Talk:Copyleft
from talk:Free software
According to wikipedia 'copyleft' was coined by Don Hopkins and not Richard Stallman. The Free software and the Copyleft articles now disagree with eachother. --arcade
- This could be tough to sort out -
- The Origin and Practice of "Copyleft" - http://www.olypen.com/harmon/fdl/copyl.htm -
- "Transcopyright: Pre-Permission for Virtual Republishing" by Theodor Holm Nelson - http://www.sfc.keio.ac.jp/~ted/transcopyright/transcopy.html - (says "the terms "shareware" and "copyleft," declared by Bob Wallace and Richard Stallman respectively, have come to represent their respective permission doctrines, both [are] now widely accepted and used.")
Both of those attributions are wrong. "Shareware" predates Wallace as well--I know that I personally saw the term applied to, for example, things like ARC (the original pre-Phil Katz version) long before Wallace's PCWrite. Both should, however, be credited with refining and popularizing the terms. --LDC
- and none of this has anything to do with viral licenses, which apply as much to patents or to music as to software.
It was coined by Don Hopkins without question, then popularized by Stallman. Stallman says so himself on his website. --AxelBoldt
- correct, Don Hopkins wrote it on an envelope - "copyleft: no rights reserved". Don validates the story - I heard it from him firsthand.
I redirected viral license to this article. There were a couple of things wrong with that article: First, it mentioned trademark and patent law, but viral licenses don't exist in that realm. Second, it was blatantly non-NPOV, in that it called viral license "desirable and popular" and used the phrase "degrade to an open source project", making the additional mistake of assuming that open source projects cannot be copylefted. AxelBoldt, Tuesday, April 2, 2002
- "it mentioned trademark and patent law, but viral licenses don't exist in that realm." This is false. You have clearly not researched the subject. MIT and the Government of China both employ viral patent licenses requiring reliable favorable terms for re-licensing improvements to works they license.
- " it was blatantly non-NPOV, in that it called viral license "desirable and popular" and used the phrase "degrade to an open source project", making the additional mistake of assuming that open source projects cannot be copylefted." This is your ideological problem. Viral licenses such as the MIT and Chinese and GNU license are all quite popular, and must be considered desirable by *someone* or they wouldn't be. If you want it clarified to whom it is desirable or popular, fine, but destroying the article over that issue amounts to vandalism.
- "open source projects cannot be copylefted" - if they could, they would be "free software projects" or "free source projects" - a much more specific designation that includes by definition the viral terms. It is your own confusion of copyleft and viral license terms that is causing the problem.
Copyleft is not the same as viral. The LGPL is a copyleft license that is not viral: all derivatives of the work must be free, but they can be combined with non-free works. GPL is both copyleft and viral, all derivatives, and anything which they are combined must be free (hence, viral, because it attaches itself to the larger project). So, if a library is GPL'ed, then any program that calls it must be free. If it is LGPL'ed, non-free stuff can use it. DanKeshet, Tuesday, April 2, 2002
- Only derivative works of GPL'd works fall under GPL. You can "combine" GPL code with anything, but if the combination constitutes a derivative work, then you'll have to GPL the whole code. The LGPL explicitly allows certain derivative works to be non-LGPL. In that sense, LGPL is "less copyleft" and also "less viral" than GPL. AxelBoldt
Agreed, copyleft is not the same as viral. That's why "viral license" was a separate article. Many patent and trademark licenses by consortia etc. are viral too. DanKeshet is right that GPL is "both copyleft and viral", and makes no reference to patents, e.g. does not require re-integration of patentable improvements, nor trademarks, e.g. does not require crediting GNU in its name, as per Stallman's absurd complaint that this "should" happen for Linux after 20 years of never so requiring such "trademark-like" recognition.
---
Suggestion for a breakdown:
copyleft more or less as is
free_software redirects to free source
free_source deals with both documentation and software and mentions patents and other instructions - there are already GPL'd patents - see openpatents.org
open_source contrasted with free source - the guarantee of "free" ensured by viral_license terms that "open" does not required
viral_license characterized more generally as the broad class of licensing terms e.g. in the Java or X11 consortium licenses which control those trademarks, e.g. in the IBM or MIT or Chinese govt patent licensing terms, e.g. in openpatents.org, e.g. in trade union use of instructions and training programs which usually require re-integration of improvements, e.g. in service licenses which require non-solicitation for competing services through the service itself, all of which seem to guarantee:
1. equal unharassed use for dedicated users of the service/license/patent/trademark/copyright who are not seeking an advantage for themselves at the cost of the other users, e.g. by extending "open source" and then patenting the extension, e.g. by bugging yahoo users with spam for hotmail.
2. a common simple procedure for extending/improving and re-integrating work, e.g. the Java or X11 compatibility tests.
3. an added-value market for features that do not affect interoperability, e.g. as in IP.