Joe Moore <[EMAIL PROTECTED]>: > Ah, here's an analogy that makes sense. Consider software that is an > anthology of works by several authors under several licenses. Clearly the > compliling (not in the software sense) author has performed significant > work. However, the license for the anthology does not necessarily trump the > license for the individual works. You can make any use of public-domain > writings, even if the copy you're reading from is a copyrighted volume. > > I can create derivative works from Homer's _Illiad_ even though the copy I'm > basing it on is in the Norton Anthology of Literature.
Most ancient documents exist in many different versions. There is significant work involved in putting together a particular text. I would guess that this work is covered by copyright, so you can't just take the text from a recent and expensive academic edition and put it on the web. Even if the compiling author is claiming to follow a particular edition that is out of copyright, I would guess that the compiling author still gets a new copyright; if a book is old enough to be out of copyright, obtaining a copy of it is a non-trivial task[*]. Small differences may be introduced into the anthology, by accident or deliberately, and these differences would be sufficient to prove that you took your text from the anthology rather than from the original edition. So, extracting a public-domain text from a non-public-domain anthology looks quite risky to me. This is why we need copyleft, of course. I hope that extracting a DFCL component from an anthology will be a safe thing to do. Edmund [*] I once wanted a reliable text of some Edgar Allan Poe. There were versions all over the web, of course, but they were all slightly different. I made a list of the differences and consulted a microfilm copy of an original edition in Cambridge university library to resolve them. I invested several hours work into that exercise, so maybe I could, as a sort of amateur textual critic, claim copyright in the final result. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
