Re: [dvd-discuss] Gedanken Experiment -Unix and Norton

[microlenz(at)earthlink.net]

On 9 Jul 2003 at 9:42, John Zulauf wrote:

Date sent:      	Wed, 09 Jul 2003 09:42:33 -0600
From:           	"John Zulauf" <johnzu@ia.nsc.com>
To:             	dvd-discuss@eon.law.harvard.edu
Subject:        	Re: [dvd-discuss] Gedanken Experiment -Unix and Norton
Send reply to:  	dvd-discuss@eon.law.harvard.edu

> 
> 
> microlenz@earthlink.net wrote:
> 
> This one piece out of order...
> 
> > ... the kind of editing that would be
> > necessary to transform database to a copyrightable work would also negate the
> > utility of a database and 
> 
> If you define database narrowly to only mean a complete transcription of
> offline records into online form, I agree.  I think the term is
> broader.  Back to the "Norton" example, Norton (check their website) has
> decided (editorially) that "Spyware" like Gator et. al. is not "a virus"
> and therefore won't include it in it's virus signature database, nor
> remove it (grumble, complain).  While the virus definition file *is* a
> database, it reflects a set of conscious editorial decisions on what
> does and does not constitute a virus.

Does it? And if someone merges the facts from AdAware does that constitute a 
NEW editorial decision. And what if one writes a program "1-end of NAV" add" 1-
End of AdAware" and "look for *.tmp" files then fix, delete and get put happy 
face on screen does that get copyright as editorial decision? What's been done? 
What creative act has been accomplished? What has been progressed beyond the 
next virus threat?

<alert reader may note. I am questioning the notion of Editorial decision. >

> 
> Now back to the top of your post.
> > 
> > Discretionary Editing of a database is an oxymoron. A database is simply a
> >  SYSTEMATIC collection of facts. A database that showed editorial
> > discretion in what goes into it would be totally useless as a database.
> 
> I was using "database" in the broadest sense.  An online encyclopedia
> would be a database in this sense.  

Not so. There is prose there, organization of material and presentation.  I can 
give you the formula for Tensor analysis -FACTS. These cannot be copyrighted 
but a description of raising and lowering indices would not be facts, although 
given the technical nature there are limits to originality. 

>A discography/bibiliography of the
> "greatest rock and roll music" would be a database in that sense.  

How so? John Z. thinks that 1,2,3,4,5,.....,N are the greatest of rock and 
roll.
Joe B thinks that 1,2,3,4,5,6,.....,N,N+1,N+2....N+1000 are. 
Originality...no..Joe B just agrees with John Z to N. But Joe B read John Z's 
book...well Joe B. just happens to agree with John Z and so is that copyright 
infringment?

>Both
> the encyclopedia and the "greatest" databases will certainly reflect
> editorial discretion. 

Oh contraire...is a lack of difference of opinion copyright infringment a 
priori? 

> One would expect to find and "Wycliffe" (the
> first to publish a bible in his local (non-latin or greek language) in
> an encyclopedia, but not entries of all of the others listed in "Fox's
> Book of Martyr's".
> 
> > Editing would be censorship ("No there was no Holocost otherwise
> > we'd have something in our database", "What of NYT, 1946?","Sorry our editor
> > decided to use his discretion and did not put that in our database") 
> 
> The winners write history.  I shudder to think what a post war Nazi
> "ministry of truth" would have done to that piece history.  
> 

Winners losers...this is not the issues of war but the discussion of how to 
continue a peace...the creation of ways of sharing information...<blast that 
soapbox..it's always so hard to step of of>
> 
> > >
> > > As for programs, these "facts" are in their entirety the "editorial
> > > discretion" of the programmer including,
> > 
> > Lets investigate this one
> > 
> > > the choice of variable names,
> > 
> > index, index1, index2..count, count1, count2? Ismpl(i),Qsmpl(i),ARRAY1(I),
> > aRRAY2(J),.....AND LETS not forget X,Y,Z, i,j,k
> 
> Yes, even while trivial these choices do reflect creative decisions and
> even cultural influence

If trivial, then they are NOT creative or that trivializes creativity. Cultural 
influences? I'm sorry my good friend who wrote FORTRAN programs for a well 
known company with variables in Russian phonetically spelled in English is not 
entitled to copyright protection even if he is a well known mathematician at a 
well known school 30 yrs later ....this is one aspect of the functional 
doctrine that Judge Kaplan thought so brilliant that has consequences he cannot 
fathom.

> 
> int todo; 
> 
> is this "everything" (Spanish)-- i.e. the total of all elements, or "to
> do" (English), the number of elements left to process?
> 
> float mach;  /* mach number or the amount to make (machen) German,
> nicht?)
> 
> <tangent>
> I once spent a long night debugging a FORTRAN77 supersonic Euler
> equation solver because I'd forgotten
> 
> IMPLICIT NONE
> 
> and for some reason the mach number computation was wrong... (bonus
> points for the solution)
> </tangent>
> 
> #define SICHER true
> 
> Even the style
> 
> if ( condition ) { 

So if I decide to put two spaces between the ( and the { does that style make 
it copyrightable....and lets look at 1000 people programming this. I'll bet 
there's only a few ways it really will get programs no infinite variety of 
creativity there more caprisiousness

> 
> is an editorial choice vs. (the evil and useless)
> 
> if (condition )
> {
> 
> Individually these are no more copyrightable than i++ or "1 whole note
> of middle C", but collectively they represent not facts, but creative
> editorial choices.
> 
> > 
> > > the structure and order of the operations,

The order of the operations on paper is merely the placement of a 
multidimensional graph depicting the variables and transformations of them onto 
a 1 dimensional manifold. Nothing unique. Noting original just stupid ways, 
less stupid ways, and easier to understand ways. Nothing copyrightable there

> > 
> > The structure and sequence of operations generally determines what is done.
> 
> If there are five tasks to be done and some of which are order
> independent, then the author is making a editorial choice of what order
> to present them.  

But also a TOTALLY IRRELEVANT ONE! Pick another one and that gets copyright 
protection. 

No more relevant than alphabetizing the phonebook. Granted that is a better way 
than using street names but so what?


>This can be base on a traditional order, an order
> which is clearest or most maintainable, or any other number of
> considerations not driven by functionality alone.

All of these are undefinable.

> 
> > There is less creativity there other than optimization but optimization is not
> > creativity it is algorithm development and discovery. 
> 
> Is there a clear line between optimization and algorithm development? 

Yes there is....one desires a faster and/or more accurate algorithm. You don't 
develop a better algorithm if you don'tneed to.There's a lot less work done in 
FFT algorithms these days than 30 yrs ago

> At some point doesn't an optimization change the definition of the
> process enough that it is a new process?

Only if a new family of algorithms are developed. That BTW is the trouble with 
trying to allow patenting of algorithms (coworkers and I have discussed this 
for about 15 yrs I might add). One should NEVER speak of AN algorithm but a 
family of algorithms. Each tends to build upon the other and some minor 
modifications of an algorithm may lend it very useful in one application but 
not another (I once saw a paper by a guy from Cray Research that had optimized 
the FFT for a Cray 1. His FORTRAN was faster than the assembly language FFT 
currently in their library!) Imagine the claims and counter claims in a series 
of say 100 interlocking patents!

> 
> > It is not covered by
> > patent and never should be. 
> 
> Optimization or algorithms?

The latter. The former is a process.

> 
> > Furthermore, since it has never been covered before
> > should it ever be, the patent office must cover the UNPATENTED prior art
> > before granting any patent (exercise for the alert reader. In expanding the
> > scope of patents, the previouslyunpatented prior art invalidates a patent for
> > it but how does the patent office check for THAT.)
> 
> Especially since source code is unpublished by and larger.  This is why
> the unpublished copyright granted to software is a disaster.  

No argument. That is totally illconceived. The bandaids put on copyright to 
appease some special interest must be removed and a new foundation for it 
constructed (I'm paraphrasing J.Litman)

>I'd rather
> have a twenty-year patent cover a novel process (something REALLY novel
> like... skip-lists
> http://ciips.ee.uwa.edu.au/~morris/Year2/PLDS210/niemann/s_skl.htm
> than hide away all useful code for 95 years (and even then it isn't
> published).
> 
> > >the selection of algorithm,
> > 
> > Bubble sort vs heap sort.....not that exciting a programming task...after
> > looking at the 1.9GHz Pentium 4, 500Mhz buss, 500MB memory and 40GB Harddisk I
> > have at work, this is not much of an issue except where people are really
> > pushing things (e.g, computer graphics or ray tracing , communications
> > simulations)
> 
> What about those of us trying to fit this same functionality on 400MHz
> processor and run at low power -- the embedded space.


Lower power in the embedded space generally means getting a decent algorithm 
and then not wasting clock cycles when i don't need them (ie put the processor 
to sleep). The things I've worked on are generally interrupt driven cyclically 
and look like 
(INTERRUPT. Do A idle....) repeat N times
(INTERRUPT DoA, DoB idle) repeat Ntimes
(INTERRUPT DoA, DoB doB idle) repeat Ntimes

and you can see the CPU cycles (and Power go up each time)
> 
> In terms of general choices, what about heap vs. quicksort vs. skip-list
> vs. ...  How about the choice of a dithering algorithm or a time-step
> CFD algorithm?  

So why should society allow one to create a monopoly that precluded everyone 
else from deciding that a good way to something is A then B. Why should such a 
choice be now labeled Novel or Original when the correct label is "better way" 

>Let's look at multimedia -- this stuff will consume
> every cycle you throw at it (and will for the next few years) Filtering
> kernels choices,  audio encoders, lossless or lossy compression, motion
> estimation approaches, bitrate control. 
> 
> Lot's of room here for creative choices that strong impact the quality
> of the
> 

What creative choices? Only better ones. So once the better set is known, you 
now have a monopoly!
> > 
> > > the syntactic style ("{" on the same line as the conditional...
> > > please!), and myriad other unique creative details.
> > 
> > Writing style is not copyrightable. How much Hemmingway or Hammitt hash was
> > created in the 30s and 40s? The placement of comments or good programming
> > style is not the stuff of copyright (although it can be used for investigation
> > of copyright infringement or theft of trade secrets)
> 
> Yes but it distinguish coding from "mere fact".  T
> > 
> > BTW I will admit to having a very distinct programming style. Look at Appendix
> > E in the back of B. Sklar Digital Communications 1ed Appendix E. Guess which
> > sections were mine and which were my coauthors.
> > 
> > But much of this applies to Source code...not object code and that is
> > generally a trade secret. But even if it is not, there is considerably less
> > originality in it than most of the court decisions would have us believe.
> 
> Don't underestimate us programmers and engineers.  

It's not a question of underestimating it's a question of what standard should 
be used to grant a monopoly! Is that monopoly desirable! Is there alternatives 
available! The question begins to be "when does this begin to be locking up 
ideas" There are simply NOT that many choices when developing programs. The 
language is too restrictive by design. 

As 25 word sentence chosen from a vocabulary of 1000 words has 
47641862536236518640933948075167736642053976275040 choices

out of 3000 words
49412580107885265514591282862261301333333977558764250322225120

Assume that only 1 out of 100 Billion make sense and that's still a lot of 
choices to be made and one has not even gotten past the opening sentence of a 
novel yet.

 programming is orders of magnitude less.

> 
> 
> <stuff we agree on ... snipped>
> 
> .002
> 
> PS when do I get MY BMW.... (motorcycle preferably) send donations to...
>