Subject: Re: How accurate is Metcalfe's law? (Was: Ximian software)
From: Ian Lance Taylor <>
Date: 07 Jan 2002 09:44:47 -0800

"Stephen J. Turnbull" <> writes:

> >>>>> "Ian" == Ian Lance Taylor <> writes:
>     Ian> You misdescribed the point I was trying to make, no doubt
>     Ian> because I did not describe it clearly myself.
> I don't think so.  You observed that some software seems to be priced
> far higher than the cost of producing it, in terms of developer time.

No, that is not the point I was making.  But since I've already given
two explanations of what I meant, I don't think it's worth going over
it again.  I'm not going to get any clearer.

>     Ian> Nobody who wants to write a new word processor or a new
>     Ian> spreadsheet or a new version of Windows has to work around
>     Ian> any IP.
> True, but anybody who doesn't give a hoot about _writing_ one, but
> does want to _sell_ one, does have to work around IP.  They can't sell
> Windows or Word without dealing with Microsoft's IP.

Dealing with it in some sense, I suppose, but Microsoft, for all its
faults, has never been a patent aggressor.  StarOffice, which is a
pretty direct copy of MS Office, was being sold for a while, and I've
never heard that Microsoft did anything about it, nor have I heard
that the StarOffice writers worried about it.

>     Ian> However, I'm a programmer, and I don't believe that software
>     Ian> [creation amounts to "accidental" discovery of obscure
>     Ian> ideas].  Although there are a few exceptions, most good ideas
>     Ian> in software are obvious
>     >> Fred Brooks says that's wishful thinking.  Programmers have
>     >> very strong incentives believe that; I'm not surprised you do.
>     >> By the same token, I have to say that on this I trust the
>     >> managers more than I do the engineers.
>     Ian> Why do programmers have a strong incentive to believe that?
>     Ian> I would think that programmers would have an incentive to
>     Ian> believe the reverse: that they were very clever in their
>     Ian> ideas.  I've certainly known several programmers who thought
>     Ian> just that.
> Right.  Those guys also believe they could reproduce any given
> software product as fast as they can type, don't they?  I have heard
> many programmers trash-talk others in the light of the other
> programmers' nearly complete solutions.  I have heard many programmers
> say "oh, we don't need to worry about that issue now; if there's a bug
> we'll fix it."  I've rarely heard a programmer talk about how
> irreplaceably clever some _other_ programmer (or the architect) was.
> Just because hubris is hypocritical doesn't mean that it isn't hubris.

I suspect that we are talking about two different things.  On the one
hand, all sensible programmers (not the sort that you are talking
about) agree that creating a large project is a lot of work.  But that
work is conceptually straightforward, just time consuming.  I am
saying that for most issues no special genius is required, that
programmers don't need to come up with great ideas.  They just need to
do the work.

For example, no sensible programmer would think that Windows was easy
to implement.  But I think most good programmers would think that they
could implement Windows, with a sufficiently large team and enough
time.  There is no secret which needs to be discovered.

>     Ian> Can you point me to where Fred Brooks said that was wishful
>     Ian> thinking?
> This is the whole point of "No Silver Bullet," as I see it.
> "Accidents" in implementation and testing according to Brooks are not
> usually responsible for schedule slippages, etc; it's the "essential"
> difficulty of design (ie, coming up with good ideas, and more central,
> the good meta-ideas that link the good ideas) that is so costly.

Yes, but I don't see how that contradicts my point.  I'm not saying
that programming is trivial; I'm saying that it doesn't require being
the one person to find a diamond.

>     Ian> I was challenging your statement: [licensing] is not the only
>     Ian> way [to link developer compensation to user value].
> Oh, I see.  I hardly think it's responsive to my point though.  I
> assumed that software would be delivered to users (whether in-house or
> through the market).  You're saying "we can compensate developers by
> withholding the product from the market entirely"!

I wrote saying that software licensing had some problems.  You said it
was the only way to link developer compensation to user value.  I
challenged that in two different ways, my point being that it is not
the only way.  We should keep trying to think of new ways, and not
just accept blanket statements that it is the only way.

> Note that the "ASP solution" nearly completely cuts the link to the
> "user value" that is the raison d'etre of this list: the value to
> other developers.

The ``ASP solution'' does not preclude using free software as the
basis of the service, nor does it preclude distributing that software.
For example, is selling an RBL service based entirely
on free software.  See