After talking with some folks about the best way to present this, it will end up being a basic overview of how lawsuits, copyrights and patents actually work in the US legal system.  My purpose here is not to espouse views on these things, but instead transfer knowledge of how it all works, in the hopes of reducing ignorance of the “way things are now” among engineers, and in particular, OSS folks.

• For the legal system in general
  • How our legal system is broken up at the federal level
  • Anatomy of a lawsuit
  • Costs of a lawsuit
  • General rules of interpretation of statutes and laws (possibly, depends on time)
• For Copyright
  • What is copyrightable
  • What the actual copy rights involved are
  • Fair use
  • How infringements of these rights are determined by courts (IE tests used, etc)
• For Patents
  • What is patentable
  • What the actual rights granted by patents are
  • How patents are examined
  • How prior art is used
  • How courts determine validity of patents
  • How courts determine infringement of patents

I only have a limited amount of time to go into this stuff, so I plan on focusing particularly on software.  Anyone who knows copyright knows that most people don’t realize how complicated the rights around things like performances and sound recordings are. Rather than dumb things down, I’ve tried to explain enough of the underlying principles and situations that one could at least come up with reasonable answers given a more complicated situation.

If you have ever taken a course on IP law, copyrights, or patents, you will likely be very bored at this presentation

Though I may also get into contracts, licensing, and EULA’s.

Of course, I’m happy to answer questions about my views, or about pretty much anything legal while I’m around.  So if you plan on being there, please bring all the questions you’ve got.  I’ve got all the time in the world to answer them.

In the current system, if two people file for a patent, something known as an “interference” is declare.  From there, both sides spend a lot of money in a bunch of complex court-like proceedings trying to prove who invented the thing being patented.  Whoever wins the interference proceeding gets the patent.  There are large battles over interferences, and there are law firms that specialize in handling them, because they can get quite complex.

In the proposed first-to-file system, if two people file for a patent, the one who filed first gets it.

That’s it.

Moving to a first-to-file system would have absolutely no effect on what is considered prior art.  If you go and publish something, it does not mean that someone can go and file a patent on it if you do not.  Your publication will still be prior art against their patent.

The real reason for first-to-file is to get rid of the weird system of interference proceedings we have, not to try to make more things patentable.

CVS2SVN has a habit of misidentifying entire-branch copies.   As a result, you will usually end up with branches that are as big as the original branch (on-disk size wise), rather than being , say, 1000 bytes.

This is not to say that SVN repositories are great at space usage.   But as a rule of thumb, proper SVN repositories are always smaller than the equivalent CVS repository.

SVN has some problems in terms of repository storage formats.  We know this.  This is why SVN will probably going to move to something like revlogs for 2.0.  But a lot of the specific size problems i’ve seen complained about on the web or mailing lists, are usually just bad conversions.

As usual, the Federal Circuit made decisions that seem to be based on what they believe the law should say, policy wise, instead of what it actually says. They generally make decisions that they can provide a sound basis in policy for, even if it means seriously stretching the meaning of the underlying laws. I usually disagree with these decisions as a matter of policy, but in any case, stretching the law to fit the often idealistic view they have of how patents support innovation is the wrong thing to do. This is one reason that they are so often the subject of harsh criticism. There was *never* anything in the patent act to support their test of obviousness. It was simply made out of thin air and their view of what risks arise when considering obviousness. These risks are not codified as factors to consider when judging obviousness, and as a result, any test rigidly based on them (as the Teaching-Suggestion-Motivation test was) was bound to be smacked down.
For KSR v. Teleflex, the court eviscerated the TSM test for this very reason.

In particular, the court said:

• a combination of familiar elements according to known methods is likely to be obvious when it does no more than yield predictable results.

  The Federal Circuit previously held there is no presumption that it would be more likely to be obvious.

• Under the correct analysis, any need or problem known in the field and addressed by the patent can provide a reason for combining the elements in the manner claimed.

  The Federal Circuit had previously held the that only the problem the particular patent was trying to solve should be looked at. I.E. A person would only look only at those prior art designed to solve the same problem.

• It is common sense that familiar items may have obvious uses beyond their primary purposes, and a person of ordinary skill often will be able to fit the teachings of multiple patents together like pieces of a puzzle.

  The Federal Circuit had previously required explicit teachings of this, rather than allowing common sense (at least, most panels. There is a one-off decision here that said otherwise, but has not been followed).

• a person of ordinary skill in the art has good reason to pursue the known options within his or her technical grasp. If this leads to the anticipated success, it is likely the product not of innovation but of ordinary skill and common sense.

  This is more or less the court’s way of saying that people are creative, and ordinary creativity that combines prior art should not be mistaken for true innovation. Patents are intended to protect novel things, not things ordinary creativity and skill would come up with. This is a rebuke of the Federal’s Circuit constant claim that patents can not be made obvious by showing they are “obvious to try”.

• The court noted the one off Federal Circuit decisions that appeared to try to modify the TSM test to avoid being overruled by the Supreme Court, saying
  

  Those decisions, of course, are not now before us and do not correct the errors of law made by the Court of Appeals in this case. The extent to which they may describe an analysis more consistent with our earlier precedents and our decision here is a matter for the Court of Appeals to consider in its future cases. What we hold is that the fundamental misunderstandings identified above led the Court of Appeals in this case to apply a test inconsistent with our patent law decisions.

• Just to show the court gets it, they said:
  

  And as progress beginning from higher levels of achievement is expected in the normal course, the results of ordinary innovation are not the subject of exclusive rights under the patent laws. Were it otherwise patents might stifle, rather than promote, the
  progress of useful arts. See U. S. Const., Art. I, ß8, cl. 8.

  Some of the Federal Circuit judges may take this to heart, but a few seem to believe the correct policy is to let people patent anything, and let licensing fees sort out the winners and losers.

This is true even though their was, at least as of two months ago, a Long Fence around the Capitol reflecting pool in order to protect you from the violent terrorist ducks that swim there with their cute future jihad-warring ducklings. Personally, this would scare me away, but we are a nation of brave patriots, unafraid of our mallardly enemies.

In any case, as most tourists are unfamiliar with the Metrorail system, I thought I would provide a guide to understanding the announcements you may hear from time to time, either on the trains, or in the stations. You see, Metro train and station operations speak a different language than the rest of us, even though it sounds very similar, so it is essential you understand the true meaning of what is being said.

In the stations:

Attention customers traveling on the Red Line in the direction of Shady Grove. We are currently experiencing minor delays due to a mechnical malfunction at New York Avenue. A train will be at your station momentarily.

This translates to the following:

One of our trains broke and they are very heavy, so it takes us a while to move them off the tracks. I hope you enjoy waiting 35 minutes for another train, because that’s how long it’s really going to be before one gets to your station. After all, why would we have thought ahead and planned a way to move trains off the main tracks without it taking 20 minutes. I would suggest you really just go to sleep and stop looking down the tunnel for lights. I’ll make another blaring loud announcement when it is time to wake up, or more likely, you’ll hear the train, being that they are very large and very loud.

On the trains:

There is a train directly behind us

This translates to “There is a train within the next 15 minutes”

Doors closing

This translates to “Doors are closing, don’t try to get on the train”

Please stand clear of the doors, thank you

This translates to “Get the fuck out of the doorway you moron, this is not rocket science”

Please do not block the doorways, there is a train directly behind us.

This translates to “To the dumb asshole who didn’t listen to the first two messages and decided to try to jump through closing doors onto a completely packed train where they wouldn’t fit anyway, even though the next train is less than 1 minute away: If you do not remove your handbag from the door so it can close, I am going to be forced to put the brake on, come over there, and beat you to an inch of your life with the aforementioned handbag. Also, as an aside we appreciate that you decided it was really important to hold up a train with 350 people on it trying to get home because you are a super important person who has to be on this train, instead of the completely empty one waiting for us to leave the station. I hope you die in a fire”

I hope this guide is found useful for all of our visiting patriots.

To give some idea of numbers, there are probably 5-10 peer reviewed journal/conference published papers on points-to analysis every year. I read all of them. I also read all the research that Google, Citeseer, and other methods can find.

Most of it, sadly is not useful, for a variety of reasons. Those that concentrate on speeding up points-to analysis rarely contribute things that make a real difference. There are a variety of reasons for this:

1. Taking shortcuts - A lot of papers deal with speedups that just aren’t sound (leaving out effects of external function calls, assuming random things about variables, visibility, whatever, instead of just handling the full language) and they leave it as an exercise to the reader to make what they’ve done sound. In every case I’ve implemented one of these unsound algorithms, making it sound lost almost all speedup the algorithm purported to give.
2. Points-to analysis research that focuses on Java - Java is rather easy as a language to write pointer analysis for. You can’t take the address of fields, you know which functions are really doing allocation of objects, etc. I’ve seen plenty of techniques that work great for huge java programs (which I guess is useful to static java compilers, but JIT’s just aren’t really going to play in this space until memory and time requirements are ridiculously low. They aren’t there yet). These techniques are essentially useless when applied to C, C++, or other common languages, simply because they use language specific assumptions that no longer apply.
   IMHO, It would be much better to do work on pointer analysis, that can handle a language like C, and then apply it to Java by making it simpler, rather than coming up with algorithms that only work on Java, and can’t be made to work on something with more complications without losing all the benefit the work done was supposed to bring.
3. Comparison to ridiculousness - This is more a combined criticism. It turns out a lot of the people write their own implementations of the algorithms they compare against, and it’s obvious they only put a minimal amount of engineering work into them. You then see comparisons against really dumb implementations of other algorithms, or using really bad data structures, and claims of speedups, when if you look at good implementations of the other algorithm, you see the time is usually the same as the new algorithm claims. Part of this is another problem - a lot of papers lack source code that implements their work (to the credit of many researchers, they are almost always happy to give source, but it would be nice to just make it available with the paper, and put in a URL in the paper to the source).  It’s interesting to note that for the really good algorithms I have found, when I looked at the source, they often had really good implementations of the algorithms they compared against.
4. Misguided notions of what is practical - I think this more or less speaks for itself, but I have read pointer analysis papers that claim “practical” algorithms, where their idea of practical is to take > 1 hour or > 4 gig of memory to analyze 100k of code. You can always claim users are willing to wait, or we shouldn’t run expensive algorithms everywhere. The problem with these claims are:

• That only about 0.1% of users are willing to wait.
• Making algorithms that only run sometimes are a maintenance burden and a can of worms for bug reproduction = unlikely to be implemented without incredibly significant benefit.
• There are algorithms without these requirements that have just as good precision.

This is not to say all the research I read is not applicable, and not a dig at, for example, people doing Java pointer analysis. I have seen good work on Java pointer analysis, where authors explain how to make it applicable to other languages, or at least have some cognizance that it is not useful outside the realm of Java. I am simply pointing out that if you want your work to be used in production compilers, it needs to be implementable without having to spend another 3 years trying to get back the speed you claim when you transform the algorithm to be sound for C/C++/etc.
Just to give examples of good pointer analysis research i’ve seen in the past few years, the work done by Paul H J Kelly and David J Pearce on field-sensitive pointer analysis for C now forms the basis for the algorithm used by GCC. In addition, Calvin Lin and Ben Hardekopf have done an amazing job of attacking the real problems of memory usage and time of points-to algorithms, using implementations they plan on making freely available. Most of their work concentrates not on better solving algorithms, but on very cheap offline detection of much larger classes of pointer and location equivalences. There are plenty of other people doing good work here (particularly in the area of scaling context sensitive analysis), I am just pointing out the ones I remember off the top of my head.

I weep for the children.

Crap.

I just ruined the ending.

A few very nice, very professional furnace installation people showed up on Tuesday, removed our old furnace (while wondering aloud how we weren’t dead from various things wrong with it and the installation of it), and installed the new one.

The new one has a 20 year warranty on all parts and the heat exchanger, meaning we have at least 20.01 years before the next service call.  Oh, and it’s not spewing odorless deadly gases into my house.  That’s always a plus.

The rabbits seem happy to be back downstairs in their normal habitat, seemingly unaware of how close to death they were.  Hell hath no fury like a woman whose house is 50 degrees or less.

As for me, well, I’m less sick.  I can actually eat food now, which is good.I still haven’t been in the office in a week (ice, ice, ice, sick, sick), but I’ll get back there this week.

Thus, everything is more or less back to normal.

By died, I don’t so much mean it no longer functions, as much as I mean that they discovered a 3 1/2″ crack in the heat exchanger, which causes a very slight hazard of carbon monoxide poisoning to you and your entire family.

So it’s not so much dead, as it’s disassembled and if I turn it back on, I may end up dead.

Except for that, it works great! Oh yes, the original problem was the blower fan wasn’t working anymore. So maybe it wasn’t working great anyway.

As everyone knows, furnaces only die ~2 days after major winter storms, causing two phenomena to occur simultaneously

1. Nobody has any space heaters left
2. Nobody will be able replace your furnace for at least 5 days.

In our case, we found 3 space heaters hidden under random baskets of easter candy at Target, after searching Lowes, Home Depot, K-Mart, and 7 other places. Our house is now a balmy 56 degrees.

Luckily, our Home Warranty Company was happy to pay the ungodly sum of money it costs to upgrade a 22 year old furnace and pieces, and was even willing to pay someone to do it this weekend, so that we didn’t freeze to death. Heck, they even had the parts and furnace ready for pickup by someone within 8 hours of finding out ours was dead.
However, this was not to be, as no HVAC service company in the area actually had any appointments until Tuesday morning. My guess is that they charge people even if they are dead by the time they get there.

To make for a more enjoyable weekend, I also decided to get the stomach flu, which means I have not actually eaten anything in 3 days. Okay, so, i’ve eaten it, but then it’s ended up outside my stomach through no intention of my own. Combined with the freezing cold, halleucinations, and fever, this is great. I have fever dreams about having a furnace.

In order to avoid a small condition known as “dead rabbits”, we had moved them from the freezing basement to the upstairs. They decided this was not fun enough, so they escaped while we were out looking for space heaters. This has caused great displeasure in my female unit to occur, and she is currently upstairs looking for recipes that involve cooking rabbit.

Oh well. I’m sure everything will go to plan Tuesday.