5 ms·
Interesting rant on expert witnesses. Mostly interesting because I've done some expert witness testimony and found it fascinating. I completely resonate with J
by ChuckMcM 10y ago
Interesting rant on expert witnesses. Mostly interesting because I've done some expert witness testimony and found it fascinating.
I completely resonate with John's issue with the expert's testimony and his understanding of it. John mentioned that their own expert testified differently but he didn't say whether or not their lawyers attempted to impeach their expert with cross examination. In the two cases I participated in, both sides agreed ahead of time on how to "handle" experts (in terms of cross examination). In one it was documents only (no testimony) and in one I was deposed by the opposing counsel after documents in which the lawyer clearly had some notes (presumably from their expert(s)) and was trying to get me to recant or change some of my points (I didn't need to).
In a jury trial I can imagine that having the jury understand your testimony is probably the most challenging.
And, like John, I consider the notion of 'non-literal copying' to be pretty ridiculous. In the limit it means "you read this code, understood how it worked, and wrote new code that could do the same function." You can stretch that to cover anything you have ever seen. Which is sad.
- madgar 10y ago> In the limit it means "you read this code, understood how it worked, and wrote new code that could do the same function." You can stretch that to cover anything you have ever seen. Which is sad. I thought the industry adopted "clean room" techniques decades ago to avoid this kind of appearance of impropriety. It's more expensive, sure, but isn't eliminating legal risk something that typically gets paid for?
- johnbellone 10y agoThis particular point worries me. Are we now not going to be able to make a clean room re-implementation of concept? Do we need to worry about our previous employer suing us because we are successful?
- madgar 10y agoThey didn't seem to do clean room here. Seems like Carmack was wrote code for both companies which is the exact opposite of clean room. (Please correct me if I misunderstood the facts of the case)
- logfromblammo 10y agoThe employer of John Carmack is buying his work product for as long as he remains an employee. The company cannot buy his knowledge and expertise from him, as that is an integral and indivisible part of him. So if you hire Carmack to write a rendering engine, then he quits and goes to work for another company, to write a different rendering engine, he is not "copying" the engine you bought. His memories of all the software he had previously written do not belong to you. You cannot contractually obligate him to forget a part of himself that makes him valuable as a software professional. The work of a software professional is constrained, because while a given mathematical problem may be solved in an infinite number of ways, only a few of those ways will be optimal for any given set of computing conditions. Different implementations may be grouped by algorithm. With respect to computer graphics, all rendering engines will tend to converge on the fastest algorithms for painting pixels on rectangles. Because of this, all rendering engines will be very similar in some respects. But they will also markedly diverge in others. There are pretty much only 4 types of algorithm: the fast way, the frugal way, the magical way (viz. both fast and frugal, but requires top-level expertise to invent or understand), and the hardware exploit. Carmack is certainly capable of writing magical algorithms. And he can remember how to make the pixie dust. He will also tend to write all his implementations in a similar style. Thus it is very possible that one person could write identical code at different times, while having no specific memory of the previous time it was written, nor any need to refer to associated documentation around that instance. The clean room is clean if no property of the other company was ever inside it. Carmack cannot be property, therefore his presence is not a contaminant. My opinion is that any non-disclosure agreement Carmack may have been a party to can only meaningfully cover the contribution of other individuals to the project, since Carmack-qua-Company-A-employee cannot meaningfully wall off knowledge from Carmack-qua-Company-B-employee. If Company A wanted to keep Carmack's raw knowledge and expertise from Company B, they would have had to get him to sign a non-compete agreement (which no one at his level of clout should ever do for less than a significant ownership share in Company A).
- ChuckMcM 10y agoIn this case, and others like it, a clean room is challenging because a principal (in this case Carmack) left one company and joined another. So the 'taint' if you will is in his head.
- madgar 10y agoIf so, then I see two possibilities here: 1) The IP in question was only reimplementable with Carmack involved, and the lawsuit was basically inevitable. Hiring Carmack in this scenario is reasonably perceived as an end-run around buying/licensing the IP itself. 2) The IP could be reproduced clean room without Carmack involved. Lawsuit is easily headed off.
- fulafel 10y ago"Clean room" is a name for one extreme of the spectrum, where you go for overkill in demonstrating and documenting there's no possible way there could have been copying. The law says you don't need to go so far, but people try to mitigate the low burden of proof in US court civil cases.
- caf 10y agoThe visual and written arts seem to get away without requiring "clean room" techniques. It's not like authors have to avoid reading any other author's work.
- jly 10y agoCleary with a lot of money on the line from one side or the other, either side can likely find a 'expert witness' who is credentialed and willing to say what they want. At least that's true in a lot of cases. Regardless of how this handled in court, it's an issue. How is a jury built of non-experts to handle such a situation? In this case they clearly believed one side's testimony over the other, against what one could reasonably argue was sound judgement on the issue. Expert witness testimony is more about theater, performance, and credibility than it is on the actual technical matter. Until we put in place a system where an expert witness is more accountable to the court than a client, we will continue to have this problem.
- yaur 10y ago> In a jury trial I can imagine that having the jury understand your testimony is probably the most challenging. What I've seen from the sidelines is that what you want in an expert is someone who make your point, which has more to do with assertiveness than competence, and the credentials to back it up but who can't delve into the topic too deeply because information that the jury doesn't understand is an attack surface that defense attorneys can use to create doubt. I imagine the rules are a little different in civil litigation where both sides are likely to have experts.
- rhino369 10y agoThat is pretty much how it works on the civil side too. Pick an older professor from the best local college who has given testimony several times before. The open secret is that a lawyer writes all the expert reports and the expert just reviews and makes edits.
- ghaff 10y agoThat has certainly not been my experience. The lawyers have a clear objective for a given expert report and are heavily involved in a lengthy back-and-forth review process. Perhaps sometimes lawyers do the primary drafting but that wasn't what I encountered.
- gregw2 10y ago^this makes a lot of sense
- Digory 10y agoIn civil litigation, my professor used to say you needed three experts: one who actually knows what they're talking about, one that the jury will like, and one to translate between the two.
- efraim 10y ago>And, like John, I consider the notion of 'non-literal copying' to be pretty ridiculous. What about translating from one language to another, like in Carmack's book example? You can't literally copy from c++ to Java, but it can look pretty similar. Would that be considered copyright infringement?
- dragonwriter 10y agoI'm actually currently involved in non-literal copying of a very old C code base into Python (legally, independent of whether that's protected or not, as it's GPLv2 and the result will be as well), and the code doeant much look the same (and wouldn't even if I was copying it into C in a more modern style.) There's certainly a point where you are just duplicating unprotected function rather than making a copy dependent on the expression of that function in the original, and that's a tricky legal line with software, but non-literal copying is certainly a thing.
- DannyBee 10y agoTo the last point, courts do actually try to solve this. The goal of the https://en.wikipedia.org/wiki/Abstraction-Filtration-Comparison_test https://en.wikipedia.org/wiki/Abstraction-Filtration-Compari... that gets used is, in fact, to abstract things, eliminate the non-copyrightable parts, and try to see if any important stuff is left. Wikipedia actually has a fairly good explanation of it in practice. So they are actually trying, and the real problem is that existing copyright law, for literary works, is just not a match for how software tends to actually be developed, which is that "most code is worthless, there is very little that really should be protected". (and before people argue that it could be millions of lines, the US, thankfully, refuses to have a sweat of the brow doctrine)
- AnthonyMouse 10y ago> So they are actually trying, and the real problem is that existing copyright law, for literary works, is just not a match for how software tends to actually be developed, which is that "most code is worthless, there is very little that really should be protected". It's not even that. It's that the debugging is the hard part. The general idea of what the code should do is easy. Getting all the edge cases right is hard. The result is that a) any two working implementations are going to be functionally equivalent (by tautology, because that's what "working" means), and b) "non-literal copying" doesn't actually help the "infringer" because anything short of literal copying or literal translation (meaning the thing a compiler does), is going to require you to redo the hard part anyway.
- amag 10y ago> "most code is worthless, there is very little that really should be protected" I think a better analogy would be that most parts of software (i.e. functions) can be thought of as words in literary works, it's not how you spell them but rather how you string them together that makes the work unique.
- jandrese 10y agoThe non-literal copying thing seems insane to me. It's like solving a math equation, then getting a new job and being told to solve the same equation a completely different way. Of course the solutions are going to look similar, it's a re-implementation of the same core principles. How far could you go in implementing a VR game if you aren't allowed to use matrix transformations or dot products?