Some question about code line count when sign GPL paper

Some question about code line count when sign GPL paper

[tumashu]

* Case 1

** I have a function in my package.
(defun test1 ()
  (let ((a 1)
        (b 2)
        (c 3)
        (d 4)
        (e 5))
    (print "test1")))


** Contributer add another new function which is very similar test1. 

(defun test2 ()            
  (let ((a 1)
        (b 2)
        (c 3)
        (d 4)
        (e 5))
    (print "test2")))  ;; the only different line from test1


should I count this countribution 7 lines or 1 line?

* Case2

If I can't let a contributer to sign GPL paper, How should I do:

1. just rewrite or revert his code and do not need to wash repo history?
2. remove his commit, by wash repo history?

Re: Some question about code line count when sign GPL paper

[Stefan Monnier]

> ** Contributer add another new function which is very similar test1. 
[...]
>     (print "test2")))  ;; the only different line from test1
>
> should I count this countribution 7 lines or 1 line?

The "line count" is approximative, it's really best to think of
"information content" (similar to kolmogorov complexity, in a sense), so
I think I'd count it more like 1 than like 7.

> * Case2
>
> If I can't let a contributer to sign GPL paper, How should I do:
>
> 1. just rewrite or revert his code and do not need to wash repo history?
> 2. remove his commit, by wash repo history?

It's best if it doesn't get into the repo's history, but it's usually
not worth the trouble "washing" the repo's history, no.

BTW, for your above example, I'd take the change as-is and then add
a new change on top of it which combines the two functions into one to
reduce the code redundancy.  It will also have the side-benefit of
clarifying that his contribution is really just that one line.


        Stefan

Re: Some question about code line count when sign GPL paper

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  >     (print "test2")))  ;; the only different line from test1


  > should I count this countribution 7 lines or 1 line?

The changes are limited to one line, but the change in that line
is trivial.  Perhaps it should count as 0.1 line.

There is no hard and fast rule here.  Ten lines of complex code
are much more significant than ten lines of trivial changes like that.

I'm trying to give a simple rule that approximates a judgment call.
There is no obvious correct way to do it.

-- 
Dr Richard Stallman
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)