On Thu, 12 Nov 1998, Anand Rangarajan wrote:
> You do not seem to realize that Microsoft could legally sue
claiming
> possible patent and copyright infringement. SAMBA, Wine, KDE,
Gnome and
> other programs are potentially at risk. Only bad PR—M$oft
goes after
> unpaid, altruistic programmers—might stop them from taking
this step.
Hmm, let’s see if we can make this another linuxtoday article,
as this is around the tenth time I’ve made the points below in
responses to individuals who still fear (with some justification)
Microsoft’s immense wealth and power. First, let’s address these
concerns; then let’s look into my crystal ball in more detail and
try to divine the Future.
SAMBA and Wine, perhaps yes since they are reverse engineers of
published MS specs (talk about the pot calling the kettle…) but
this will not really hurt linux and indeed attacking SAMBA, or even
(much more likely) messing with the SMB spec to try to break it
would probably hurt them more than it helped them — network
managers don’t like it when companies break their client/server
operations and if anything will be pushed further towards true open
standard products.
I don’t see how Gnome is at risk, and if they tried to attack it
they couldn’t get away with just suing linux developers — COSE,
CDE, and a whole slew of related products are being developed by
various “open” consortia. Many of these have pockets just as deep
as or deeper than MS’s. And finally, in the past, courts have been
fairly unsympathetic to “look and feel” protection or “published
spec” protection — something that has benefitted Microsoft
tremendously as it stole Apple/Xerox-PARC’s GUI lock, stock and
barrel to turn it into “Windows” and benefitted everybody as the
Gnu project cloned Postscript in a carefully documented open
development process that resulted in ghostscript.
It is not illegal to develop from published specifications. If
one can document that one used NO COPYRIGHTED CODE in the
development of your product and leave off precisely duplicating the
interface of an interactive product, I believe a judge would simply
throw the suit out after a simple review to establish those facts.
Patent protection (these days) is a murkier issue — one can
(sometimes) patent something like an ABI or API and then publish
it. As I understand it, if a company does this anyone can use the
standard for themselves or work with it in the public domain, but
commercial implementations of it must pay a license fee. (That is,
you legally can (if you are able) go out into your garage and build
a better mousetrap from its published patent description, as long
as you don’t try to sell it without a license. I could be wrong
about this, but I think that’s how it works.)
I believe that this is what the Halloween document is
effectively advocating as a strategy — subvert the open standards
process by modifying an existing open standard with “patented,
proprietary modifications”, building must-have software that uses
the hacks, and freeze out competitors stuck with the open standard.
Pretty much what they are trying so hard to do with netscape,
email, and Internet standards in general right now, actually.
Standards are the Enemy to a company like Microsoft, because it is
so easy, really, to craft a product that meets them that
competition is high, prices low, and real innovation or added value
is necessary to make money, at least compared to the
fish-in-a-barrel ease of “competing” with a proprietary product in
a closed market.
Anyway, with the DOJ already on their case and occupying a great
deal of their legal energy, I don’t think they’ll find it a really
good time to try to quash “competition” with legal suits of dubious
merit. I expect their lawyers will limit their “punishment”, if
any, to getting fined and their wrist slapped, but the knowledge
that they are under ongoing scrutiny and getting bad press as it is
ought to be enough to keep them out of court fighting what they
still perceive of as fleas.
The reason I presented such a humorous view of the Halloween
document is that I really don’t think that ANY of these
possibilities — legal action against OSS developers, further
subversion of the standards process followed by FUD tactics, or
painting themselves blue and rubbing mud into their bellies will
enable them to avoid their Doom, because this isn’t where the Doom
is coming from. They’ve finally gotten the point that linux,
freebsd, gnu, and OSS in general are real competitors but it is
clear from their document that they don’t really understand
Why.
In my opinion, the WORST possible outcome for the linux
community is that they could actually lose the DOJ case and be
broken up like ATT, forced to separate their OS/Compiler division
from their consumer software division in separate companies. It is
the one thing (other than a fundamental change in Microsoft’s
corporate strategy that accepts far lower margins) that might save
them. Here is Why Microsoft Should Fear The Penguin, by rgb.
Remember, folks, you heard it here first.
MS has finally succeeded beyond their wildest dreams in creating
a software monopoly any Rockefeller or Vanderbilt would be proud to
own by ruthlessly co-opting each and every really successful piece
of software invented and developed by somebody else and sold for
their operating system. In many cases they have driven the actual
inventors right out of business, in others they have just bought
them out. We can go down a neat little list of companies that have
been driven bankrupt or marginalized while the immense profits that
they SHOULD have earned from their innovative products have gone to
line Microsoft’s already swollen pockets. Remember Borland? It
(Phillip Kahn) invented the integrated compiler and failed utterly
to keep MS from co-opting it so that Borland has a relatively small
share of the total integrated compiler market today. Remember
spreadsheets (Visical, Lotus, Borland, the list goes on) before
Excel? Remember the many word processors (Wordstar, Word Perfect)
before Word? Novell struggles on, but they don’t have a chance
(heck, even if they manage to hold out against Microsoft, there is
linux:-). Who sells more encyclopedia software — Brittanica or
Microsoft? Don’t know for sure, but I’d be willing to bet I know
the answer. Who has muscled their way directly into the web browser
market by giving away “just a little sniff” of their product on the
streets, just as at one time they “encouraged” vendors to bundle
Works or various elements from their current Office suite to coax
folks away from Lotus and WP? Who has a direct challenge to Oracle
mounted? Who has somehow mysteriously “broken” Apple’s Quick Time
multimedia product (making it vulnerable to the usual FUD strategy)
by making small, undocumented changes in the Windows protocols
while ensuring that their own competing multimedia products
continue to work? And I’m sure that the public half of the story is
only a tiny bit of the whole story.
Only games remain relatively untouched as the market is too
volatile — Microsoft only buys out games when they’ve been around
for long enough to be considered classics.
Why, then, is this immensely successful strategy self-defeating?
Why are they doomed?
Because at this instant in time, there are two strategies
realistically available as corporate/marketing plans to software
developers writing for windows. If you succeed in coming up with an
Idea, invest money, blood and sweat in developing it and creating a
market for it, and actually get to the point where your market
starts to make you a real profit that, under ordinary
circumstances, you could exploit (given your head start) basically
unchallenged for 5-10 years, you can:
a) Sell out to Microsoft. No kidding, there are companies out
there where this is basically their openly stated business goal.
Quite a few of them, even today. What wimps.
or
b) Accept the fact that when you succeed, Microsoft will clone
your software using the talents of its immense farm of programming
talent (carefully skirting Look and Feel issues with the aid of the
many very expensive lawyers on their staff) and take maybe half
your market away in the first year their clone hits the streets. If
necessary (because your product is still better), they’ll just give
it away until your sales are driven down enough to starve you for
revenue and reinvestment capital, or perhaps they’ll will “need” to
make just enough perfectly reasonable changes in their OS/Windowing
system that your developers are always behind their developers (who
of course, have unparalleled access to OS/Window support) in the
development cycle. Eventually their clone product will outperform
yours or yours will break, making you vulnerable to FUD.
Don’t worry, you can “survive”. If you are lucky, they’ll let
you hang on in some sort of half life for those 5-10 years, and you
might even stabilize with a steady market share of 1/4 or so if
your product is phenomenal and achieved significant name
recognition early enough. Microsoft recognizes your value, after
all — if you innovated once you might do it again and they were
never ones to kill off the goose…
Microsoft should abandon the funny looking Windows logo and just
hoist the Jolly Roger. Blackbeard was a piker by comparison. The
famous tactics of the Roaring 90’s robber baron capitalists (the
Company Store, the manifold shenanigans associated with the rail
lines and oil wells and coal mines) are crude by comparison. Even
the moonshiners and gangsters of the prohibition era, with their
“protection” schemes and shadow governments are romantic idealists
compared to Microsoft and its Business Plan. Even Intel has been
directly threatened by Microsoft, and you can bet that if MS were
to delay software support for Merced even six short months it could
cost them billions in profits, let alone revenues — they are
indeed vulnerable.
From the above, one could easily be tempted to conclude that
even though there >>is<< a temptation (to wimps!) to
take the mouse’s share of a really large number of covered desktops
and just give Microsoft the lion’s share of your hard work (a mouse
can certainly get fat — for a mouse — on the crumbs from the PC
desktop table) a software developer would have to be brain dead
stupid to develop any really GREAT idea for a Microsoft platform.
Really smart computer scientists and software developers, the ones
who like to work for themselves instead of being a vassal in a
feudal corporate structure, the ones who have the Really Great
Ideas, will Not Develop for Microsoft’s OS or Windows Environment.
Period. I think that we’ve been seeing this trend develop for
several years now and that it is a major thing that has fueled the
development of linux and freebsd.
This is what I REALLY think will cause Microsoft to change or
die. Because right now, developing software for Microsoft is like
brushing the teeth of a Great White Shark with a piece of raw
steak.
Microsoft, contrary to their public press, is not much of an
innovative company — they don’t have to be. They simply engulf
amoeba style all the real innovations as fast as they appear. Like
any big, amorphous predatory blob, they need a steady diet of Ideas
just to maintain their existing overgrown body mass. But they’ve
destroyed the symbiotic relationship they once had with their food
supply and are WAY out on the predator-heavy part of the
predator-prey cycle. If you know this particular set of couple
ODE’s, the inevitable outcome is starvation and death for the
predator. linux and freebsd represent an additional dimension in
the predator-prey landscape into which the smarter prey can
“escape” without being eaten, and has what I believe to be a
healthier and more stable ecology. And of course, the
better/smarter prey are doing just that; it’s just (market)
evolution in action.
From this analysis, one doesn’t have to have too much
imagination to see why Intel, Oracle, and basically ALL the big
software houses that haven’t been absorbed by Microsoft (yet) are
backing linux in a big way. They have finally realized what years
of dithering and pointless squabbling over proprietary advantage in
the OSF and Unix International communities didn’t teach them — it
isn’t just hype: they NEED to have an open standard, open source
operating system as the basis for their software to be even
moderately assured of actually reaping the full benefit of their
development efforts and making a fair profit on their innovations.
The transient benefits of being the setter of a proprietary
standard are not worth the long term costs of competing against a
whole raft of companies who are ALSO trying to set a proprietary
standard.
Unfortunately, Microsoft has forgotten that it was the software
written and sold by OTHER people that made them the company that
they are today. As of right now, a whole lot of the same KIND of
folks that were writing software for the early PC’s (gimme ROM
basic and I’ll make a Million) are now writing software to run
under linux or freebsd. There is the same sort of gold-rush, can
make a fortune fast feel to it.
I give Microsoft at most two years of free running before they
really start to feel the burn.
Pass the Crystal Ball, please. Thank you. I see it, what the
spirits reveal — Aha!
Linux will be the first (or possibly tied for first with
freebsd) to provide >>absolutely transparent<<
migration to Merced when it comes out and Intel realizes this. (So,
by the way, does Sun. In the latest version of Sun Expert Sun
Experts crow about the fact that there is expected to be a 6-12 lag
porting “popular 32 bit Microsoft software” to the new 64 bit
environment, and that folks might be understandably miffed at
spending top dollar for a 64 bit cat’s meow system to get Pentium
performance from an emulator — if the emulator works. They view
this as an opportunity for their Sparc/Solaris migration pathway to
make up lost ground.). With linux (or, to be fair, any other
properly written Unix code), all that will be required to enable
“instant” ports is for the kernel itself and the compiler itself to
work. I forsee that Intel will simply ensure that Linus et. al.
have whatever resources they might require (including early release
merced systems) to make this happen. Intel of course, has precisely
the resources required — they are undoubtedly writing the
reference compiler for the CPU, for example — imagine what will
happen when they provide their reference compiler, in spec or in
source, to Gnu — gcc will be THE premier compiler for the
platform.
With kernel, compiler, and device drivers ported, any linux
developer, using gcc and their existing 32 bit sources, can have
“64 bit applications” on the market THE DAY MERCED IS RELEASED with
a few simple modifications and a single recompile. I’ll bet you a
nickel that on that day, Red Hat Linux (among others) stands ready
to deliver the canned operating system and the entire suite of
commercial linux software ready to run. If Intel times it just
right, linux software developers can actually start distributing
merced-ready binaries six months before the actual release date on
their non-merced CD’s, so that one doesn’t even have to “upgrade”
(although the temptation to gouge just a leetle bit from their
customers for the upgrade will probably prove irresistable to
some).
History has shown us just how fickle and volatile the computer
marketplace is, and how much it has become and is continuing to
become a commodity, low margin operation. Microsoft’s ascent has
been rapid, their position is unbelievably high — they stand on a
peak far above where any have gone before — but they abandoned
their climbing pitons below in the rock face in their rush to reach
the top.
The top crumbles, their footing is looser than they think, and
it is a LONG way down. And scrabbling around underfoot, tripping
them up, there is a penguin…
rgb
Robert G. Brown http://www.phy.duke.edu/~rgb/ Duke University Dept. of Physics, Box 90305 Durham, N.C. 27708-0305 Phone: 1-919-660-2567 Fax: 919-660-2525 email:rgb@phy.duke.edu