Francesco Paolo Lovergine dijo [Sat, May 01, 2004 at 11:40:08AM +0200]: > > To adapt an analogy that someone used earlier, when you go to a store, > > you might find fonts, images, or other data in a box in the software > > section. However, you are not likely to find a specification for > > TCP/IP in the software section, and you are not likely to find a print > > of Starry Night in the software section either. > > > > Ah that's an interesting point. TCP/IP is a standard, so it's non free... > Maybe all implementation of that should go in contrib so, because > they 'depend' on a piece of 'something' which is not free. So, we > have to move the whole kernel there, and oh sure, libc too... > > Probably someone should clarify me better what's 'depending' means, > why a document which define a standard is non-free but > a program based on that standard is not in contrib? Who wrote that > program did read the standard and use it to write the program. > So, the program indeed _depends_ on that standard. And the standard IS > the document which describe it. So there's a direct dependency. > And POSIX? Mmmm...
Exactly, it seems you understood this perfectly. The standard documents defining TCP/IP and POSIX are non-free. Take a look at them [1], you will find that many of them say that 'distribution of this memo is unlimited' or similar statements. Of course, arbitrary modification is _not_ allowed (need I explain why? :-) After all, the documents are a published standard, not something to play on). They are _not_ DFSG-free. Now, about implementing them... You can implement them freely, and put your implementations under a DFSG-free license. Greetings, [1] http://www.faqs.org/rfcs/ -- Gunnar Wolf - [EMAIL PROTECTED] - (+52-55)5630-9700 ext. 1366 PGP key 1024D/8BB527AF 2001-10-23 Fingerprint: 0C79 D2D1 2C4E 9CE4 5973 F800 D80E F35A 8BB5 27AF
