[Savannah-hackers] [ 101951 ] Legality Issues

[nobody]

Support Request #101951, was updated on Thu 03/13/03 at 08:21
You can respond by visiting: 
http://savannah.gnu.org/support/?func=detailsupport&support_id=101951&group_id=11

Category: None
Status: Closed
Priority: 5
Summary: Legality Issues

By: yeupou
Date: Thu 03/13/03 at 12:27
Logged In: YES 
user_id=1896
Browser: Mozilla/5.0 (compatible; Konqueror/3.1; Linux 2.4.21-pre4; i686; fr, 
address@hidden)

Hi, 
 
About legal issues,  I think you are right. In doubt, you may ask to 
address@hidden but they are busy people and the anwser may 
be late. 
 
About CVS, there is not absolute rule, it may depends on the 
situation, but usually you commit when you are satisfied with 
something. Your function work, you may commit.  
You begin to work, you do "cvs update". When you stop, if what 
you planned is done, or even if it's not done but if someone else 
may work on it, you should do "cvs commit" 
 

----------------------------------------------------------------------

By: qckcode
Date: Thu 03/13/03 at 08:21
Logged In: YES 
user_id=16521
Browser: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.0) Gecko/20020623 
Debian/1.0.0-0.woody.1

I'm not too sure how on-topic this is; if it's off
topic, a pointer to a doc or an area to research would
be very helpful.

Before I start developing/cvs-committing my code, I
thought I'd make one last check to avoid possible legal
issues in the future:

Ideas are non-patentable right?
Especially:
  * if I read an algorithms text, take the pseudo-code;
and transform it into real-working C++ code, then this
is not a copyright violation of any sort right? In
particular, the pseudo code form "Introduction to
Algorithms" by Cormen, Leiserson, and Rivest

  * if I read an idea on how an piece of code is
implemented (regarding of license, as long as there is
no non-disclaimer agreement), and learn how it works;
and implement/code my own version, w/o taking/copying
any lines from previous code; then this is not a
copyright violation of any sort right? Particularly, I
have: the "visitor concept" from Boost Graph Library
(http://www.boost.org/libs/graph/doc/visitor_concepts.html)
in mind.

Some highly optimized / very classical ideas seem to
have implementations that lend themselves into one
"trivial" form of code, which might result in multiple
coders coming up with the same idea -- what should I do
when writing/committing such pieces of code.

For example: a breadty first search on a weighted,
directed graph -- one essentially must have a priority
queue, a few arrays (distance, reachable, parent); and
pretty much a 4 lines of code per 1 line or pseudo code
replacement from books/websites.

Lastly, along the same line, how often should I
cvs-commit my changes to show proof of "development" of
code? Every file? Every function? Should non-debugged
code by CVS-committed as well? Code that generates
warnings? How about non-compilable code? Should I have
a separate (already created, heh) TESTING directory
where stuff are _buggy_ and periodically move stuff
from TESTING back into the main "liblia" as it stabilizes?

Insight/information/docs/pointers for the above will be
much apperciated.

Thanks in advance,
--qckcode

----------------------------------------------------------------------
You can respond by visiting: 
http://savannah.gnu.org/support/?func=detailsupport&support_id=101951&group_id=11