Lists Home |
Date Index |
From: Rick Marshall [mailto:firstname.lastname@example.org]
>just for the record len, seeing as we've spent a lot of time on this....
>redhat is now suing sco, essentially for mischief, for claiming ip when
Yes I read that in the evening paper here. They are also setting
up a defense fund for others. So now, they are willing to pay
the legal costs for customers rather than expose them to the GNU obligation
to defend themselves. An odd twist on indemnification insurance.
I think the open source vendors are in trouble although
not open source itself. Customers will reconsider GNU.
In the next story, they were touting that a version of
linux has finally been certified for higher security mission
critical systems at a cost of half a million dollars, but that
this only improves the market position for government buys given
the dearth of desktop products and the continuing complexity of
working with linux on the desktop. At the same time, the indemnity
issue for government buys is getting increased notice as the
procurement officials realize that GNU makes them the litigated
party. I doubt it's fatal but a lot of people will keep an eye
on those suits and counter suits.
>so ip can protect what you've done, it can also, it is alleged, be used
>as a threat and to manipulate the market
Welcome to business.
>imho the w3c needs to be very careful about this binary business
That is what some are trying to work out. One common approach is to
ensure that each document in the process covers the contingencies
for the next step. For example, the workshop is a little dangerous
if any of the presenters provide information which taints further
work. In an RFP, one details all obligations for submission including
for example, recognition that the submitting the material obligates
the submittor to the IP policies of the organization posting the
RFP. If a submittor takes exception to any provisions of the RFP,
they must do so in a cover letter so that the RFP evaluator can
assess these and decide whether or not to open the body of the
proposal. This avoids trip wires. I suspect no one has thought
through these issues for a workshop. This does not mean that good
technology won't be offered. It means that the offerors who
are basing the offer on IP they own have to acknowledge in advance
that they are accepting the terms and conditions of the W3C.
>i'd rather see the rdbms model where the external view (rows) and the
>access method or api (sql) are are ascii and in the case of sql - a
>standard; while the storage of data - possibly binary, and the
>optimisers that go with it are as proprietary as you like with all the
>ip and other stuff that suits
To me, that is almost the same thing as saying XML binaries (the
XML infoset items such as elements, attributes, etc) use a completely
open specification or standard even when compressed. Because technically
we must do this, the text node content data types can be treated
individually. For example, in X3D, geometry compression is treated
individually. I still want these to be open. I'd rather bite the
bullet on performance than pay the toll, but that is my opinion only.
I believe 'good enough' technology will be offered for that and that
binaries for some XML application languages are inevitable. The
question is what is sharable among them.
>the important principle being that the data access and interchange are
>not ip owned by any one company or individual
We can state that principle but can't make it stick unless an offeror
signs up to the IP policy of the W3C or any other organization working
with it. Remember, it can mean standardizing on 'good enough' and
then watching proprietary solutions walk away with the market.
>xml really should stay the same - the formats are public and specified.
>if a company wants to develop and market a system for high speed
>transfer of xml data between systems - good luck. if they want to
>release a high speed xslt processor - good luck. if you want to offer a
>publicly available service - then stick to the ascii, xml 1.0 and
Even today, nothing stops a company from copyrighting a DTD or Schema
or even patenting based on that. It just can't be submitted to the W3C
unless Tim Berners Lee allows it. The W3C patent policy still comes
down to Director's choice. I'd hate to have that job but I'm glad
he does. TimBL has a long history of doing the right thing.
>fwiw i think standardisation in this area will only limit innovation and
>increase the costs of participation for smaller companies and
At some levels and for some companies and individuals, yes. As
with gif, if any company pays the license costs and others buy the tools, it
isn't that much of a problem. What is happening is that for core
that reach a large scale of acceptance, there is very little if any profit
or incentive for individuals or small companies to produce these. The
fact of standardization at scale limits innovation. I don't see that
doing bad things to the market; it is doing bad things to companies who
don't get off the lily pad before it sinks. Commoditization favors the
large vendors in any market.