were fewer
software developers, just as many people suspect that society as a
whole would do better if there were fewer lawyers.

> Yes, I agree. Although I didn't hear about the whole "Mundie Affair"
> until a week after it had already hit the O'Reilly Network, I've since
> read up on it, and had to pick my jaw up off the floor. Could any of
> us have forseen the Microsoft of today (or rather the very near future
> Microsoft) 5 years ago? Heck, even 2 years ago would have been a
> stretch. The Microsoft that at the time was world-renowned for heavily
> integrating their internet browser into their OS which effectively
> destroyed Netscape's last hold on the desktop, for taking the
> protocols and systems of other developers and claiming them as their
> own after giving them their own M$ twists and dependencies. And now my
> shock when I heard Microsoft would not only use XML in it's future
> products without trying to alter the standards in their favor, but
> would be embracing it to such a degree as they seem to be doing in
> their Office suite and .Net initiative. True, they are still a long
> way from winning "Fair Player of the Year", but with this recent
> announcement by Mundie, they are showing they can change, and with
> them, the entire industry.

I don't want to get into this detail, but I don't yet see any evidence
that Microsoft can change.  Microsoft is saying that they won't alter
the standards in their favour, but they have said that before and
lied.  I do agree that rabid anti-Microsoft tirades serve no useful
purpose.  But I believe that as a company they have shown themselves
to be untrustworthy, and since the same people are still in charge
it's going to take a good while before I'm willing to trust them.

> I understand your use of Red Hat (who can mention open source in a
> business model nowadays and not bring up Red Hat?), but as I briefly
> stated above, I am confident multi-million dollar corporations can
> find a way to profit. They have the resources to be able to widely
> distribute "packaged" software at a reasonable price and make money
> soley on the price of packaging. They also have the resources at their
> disposal to staff the number of knowlegable people needed to provide
> support for these works.

I really don't understand this point of view at all.  You say that Red
Hat is a multi-million dollar corporation, and they are.  But do you
think they were created as a multi-million dollar corporation?  The
company was founded in 1994 out of the back of Bob Young's van.  They
didn't have any of the resources you are talking about to start with.
He and others built it into a multi-million dollar corporation, one
step at a time, in just five years.  They were smart, they worked
hard, and they were lucky, but they weren't fundamentally different
from you or me.

You're looking at the Red Hat of today and saying it is easy for them
to make money.  But why not look at the Red Hat of seven years ago?
It's the same company.

> 4.  The software Red Hat sells support to is very complex. Trying to
> sell support to more simplistic works such as what I currently have
> would be near impossible. When the open source work is simple, a
> support-driven revenue just wouldn't hold up.

In my last message I explained that though Red Hat makes a significant
amount of money from support, that is not where they make most of
their money.

> 5. (related to 2) Many people point to Linus Torvalds as an example of
> how people can become wealthy off of developing open source. I suggest
> he is the exception, not the rule. He is wealthy mostly from stock
> options freely given to him by start-ups that make heavy use of, and
> greatly admire, his work. Can we think that these same start-ups would
> give to all (even most) developers of the software they use or sell?
> When donations become the primary source of personal wealth, most will
> fall through the cracks, while the few that happen to become famous
> sit on a nice sized hill of stock and cash.

No, if Linus is rich, he is rich because he is employed by Transmeta,
a company which is so far reasonably successful.  I'm sure he has made
money from donations.  But I'm also sure that the money he has made
from Transmeta is far more than the money he has made from donations.

> 6.  Red Hat, and comparable companies, are quite large compared to the
> small garage business that is hoping to make some money from a nifty
> new technology sprung up by one of it's genius owners. Being larger
> has it's benefits here. Namely, access to critical resources that
> simply are not available to this garage co. or other small
> business. Resources like cash and credit lines, but also more
> importantly manpower and the all-powerful "contacts".

Again, not very long ago, Red Hat was a small garage business too.  Do
you think that companies leap fully formed from the brow of Zeus?

> "Glen Starchman" <> wrote:
>     That is also a case in point. Sure, there are people who write
>     software for the love of it. But there are also people who write
>     software to make a living. If I run a bakery, I am not going to
>     give out my bread recipe.
> "Ian Lance Taylor" <> wrote:
>     Not a great example, since 1) bread is not permanent, so people
>     will always come back to you for more; 2) the marginal cost of
>     making and selling another loaf of bread is significant relative
>     to the purchase price, unlike the marginal cost of making and
>     selling another copy of a typical software program.
> True, not a great example, but it suffices. Look at it without taking
> it literally. Mr. Starchman is basically saying no-one really wants to
> give away the very thing they make their livlihood on. "Want" may not
> even be the best word for it. "Should" might be more fitting.

OK, granted, but still: software is not bread.  If you reason by
analogy, you must choose appropriate analogies.  You see,
Mr. Starchman's analogy leads you to think that you should not give
away your source code.  Your version leads in a completely different
direction: to figure out just what it is to you make your livelihood
on.  You may come to the same place in the end, but you will use a
completely different reasoning process.

> "Dave Blankley" <> wrote:
>     Which brings me to a more focused question than my earlier
>     challenges: How does a developer that wants to invest his time
>     developing, get compensated in the open source arena?
> "Ian Lance Taylor" <> wrote:
>     Much the way it works in any other area: start by teaming up with
>     somebody with marketing and business knowledge.
> Yes, true, however I do believe there is that brick wall, so to say,
> between the small company in open source and the large one. That brick
> wall makes all the difference, since it represents the manpower,
> credit lines, and contacts that would be essential to muscling a
> profit out of the open source field. Besides, partnering up with
> business and marketing savvy people still doesn't solve *how* they are
> going to turn a profit. *That's* why this thread was started.

That brick wall is one that almost every large company crossed (some
companies were created large by spinoffs, but most simply grew).

I mention partnering with somebody with marketing and business
knowledge because it is that person's job to figure out how to make
money.  You can learn the business knowledge yourself.  But if nobody
in the organization has marketing and business knowledge, you are
almost certainly not going to figure out how to turn a profit.  Don't
put the cart before the horse.  First you need the business analysis,
and before that you need to learn how to do it.

> "Ian Lance Taylor" <> wrote:
>     Alternatively, developers who do not do this can make money by
>     consulting (this require a modicum of business knowledge).  It is
>     possible to consult on free software projects; our host, Russell
>     Nelson, serves as an example, and there are many others.
> yes, developers can make money at consulting, and that is so far the
> most assured source of revenue for a programmer, but it is not enough
> when more and more programmers migrate over from the closed source to
> the open source fields. As the consultant population balloons, the
> only way for any large amount of them to make a living at it would be
> for consulting fees to increase greatly. Unfortunately, as basic
> economics shows, the opposite happens. A bigger pool of workers, no
> matter how experienced you are compared to the others, decreses what
> you can get.

Isn't the same true in the proprietary software world?  Why hasn't it
happened there?  (Of course I think it will happen eventually, but I
also think you are begging the question.)