|
|
|
| Authored by: Anonymous on Tuesday, May 22 2012 @ 09:48 AM EDT |
From the notes in update 5:
A computer can not make a numeric
reference symbolic later. But he told the jury that to figure out if a reference
is numeric or symbolic, they have to consider what happens downstream with it.
They should look at the value contained in the instruction stream. If this value
refers to an actual memory location, it is numeric.
This is
worrisome. What happens downstream has NOTHING to do with whether a reference
is numeric or symbolic. Whether a value in an instruction stream "refers to an
actual memory location" also has nothing to do with whether it is numeric or
symbolic.
Despite valiant efforts, the court's interpretation of technical
concepts is drifting farther away from reality, and I fear we're going to end up
with a result that is based on these incorrect
conclusions.
---
Regardless of who wins, I imagine the appeal process
will be even more divorced from reality. Watching it happen is going to be
painful. It will probably be won by one side or the other on a gross
technicality, and there will probably be sound technical arguments excluded
solely because one party or the other failed to make them now during the
trial.
We techies have a phrase we use to describe the onerous experience of
complying with processes that are a bit divorced from reality (anyone who has
worked for a large company has probably encountered this phenomenon of "process
run amok".. for example, IBM has an HR department of more than 10,000 employees,
with all of the process and bureaucracy you would expect from such a
juggernaut).
Anyway, the expression we use in this situation is "pleasuring
the process". It means "do the least painful thing you can think of to fulfil
the process's requirements, so it will leave you alone". [ Reply to This | Parent | # ]
|
| |
| Authored by: Ian Al on Wednesday, May 23 2012 @ 05:23 AM EDT |
I just searched out three books left to me by my engineering father.
The one I want to discuss is 'The Solution Of Triangles, Vol. Two,
Trigonometrical Tables'. Essentially it was five figure tables of Sines,
Cosines, Tangents and Cotangents by degrees and minutes. The book was from
'Machinery's Yellow Back Series'. I'm not certain whether the degrees and
minutes constitute numeric values or text labels.
He used them to bore precisely positioned locating holes for plastics moulding
tools. They are lookup tables. They could be used to make 'A' frame trusses for
roofs, drilling cylinder head bolt holes, or drawing abstract triangles on a
piece of paper with a compass. Using them is a math activity.
In patent '104 we are told that looking something up identified by a text label
in a table is a feature of the invention. If the lookup is done to find the
location in memory of a variable name or a method name then this is not
protected by the patent.
If the lookup is done to find the location of a virtual processor machine code
instruction, then the timing is important. If it is done just after compilation
of the source code then that is defined as 'static' and not protected. If it is
done before the program is offered for execution then that is 'static' and not
protected. If it is done only when the 'phone user asks for the app to be
executed and is done just before the app is executed, then that is 'dynamic' and
is protected by the patent.
The penultimate and the ultimate use require more or less the same resources in
the 'phone. The protected method means that the first run of the application is
slower than the unprotected version and then there is no difference because the
executables are then identical.
The app is any app and it does not matter whether the lookup is done
'statically' or 'dynamically' because the result is identical. The patent is not
limited to any particular virtual processor. It is not protected if the
processor is not virtual. It is not protected even if the non-virtual processor
uses a lookup table to micro-code to execute the instruction.
'104 is a patent on the use of math tables narrowed down by arbitrary terms
defined within computer science by the judge and also narrowed down by the time
and place that the math tables are used. It is further narrowed down to an
abstract concept of a Virtual Machine which is only distinguishable from a
platform to run interpreted Basic source code by assigning computer science
terms to the timing and the nature of the workings of the platform. It is
further narrowed by the abstract concept of reducing the program language
functions down to a set list of functions analogous to a microprocessor
instruction set.
However, all intermediate forms of object code must, by definition, do the same
because the final execution is on a physical processor with the 'virtual
processor' only being a software artifice to help porting the platform to
different processor based computers.
The bottom line is that it did not matter when my dad used the tables or what he
used the tables for, the tables did not magically turn into new inventions
depending on the use.
I won't quote from the book: it is protected by copyright by the Machinery
company, and dated 1958. It doesn't seem to have any patent marking, though.
---
Regards
Ian Al
Software Patents: It's the disclosed functions in the patent, stupid![ Reply to This | Parent | # ]
|
|
|
|
|
