[Proj] Method of reference to license within software code modules

[Gerald I. Evenden]

On Monday 05 January 2009 7:49:39 pm Frank Warmerdam wrote:
> Gerald I. Evenden wrote:
> > Not a bash nor editorial.
> >
> > Question: does/should a copy of the license that covers a program module
> > be placed within each and every source file??
> >
> > I had been putting a copy of the MIT in every file of whatever that ends
> > up in a distubution tarball.  Is that necessary?  I doubt that a copy of
> > the 10,000 lines of GPL ends up in each of Stallman's files.  ;-)
> >
> > Is a reference and url (such as
> > http://www.opensource.org/licenses/mit-license.php
> > sufficient?
> >
> > Anyway, I feel that I have been a bit excessive.
>
> Gerald,
>
> It is my practice to include a copy of the MIT license in each source file
> since it is of a managable size.

For some of my shorter files the license is longer than the code.  In 
[lib]proj I would not be surprised that 35 t0 40% of the text in the code 
distribution is license text.

Counting the characters in the RCS files (RCS adds some overhead to the bare 
file length) of the libproj4 RCS directory and the character count of the 
license times the number of files shows that 35.9% of the directory's content 
is repeated license text!  So my guess came damn close.

> However, for GPL projects it is common practice to name the license and
> version and refer to a COPYING file with the full license so that only one
> copy of the GPL (etc) needs to go with each project.

I am begining to feel that about 3-4 lines of copyright and reference to a 
COPYING file is the way to go.

> It is generally not considered good practice to refer to a license only by
> URL since web resources can easily be lost, change, etc.  So I'd encourage
> having at least one copy of the license in your source distribution.

I agree with that!

> Best regards,

Any other comments, please.

-- 
The whole religious complexion of the modern world is due
to the absence from Jerusalem of a lunatic asylum.
-- Havelock Ellis (1859-1939) British psychologist