[
Lists Home |
Date Index |
Thread Index
]
At 18:13 -0400 2005-04-13, Ryan Tomayko wrote:
>On Apr 13, 2005, at 6:02 PM, Bob Foster wrote:
>>Ryan Tomayko wrote:
>>><http://tinyurl.com/54f52>
>>>United States Patent: 6,880,125
>>>System and method for XML parsing
>
>>There are push and pull variant APIs for *all* event-driven
>>processes and any push variant is trivially converted to pull by
>>multi-threading or coroutines. This fact has been exploited in
>>communications since God was a boy.
True (well, not *quite* that long -- but very long before this
patent, which is what counts).
NOTE: I AM NOT A LAWYER.
If I'm reading the patent correctly, it's not much of a threat. It
seems to only claim cases where you have multiple APIs for invoking
the same parser; or possibly multiple parsers invoked the same way;
all in the same application. If I'm reading that correctly, then the
patent is very easy to work around.
It also sounds an awful lot like technologies that were floating
around in the mid-90s. For example, there was a product that could
either browse an SGML document straight from disk, or from a compiled
DOM-like binary.
Of course, that was *SGML*, not XML -- since this patent claims only
XML, that product (and similar ones) wouldn't absolutely preclude
this patent. But they'd make it relatively easy pickings on grounds
of obviousness.
Or, you could work around it by making your *parser* not read XML:
just have a low-level routine that inserts "****" or something at the
beginning, and *then* passes data on to the parser -- the parser
*itself* requires the ****, therefore it's not parsing XML.
Very weakly written patent, IMHO. Though it at least does cite *some*
non-patent prior art, which is better than many these days.
Steve
--
Luthien Consulting: Real solutions to hard information management problems
Specializing in XML, schema design, XSLT, and project design/review/repair
Steven J. DeRose, Ph.D., sderose@acm.org
|
