But you only risk getting your employer into trouble because they're committed to releasing proprietary software. The FSF considers proprietary software evil; the GPL inconveniencing developers of such software is a central goal, not collateral damage.
Now, if the concern is that you can't modify a GPLed text editor or compiler or browser or other such tool at your workplace, that shouldn't be an issue. Unless you're releasing modified binaries from GPL code, you are under no obligation to distribute your changes to the source.
> Now, if the concern is that you can't modify a GPLed text editor [...]
That is indeed the issue. The concept is that I might read some algorithm or solution as a whole while working with the GPL code. This solution, even if not directly copied, might end up in the code of the proprietary (and patented) product.
If prosecuted in America, where the court is interpretive on new issues, this could result in trouble for my employer. Having no previous rulings to go by, there is no way to approximate the outcome of a case such as this and hence looking at GPL code could constitute a risk.
A good "inverted" example of this is when Oracle went after Google/Android for supposedly having copied a hashing function. The obvious question and defense is: common data structures have obvious hashing functions and there is a good likelihood of coming with identical or near-identical code, i.e. serendipity.[1] Even though there is no proof that the code was copied, the court ruled in favor of Oracle on the grounds of "structure, sequence and organization of an API is copyrightable."
How this applies to GPL code, from my situation, is uncertain. GPL has patent clauses, so it's really risky.
In that case you can't possibly read any code whatsoever.
Yet again the GPL does not restrict your freedoms. It just adds to them if you're in the pro-user freedom camp of software development and leaves you no worse off if you don't.
> In that case you can't possibly read any code whatsoever.
Hyperbole isn't really called for. I can read MIT/BSD because that has patent allowances.
I'm not saying that GPL is bad, I am saying that GPL serves different freedoms to MIT/BSD which in turn serves different freedoms to proprietary (which does have freedoms, e.g. liability).
There is no one size fits all solution to licensing and there will never be.
Without reading the code, how do you know you're not inadvertently reimplementing something the same way a GPL program does?
If you are truly concerned about this, the best way to avoid copyright complaints is to be aware of what's out there already and deliberately build a different implementation. I would argue that it is more beneficial -- from your point of view -- to read as much GPL code as possible.
Let's stipulate that this hashing scenario (I don't know much about it) is an example of a common algorithm that is the best way to solve a specific problem... it probably is. The only way in the US to defend that premise legally is to go to court. If the court decides against you, you have to lobby the legislature. Is the process convoluted and costly to the point of being broken? Probably, but this is how the the law comes to grips with changes in society, and there's really no alternative if you want to live and work in the light of day.
> This solution, even if not directly copied, might end up in the code of the proprietary (and patented) product.
I and many others are deeply opposed to proprietary and patented software. I (of course) understand you need resources to build software, but I think there are other ways to acquire those resources. The vast majority of software that people use is open source, free software (Firefox, Chrome, Nginx, Apache, Java, PHP, .NET, Linux, Android, MySQL), and this has been amazingly great for society (admittedly mostly rich, western society though). Can you imagine the cost of a browser if everyone had to license it and all the technologies contained therein? PNG, JPEG, GIF, SVG, SSL, JSON, XML, 2D rendering, font rendering and layout, hardware accelerated rendering, JavaScript, JavaScript interpretation & JIT, DOM processing, CSS application, user interface toolkits, video codecs, audio codecs, network protocol licenses, memory management algorithms, blah blah blah. By the time you were done, Firefox would cost $60 a copy - if you were lucky.
> Without reading the code, how do you know you're not inadvertently reimplementing something the same way a GPL program does?
Fair point. At the end of the day it's complex so I go for the simplest solution: avoid GPL code in any way possible.
> I and many others are deeply opposed to proprietary and patented software.
I'm starting to understand that from the responses, now. Although I work in proprietary, I try to look at software licensing from a more neutral standpoint - which might not be popular around here.
Please cite a single court case in which a engineers reading of source code has been used as proof of intentional infringement of a patent or copyright.
I hope you do not read newspapers, HN or other places which publish information about new technology. Newspapers seems much more dangerous than source code, as they often publish news about patented technology. Think of the risk that you inflict your company, as a single miss clicked link can cause ruin in a future patent case. Better to cut down on any contact with the outside world.
> Please cite a single court case in which a engineers reading of source code has been used as proof of intentional infringement of a patent or copyright.
It might help if you read the entire comment, instead of patronizing it after reading a single sentence. I did actually address that.
> Better to cut down on any contact with the outside world.
>> there is no way to approximate the outcome of a case such as this and hence looking at GPL code could constitute a risk.
contact with the outside world also constitute a risk, and there is no way to approximate the risk. Beware!
The Oracle vs Google case is not about google employees reading source code. They copied header files, which neither party denies. The case resolve around the question: Does copying header files constitute copyright infringement. Talking about reading source code in this context is FUD, likely from someone who has an agenda.
Now, if the concern is that you can't modify a GPLed text editor or compiler or browser or other such tool at your workplace, that shouldn't be an issue. Unless you're releasing modified binaries from GPL code, you are under no obligation to distribute your changes to the source.