"RAND" licensing concerns...

As a developer writing software that is intended for use with the Internet
that is developed under Open Source licensing terms, I write to you with
concerns about the proposal to allow reasonable and non-discriminatory
(RAND) licenses on patents for W3C specifications.

Much of the software that makes the Connected Internet, commonly
referred to as "The Internet", work is of an open source nature as are
most of the protocols proposed to date or that are in use.  The development 
of the Internet as a whole as well as the World Wide Web has benefitted from 
the philosophy of non-patented or non-royalty technologies and 
specifications.  To bring RAND as an option for protocols is to close off the 
Internet and development of systems for the WWW to those who have large sums 
of money and corporations.  While it benefits those groups of individuals, it 
doesn't benefit the world at large (Claims of "better" technologies being
offered because of them being patented, etc. not withstanding, it needs to be 
stated again- we got where we are right now WITHOUT this, so why do we need 
it again?  Also, there is nothing said at this point about whether or not 
being patented really makes it any better than an unpatented implementation.) 
as there are groups of people that do development of technologies for 
educational and altruistic reasons.  

These groups would most likely prevented from participating because most 
"reasonable" licensing tends to involve some money outlay or licensing often 
includes extremely narrow allowances for use when it doesn't.  Witness the 
"reasonable" licensing that MPEG uses for the Layer3 audio codec- the open 
source developers were required to pony up some $5US royalty per encoder 
developed and downloaded, hence they had to stop development even though they 
didn't use Fraunhoffer or Thomson's code.  They haven't pursued the decoder 
developers yet, but the licensing is equally problematic- $50k-100kUS up 
front or $0.75US per player.  The result from this was the development OGG 
Vorbis codec which is open source and has not patent encumbrances.   Witness 
the "reasonable" licensing that is in place with the GIF file format that 
uses Unisys' LZW algorithm for compression.  The royalty is 1.5% of the 
purchase price or $0.15US, whichever is greater, for the developers of 
software using the algorithm.  If you run a web site, it's some $5000 dollars 
per server- IF you're running a non commercial site.  If you are, it is much 
more expensive.  The result was the development of the PNG image format.  
These royalties seem "reasonable" if you think in terms of traditional major 
corporation business.  They aren't so in the context of Open Source or Free 
Software- in most cases, they are the kiss of death.

The majority of the WWW is served by an Open Source program, called Apache.  
With RAND, it is very conceivable that it would not exist as it would very 
likely require a per connection or per server royalty charged to the 
_developers_ of the software.  Most open source projects are made up of 
volunteers developing the code and do not comprise a group of individuals in 
the sense of a company doing business.  Patents and royalties make no 
distinctions in most of these cases, so the work would very likely not get 
done or someone would come up with another defacto standard that doesn't have 
patent encumbrances.

RAND only benefits a company or a group of companies that seek to
gain by getting everyone use THEIR patent and by doing so pay them
money.  In many cases, it doesn't benefit the rest of the community because 
there's often no room to develop improvements and moreover no desire to do so 
(If you owe, even for a dramatically improved solution, why even bother...) 
or they spawn efforts to come up with non-patented alternatives that might be 
better spent developing improvements on the current ideas.  RAND does NOT 
prevent discriminatory licensing- a company can sign any agreement saying 
that they'll license it, but it requires an expensive court case to prove 
that they didn't.  Also it takes an expensive court case to prove that a 
given company cut another company a break on the licensing, and didn't on 
another, and so forth.  The only way to really prevent patent problems is to 
require that the company submitting patented technology as part of a W3C 
specification officially grant a royalty free license to anyone implementing 
the specification as it's outlined in the documentation or any future W3C 
specifications that touch on the patent's coverage.

I urge you to not consider RAND as a viable alternative to what is currently 
in place with regards to patents that is in place right now for W3C 
specifications.

-- 
Frank C. Earl
Earl Consulting Services

Received on Saturday, 29 September 2001 22:57:07 UTC