Comparing OSI to OASIS and W3C

[Forrest J. Cavalier III]

David Webber (XML) wrote:

> Sweet.
> 
> What this is missing is the matrix of checkboxes - so you can easily see the
> crossreferencing - and determine by that - which license(s) is more closely
> suited to your purpose.

The problem is that licenses are messy.  The matrix you want may not be
the matrix someone else wants.  There are many, many columns and many, many
rows in what amounts to be a sparse matrix.

I know you want a table.

There is a full list of symbolic license characterstics (so far) here:
      http://mibsoftware.com/librock/lidesc/tags.htm
(Click on the "example" link at each line for example licenses with that
tag.)

If you grab the .lh files from LIDESC, you can easily create your table
with any columns you want. Look for the .lh files in
     http://mibsoftware.com/librock/librock/license/
The #define librock_LIDESC_TAGS_[license name] string is the one you
are interested in.

LIDESC can also be used to pick an appropriate existing license.
Instead of writing a new license, if you check off the tags you want and submit,
at tags.htm. LIDESC will offer a report and example licenses ranked by similarity.
That gets you a short list.

 > I love the registration of license certificate too. Killer.

No registration.  That's the SHA-1 of the license text.  LIDESC was designed
so there is no need for a controlling authority.  You do want competent people
matching up tags, but what you keep in your conflict and tags DB is up
to you.  (If you create different symbolic tags, please do not use the
librock_ prefix.)

Because there is no central authority, that does place some practical limitations
on how LIDESC can be used : canonical license texts are sometimes hard to create
(some OSI-approvals had non-ASCII in the submission to the list.)

And starting last month SHA-1 is assumed a bit more collision-prone than when the
LIDESC software was written three years ago.

But, all in all, problems like LIDESC collisions are probably pedantic.

The LIDESC system is not meant to thwart plagiarism and perjury.  It is not a
license manager. Instead is meant to let reasonable people keep track of what
they have and report license conflicts when creating derivative works.

The licenses are referred to by a recognizable and unique one-line reference, the
license stamp, so it can appear everywhere it is inconvenient to put all the license
terms, (like every function body of source code, for example.)  That catches some
cut-n-paste derivative works if you use it responsibly.

LIDESC works especially well when you set up a personal profile of license terms you
find unacceptable. Then you can report if license XYZ is incompatible with your
personal profile, and avoid tainted source.