Re: Rohit -- don't hack too hard :-)

Date view Thread view Subject view Author view

From: Kragen Sitaker (
Date: Sat Mar 25 2000 - 06:43:04 PST

Adam Beberg writes:
> On Fri, 24 Mar 2000, Kragen Sitaker wrote:
> > "0" is not exactly right; they actually have substantial sales in
> > China, many of which are site licenses. And even in the
> > least-copyright-respecting parts of Asia, the ratio doesn't go much
> > above 30 unlicensed copies per licensed copy. A thirtieth of a billion
> > is still a substantial number.
> I doubt there will ever be anywhere near 1/30th of a billion Windows
> PC's in China.

My CIA World Factbook 1998 says China's GDP quadrupled from 1978 to
1997, a 7.6% annual growth rate, with the fastest increase in
1992-1997. Per capita GDP for 1998 is estimated at $3,460; the
corresponding figure for the US is $30,200.

If the majority of people in the US have cars (costing $9,000 and up),
and a large number of families have two or more cars, it is quite
plausible that computers (costing $300 and up) will be widely
affordable in China.

According to the same source, China has 89 million telephones for its
1.2 billion people, about one telephone for every 13.4 people; the US
has 182 million phones for its 270 million, about one telephone for
every 1.5 people.

According to these figures, there might be 33 million Windows PCs there
already --- let alone "ever".

I don't think there are yet.

Michael Dell's 1998 speech, in which he looked forward to a future of
one computer per 500 Chinese people, suggests we're a long way from
that now. He says he thinks the market will grow at 30% per year for
the next several years, and expects China will become a bigger computer
market than Japan within a few years.

There's a USIA press release that asserts that, in 1998, there were 1.2
computers per 1,000 population in China
( If this is
true, and if Michael Dell's statement is true, we can expect there will
be 33 computers per thousand population somewhere around 2010. (I'm
fudging by assuming all of those computers were bought *in* 1998, a
very optimistic assumption.)

I'm curious to hear your reasons for doubting this will ever happen.
(Do you just think Windows will be passe by 2010?)

But, back to the issue at hand. If the USIA press release is about
right, that means there were 1,236,000 * 1.2 = 1 483 200 computers in
China in 1998. Leaving aside how many more there might be today,
that's still a $45 million market for Windows and a $45 million market
for Office. Assuming the worst-case 30:1 ratio (which is more
appropriate to Thailand than to China), that's a $3 million market.

> A 30:1 piracy rate isn't anywhere near acceptable enough
> to justify costs. I can't imagine what the MSFT stockholders would have
> been happy about if instead of the million win2k sales mark they settled
> for 30,000.

$3 million can justify a lot of costs. $3 million with a 30%-per-year
growth rate can justify a lot more. $3 million with a 30%-per-year
growth rate and lucrative consulting opportunities with what is, in
effect, the biggest company in the world can justify still more.

> > There are a few US software companies founded by immigrants --- Borland
> > comes to mind, and perhaps Be --- but the vast majority of US software
> > companies are founded by native-born US citizens. The "everyone comes
> > here to pay the bills" argument simply doesn't hold water.
> Yea, so there is no reason at all for the software industry to be
> completely paniced about the visa situation, there are no foreign
> WORKERS in the industry. And no foreign offices of american companies
> either. No worries.

Foreign sales offices of American companies further weakens the
"everyone comes here to pay the bills" argument.

I think there's a very good reason for the software industry to be
panicked. Because of the Internet --- a big glob of free software ---
they are suddenly in a position to deliver tremendous value to
consumers, value they never could have delivered from their own
proprietary R&D. But they need workers to take advantage of this
opportunity. This creates a seller's market for labor.

Labor buyers --- employers --- would obviously like to increase the
labor supply, and reduce the market choices of labor sellers, so they
don't have to pay so much for labor. H-1B visas are a perfect
mechanism for defeating the market and cheating employees.

But this is sort of a tangent. Your point, I think, is that there are
a lot of prosperous software companies in the US. That's true. Your
implication is that this is because the US has restrictive
intellectual-property laws and practices. I suggest that there are a
lot of countries with IP laws that are more restrictive with fewer
people working at prosperous software companies --- such as many
countries in Europe --- and some countries with IP laws that are less
restrictive, with equal concentrations of software-selling prosperity.
I could be wrong. Do you know where the numbers are?

> > > No IP rights, no point in writing software/music/movies, no software
> > > industry, no anything digital industry.
> >
> > I think this might be false. The vast majority of hours of work spent
> > by programmers are on bespoke projects, where the software user pays on
> > a time and materials basis to develop the software. (I think the
> > fraction is something like 70%.)
> So lets ignore that hidden 70% of internal top secret software used only
> by other geeks. What about the other 30%? Who will write apps for mom
> knowing there is no money in it? I can't think of one opensource(tm) app
> that's mom useable; the only ones that are even close are exact clones
> of commercial apps.

Mozilla. Although there are proprietary clones of it, including a
well-known one from Microsoft.

I note that you have just changed the argument; before, you asked,
"Would geeks go homeless if there were no intellectual property?" I
think the 70% argument pretty much lays that fear to rest. Now you've
changed the question to this: "Would there be software usable by
non-geeks if there were no intellectual property?"

The exact-clone argument is sort of a red herring, anyway; most
proprietary apps are exact clones of other proprietary apps. AbiWord,
the GIMP, kfm, gnumeric, and Mozilla are probably mom-usable --- or at
least as much as proprietary apps are.

I recently talked my mom through saving an email attachment from AOL,
unzipping it with WinZip, and opening it with WordPad. After doing
this kind of thing once or twice, one learns that proprietary software
is pretty unusable by most people.

You're also getting dangerously close to circular logic; the debate is
(was?) over whether or not there's money in writing free software. I
thought Cygnus, Cyclic, and Sleepycat (not to mention this whole
Internet thing, which is starting to look like it might be more than
just a passing fad) answered that question pretty definitely in the
affirmative, but your statement above takes as its premise the
opposite: that there is no money in writing free software.

In summary: you're changing the subject, and your new argument seems to
contain the following false premises: there's no money in free
software, proprietary apps are innovative, proprietary apps are
mom-usable, free apps are either unusable or not innovative. It also
comes to the false conclusion that non-innovative apps are necessary
for "mom", who would probably do just fine with Wordstar 2.2 on CP/M
running on a 4MHz Z80 once she memorized the menus, and will do much
better with AbiWord, which has amenities like online help.

(AbiWord is a project of AbiSource, Inc., by the way. It's GPLed.
AbiSource is not a stock bubble baby.)

> > Cygnus was wildly profitable for ten years before they ever had any
> > outside investment, selling nothing but support and bespoke software
> > development for free software. EDS, Keane, Compuware, OriginIT, Ciber,
> > and Arthur Andersen sell the same things, albeit mostly for proprietary
> > software.
> Bingo, supporting proprietary businesses and software. That 70% that
> wasn't my topic in the first place - their product isn't software,
> software is a cost. But you can bet the companies they consult for
> depend on IP rights to survive.

You could bet that. In some cases, you would win. In others, you
would lose.

For the most part, companies depend on providing value to customers to
survive. A significant number of companies depend on IP rights to shut
out competition. Most big companies don't.

> You cant tell me for a second all those
> genome companies are looking for genes just for fun. They do it to get
> patents so they can make money. No IP rights, no gene hunting, no hired
> programmers, no hope of keeping up with the machines.

I don't know much of anything about "genome companies", but I just heard
that Celera just finished sequencing the Drosophila genome and is
making that information publicly available.

> > IP rights were necessary in the age of printing presses and movie
> > reels. They're probably still necessary for those media. They are
> > nothing but a crippling load of overhead for digital data on the
> > Internet.
> Crippling to whom exactly? If IP rights didn't exist for software, geeks
> would simply find something else to spend their creative energies on
> becasue they have to eat. Rootbeer isn't free, never will be.

Crippling to people writing new software, trying to get old software to
work, and trying to learn how to write software --- not to mention
people whose privacy is being invaded by intrusive spy software like
RealJukebox, whose wide usage was only possible because it was proprietary.

But, again, you're taking as a premise that software freedom precludes
income, a premise I thought you might at least treat with a little
doubt after I listed several companies that have been profitable for
many years in a row writing free software.

> > Free software enables programmers to deliver much greater value per
> > hour of work. Partly as a result, Cygnus's income per programmer is
> > considerably greater than Keane's.
> Sure it is, all that free help goes right to the bottom line :)

Well, Cygnus actually wrote the vast majority of the software they
support and extend. They don't get a huge amount of "free help",
although it's not insignificant. What their *customers* get is the
best compiler in the business, supporting the most platforms, with the
knowledge that they could fix bugs in it if they had to and the
knowledge they can go somewhere else for gcc support if Cygnus starts
to suck.

> > The quality of proprietary software is a disaster. It has always been
> > a disaster.
> Most software is a mess, proprietary or otherwise. Noone likes to test,
> or document, or any kind of quality control - the last 10% that takes
> 90% of the time. At least with proprietary software someone will get
> paid to do it eventually.

Have you seen the "fuzz" study? It's an empirical software-reliability
study; they took the basic Unix utilities shipped with many different
versions of Unix and fed them random input streams. A lot of them
crashed; the GNU and Linux versions were less than half as likely to
crash as the proprietary-Unix versions, despite the GNU and Linux
versions having more features.

They redid the study --- five years later, I think --- and found the
same pattern.

There is lots of anecdotal evidence that points in the same direction.
Just ask any Linux fan. :)

Again, you're making the dubious assumption that free software means
nobody gets paid.

<>       Kragen Sitaker     <>
The Internet stock bubble didn't burst on 1999-11-08.  Hurrah!
The power didn't go out on 2000-01-01 either.  :)

Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Sat Mar 25 2000 - 06:43:07 PST