[Commons-Law] Re: The gates of intellectual property (Lawrence Liang) (Hasit Seth) (Sudhir) (Hasit)

[Hasit seth]

Hi Sudhir and Everyone,

    This is just a subjective take on comp. sci. research in an
attempt to answer Sudhir's query.  I do not believe that there is an
artificial divide between "basic" computer science research done in
Universities and "applied" stuff being done in the industry. 
Mathematics that relates to computer science is done largely in
Universities, but fundamental theoretical computer science (which is
some math anyway) is done everywhere.  Best of computer science
theorists Donald Knuth of Stanford is an active programmer though a
mathematician by training.  But Al Aho, who crystallized compiler
research, works at Bell Laboratories and does a lot of theory.  Some
esoteric stuff such as "hard" Artificial Intelligence has only
academic interest because industry sank a lot of money in it without
any good results.  But fields such as searching (Google), computer
vision (Siemens and tons of other companies), speech recognition (IBM,
Microsoft, Matsushita...) have a lot their theory and applications
being done both in academic and industrial world.  Historically,
terrific stuff has come out of  industry "incubators" such as Bell
Labs (UNIX) and Xerox PARC (Ethernet, Mouse, GUI, Laser Printers...).
 
      Bigger issue than academic/industry divide is the lack of major
breakthroughs in research.  I understand that many readers of this
list are interested in
Legal/Social/Ethical/Political/Emotional/Activism side of innovation
and not the technical aspects, but any discussion of these aspects of
invention without considering technology issues is bizarre.  Open
source is a reaction to this lack of major innovation in computer
industry.  For example, from a user's viewpoint why should I upgrade
from Windows 95 to Windows XP and pay any extra money for it when all
I get is better eye-candy on screen but nothing significant
functionally?  A group of hackers can get together and clone this OS
and applications and not pay the "windows upgrade tax" to Microsoft. 
That is what is Linux today.  Earlier such cloning was for academic
purposes or just for hacking fun - BSD is a clone of Linux, GNU is a
UNIX clone, GNU/Linux is still a UNIX clone.  Yes, it is "free" - that
is the innovation here! but technology wise? not a major breakthrough.
 To say this is NOT to undermine the TREMENDOUS efforts of
BSD/GNU/Linux  hackers,  because in first place writing the code and
then making everything work through a community effort is just a
fantastic achievement. For those who are interested in state of system
software research, Rob Pike of Bell Labs, a respected UNIX hacker,
wrote a paper in 2000 about how system software research has become
stagnant (here: http://www.cs.bell-labs.com/who/rob/utah2000.pdf)

     Now what could be a breakthrough in software?  No predictions. 
But take an example:  Continuous speech recognition where user speaks
normally to computer (with no special microphones, no a priori
training of speech models, etc.; this does mean an intelligent
conversation, but recognizing continuous spoken words syntactically
and semantically) is a tough problem, and any operating
system/application solving that will be a major breakthrough. 
Everyone - academia, industry and open source community is trying to
solve this problem since years, and patents are least of hurdles here,
though a lot of approaches (neither of which is entirely workable) are
patented.  And users will pay for such an invention, because they will
see it is a major breakthrough.  If this comes through
industry-patented tech route, people will still pay for it but its
spread will perhaps be a little slower.  Academic debates will
continue about why it is evil to have this patented, but people will
be more interested in getting hands on this technology rather than
debate about it.  If users pay for a better toaster or television,
then why not better software?  If open source community achieves  this
first, the spread will be perhaps much faster.

The point is not who does it, the point is how to do it? when will it
happen?  Right now the state of research in continuous speech
recognition is like Edison trying thousands of chemical compounds to
build a storage battery with just failures in sight ahead, but a lot
of hope ahead.  In many cases like these, invention is not a David vs.
Goliath fight between open source vs. patents, industry vs. academia
or industry vs. open community.  There may be skirmishes on the way, 
but in a bigger picture it is rather every-kind-of-effort against the
scientifically unknown and uninvented. For business to chase such
tough problems it needs money to keep payroll for programmers going. 
For open source community, it needs interest of the community to be
kept going to solve such long term non-cloning type demanding
problems.

       In short, invention is not simply Archimedes getting an
"Eureka" idea and running out of his bathtub (modesty and clothing is
not at issue here!)  straight to the Greek patent office and blocking
all invention related to the Eureka and hoarding billions in money by
monopolizing the use of "Eureka" through a patent.  This could happen,
but usually it is lot more complex than this.

Regards,
Hasit Seth

> 
> Message: 2
> Date: Wed, 16 Mar 2005 11:00:41 +0100
> From: <sudhir at circuit.sarai.net>
> Subject: Re: [Commons-Law] Re: The gates of intellectual property
>         (Lawrence       Liang) (Hasit Seth)
> To: Hasit seth <hbs.law at gmail.com>
> Cc: commons-law at sarai.net
> Message-ID: <5cbaf66cc761ec00466ab779c2768c84 at sarai.net>
> 
> Dear all
> 
> Thanks Hasit for that interesting post on the early computer industry. I
> would suggest that there is one important actor left out of this story -
> the University and Research Institutes!
> 
> Most accounts of computer software innovation rely on industry based
> innovation models. As you rightly point out a large part of such innovation
> is generating clones, marginally differentiated for the market. I am
> curious whether innovation in computer software breaks down into the Basic
> and Applied Science categories - where basic and fundamental innovation
> more often occurs in the University and Research Institute rather than the
> software company.
> 
> Best
> Sudhir
> 
> many firms are the level of innovation  On March 15, 5:04 pm Hasit seth
> <hbs.law at gmail.com> wrote:
> > Hi,
> >
> > Interesting article.  The New Zealand patent if it tries to claim XML
> > as being proprietary then it is utterly stupid.  Cannot comment more
> > without reading the text of the patent.
> >
> > I am not a big fan of Microsoft, but as a computer enthusiast I find
> > the characterization of Gates and Microsoft in early PC history is
> > terribly twisted.    Microsoft's DOS OS was the prime reason for
> > personal computing being affordable today.  CP/M the only major "PC"
> > OS around at the time was proprietary.  IBM's PC-DOS (developed by MS)
> > was expensive (or impossible at one stage) for PC Clone makers to
> > license through IBM.  Apple being the paragon of proprietary stuff had
> > no interest (and still doesn't) in licensing its OSs.  MS-DOS is what
> > ran the PC-Clone market.  Technical merits of MS-DOS is a different
> > topic, and GNU certainly didn't make personal computing happen.  If
> > one understands the difficulty that Linux is facing today in
> > supporting thousands of different types of hardware devices from
> > multiple vendors, therein lies the historical value of MS-DOS and
> > Microsoft.  Linux enthusiasts are doing an admirable job of supporting
> > drivers for a lot of hardware but still a lot needs to be done (e.g.,
> > Wireless cards with proprietary firmware are a pain to setup).  DOS
> > ran on every kind of PC clone, and Windows still supports the widest
> > variety of garden hardware.
> >
> > Yes, MS monopolized the Windows Application market, but this deserves
> > a closer look.  But before that on DOS platform, MS just provided the
> > OS - The Wordprocessing package was Wordstar, The Spreadsheet was
> > Lotus 1-2-3 and The Database was dBase III+  - all non-MS
> > applications.  All these non-MS application makers very quite slow in
> > releasing their Windows versions.  Most were so wedded to the text
> > based MS-DOS that they just could not make the transition fast enough
> > (e.g, dBase III+ or Wordstar).  Nobody wanted to develop applications
> > till Windows 3.1 became popular.  MS filled in this space (of course
> > with head start from knowing Windows codebase) with acceptable Apps
> > such as Word and Excel.  Developing these Apps is hardly a trivial
> > task where one can say that MS just hacked together a few pieces of
> > off-shelf software.  Just look at the effort required in OpenOffice to
> > understand the nature of the task.  Other OS Vendors lost faith in the
> > PC market - principal being IBM with a much superior offering of OS/2.
> >  MS continued developing both OS and App market, with entry barriers
> > getting higher and higher for other players.  Sure they were
> > monopolistic, but just look at Apple or IBM in similar circumstances
> > and they are no different.
> >
> > Software is not easy to develop, especially complex and large
> > applications.  For example, GNU's Hurd kernel is years in making
> > without any significant results.  It is just a damn tough job.  Linux
> > enthusiasts will rave about Vi and Emacs, GNU CC and Make as all that
> > you need to develop applications.  May be.  But MS's contribution,
> > from a purely technical and developer producity viewpoint, in Visual
> > Basic is just immense.  Nothing comes close in ease of developing a
> > GUI app - whether free or proprietary application.  I know there is a
> > ongoing Gambas project to close Visual Basic, but then like most open
> > source projects it is a cloning effort.
> >
> > If we accuse of MS of rehashing and repackaging existing technology,
> > open source software is no different with their best efforts being
> > admirable clones (some better, some worse) of existing proprietary OSs
> > or Applications.   While MS is a terrible company, a classic
> > monopolist at many times, its historical role in bringing affordable
> > computing to masses cannot be just wished away.
> >
> > Regards,
> > Hasit Seth