Somebody, please save RMS, write an Emacs mode that inserts spaces into JavaScript and converts single char names into made up longer ones! For extra bonus, insert comments that talk about how if you are not using GNU, you are a slave! A slave, I tell ya!
Just so you know, if you plant an apple tree from an apple you buy at the grocery store you will likely be sued for patent infringement. Though, it's probably sterile anyway.
But your planting of a tree doesn't automatically prevent the farmer you bought the apple from from making any money.
If I have to release my code under GPL, I can't make money charging thousands for it (low volume type of app). If I can't make money from it, I won't spend years developing it. Not having the app would make it harder for some pharma scientist somewhere to cure cancer. Everyone loses. Communism doesn't work.
You are not forced to release your code under the GPL. You can spend years developing it and release it however you want.
However if you choose to release under the GPL you will be standing on the shoulders of giants. You will have whole new avenues of development opportunity opened up to you as a result.
"Communism" does work in the sense of the Linux ecosystem that you see around you every day. Not only that, but companies like Microsoft are forced to innovate because open source eventually catched up to them. Without free software all software would be tied up in a morass of patents and licensing fees. Progress would have ground to a halt, and we would wait with baited breath to pay $500 for the latest version of Windows crafted upon their latest business model to coerce upgrades from all their regular customers.
If you are adding 1000 man hours on top of 50,000 man hours of development that you are getting for free, is it moral for you to give nothing back? The GPL is simply a way for the contributors of the 50,000 hours to say that you can use their code if you share in return. It's not clear at all why you think you should be entitled to that code.
you're legally allowed to use it to grow yourself an apple tree.
Legally, sure (modulo plant patents). Practically, you can only do it as long as you don't mind it being a crab apple tree. (Apples tend not to breed true; if you want a tree that is sure to produce edible apples you have to get it by grafting.)
That's what they want you to believe. Try reading the license.
You have to know how many lines from the header file you are "distributing" (in the GNU lingo). If you are using C++ templates, you are probably screwed. You have to make sure you link dynamically (or else). LGPL seems to require advertising the library. And if you want do static linking, talk to a lawyer. Linus and RMS disagree on what constitutes "derived work", btw.
Edit: I'd like to know what motivated people to upmod parent, when it's factually incorrect and can cost developers dearly if they take the message at face value.
You are talking about Section 3 of the LGPL? IANAL, but this section seems pretty logically simple to me, and I think you are misreading it. If you read the terms, you are required to do either of the following:
* a) Give prominent notice with each copy of the object code that the Library is
used in it and that the Library and its use are covered by this License.
* b) Accompany the object code with a copy of the GNU GPL and this license document.
How is needing to provide a mention of the LGPL'ed libraries and the fact that they are LGPL'ed in any way screwing you? You also have to provide a copy of the GPL and LGPL amended your own legalese. You are not required to distribute the source for the library, nor the source of your own application. I've seen several large commercial applications make mention of included libraries, some under the LGPL, and it has never bothered me, nor anyone else who purchased it.
I understand that you want to be able to charge money for the software you write, but that you seem to perceive the LGPL as a threat to this ability seems excessively fear-mongery. I agree that anyone concerned about licensing terms should consult an attorney, but because you want to understand things better, not because you are fearful of software development turning socialist.
If, as you say, LGPL, is clear to you, what do you think "object code" refers to in the (b) clause you just quoted? How are the "ten lines" counted in the paragraph above (you didn't quote): separately or total? etc. etc.
2. I'd assume as lines are normally counted. The number of line-feeds and/or carriage returns in a file. I don't see what the big deal is anyway; if you are at all concerned, just include the license text. Is doing so that big a deal?
No. Whose object code? Is it talking about my object code or the library's?
Ten lines per function, per header, per library, per program?
Edit: I think you didn't understand LGPL yourself either. I think in the quoted lines, "object code" refers to your code, and accompanying them with a copy of GPL refers to having to release your code under GPL. Or do you think you only have to say you release it under GPL and not actually release it?
The object code is the aggregate of your code and the library's header code.
If you have just your object code, you are not bound by the license of the library, then you don't need to include it.
If you have just the library header code, then you are not incorporating it into anything else, and have no obligations by the license.
If you have both the library header code incorporated into your object code , then you are bound by the license to provide attribution along with that incorporated object code.
As for you second question, I'd guess lines per file. But it is a bit ambiguous, as almost all legalese is. Again, you could just err toward caution, and include the license text regardless. Are you averse to this for some reason?
Oh, and as per the wiki link, object code is the compiled binary of your source code. At this point, whatever headers you included would already be integrated.
If that were true, why would would the FSF say that
"using the Lesser GPL permits use of the library in proprietary programs; using the ordinary GPL for a library makes it available only for free programs" [http://www.gnu.org/licenses/why-not-lgpl.html]? The LGPL is not designed to trick you into using the moral equivalent of the GPL; it's designed to occupy the middle ground between GPL and BSD licenses.
> Lesser GPL permits use of the library in proprietary programs
That means "in some cases", not "in all cases". Logical fallacy on your part.
> ordinary GPL for a library makes it available only for free programs
Here, GPL'ers contradict themselves (for both senses of "free"). GPL code can, technically, be non-gratis. And GPL code can be used in non-GPL programs (you just can't distribute them easily).
I think you're entirely wrong about the "copy of the GPL" business. The LGPL consists of the GPL plus a bunch of extra stuff. The LGPL document found e.g. at http://www.gnu.org/licenses/lgpl.html lists only the extra stuff. Therefore, whenever it imposes any requirement to distribute a copy of the licence, it says "a copy of the GNU GPL and this license document". That's all. If there were a requirement to distribute your stuff under the terms of the (not-L)GPL, then it would say so. (Merely distributing a copy of the GPL along with your code would not put the code under that licence.)
Those terms to which you object mean, I think, the following: If your object code includes bits of the library header files, and if those bits are substantial, then you have to (1) say that you're using the library, (2) say what licence the library is covered by, and (3) include a copy of the documents that define that licence.
Exactly. I'm unwilling to pay the uncertainty tax. (Ironically, commercial software is often much easier to justify than the LGPL equivalent is! "Fork over $200 and you're OK" -- sweet, done! Spend a few hours reading legal code and still coming no closer to understanding whether a plugin in application A accessed over REST by application B is "linked" to A, B, A and B, or neither... bah.)
Yeah, and I'm not sure I follow this (the Obfuscript bit). You _do_ have the source code, albeit a bit difficult to read. I'm guessing there are plenty of GPL'ed programs out there with horrible style and horrible documentation.
That's because the LGPL is only meant for very specific situations were licensing under the GPL would hurt the free software movement. This is hardly the ever the case (a notable exception would be a high quality library) and thus the encouragement for the GPL.
I'm obviously unfamiliar with US job dynamics, but a YC applicant team, if funded, would not arrive in the US "looking for a job"... they would have it (YC helps with paperwors for incorporating after all, doesn't it?)
I'm a EU resident as well, and I'm tyring to apply for this round.
There may be a circular dependency here. AFAIK, to found the company you would already have to have a legitimate visa but to get a visa you would need to already be employed by the company.
IIRC Miguel de Icaza originally came into the US on a tourist visa and thus couldn't be an official founder of Helix Code/Ximian since he didn't have a work permit.
We'll overshadow that eventually :) That's an unfortunate side-effect of starting out now; Quantios was in part chosen by virtue of domain-name availability, so "Did you mean..." is a predictable side-effect of that while we get off the ground.
In the States, PBS does not run commercials. Yet the scarcity of people who watch public broadcast may be evidence against that claim. (Maybe average people prefer commercials?)
> There are plenty of examples of successful companies that have used this model
Do you mean they used their blog as the main marketing effort? I'd like to hear more about those.
It seems if the users are looking at your blog, they've found you already. Or perhaps those companies created thousands of other blogs that were only used for search engine optimization (linking back to their money blog)? That I can understand.
You're assuming that a purchase is done the instant a company/person is found. What a blog does is build trust, familiarity and credibility in a particular field. So there's two scenarios : one is a reader becomes familiar with you and purchases a new offering (creating a channel and selling into it), the other is a reader finds your product, skips through some of your blog posts and decides to purchase based on the fact you appear to know what you are talking about. (creating credibility)
Of course this only works if your blog posts are related to your product offering.
We'd be using OS X and FreeBSD instead of Linux. Who's laughing now?