Re: typo in the new W3C Document License? It does not allow "use"

Oops! Thanks for the reminder. I thought I'd sent this out, and have 
just realized that I didn't...

Here you go...

Kirill Gavrylyuk wrote:

>Hi Patrick, 
>I'd like to look at your submission prior to the teleconf - is it
>publicly available?
>-----Original Message-----
>From: Patrick Curran [] 
>Sent: Thursday, February 20, 2003 8:27 AM
>To: Kirill Gavrylyuk
>Cc: Joseph Reagle;
>Subject: Re: typo in the new W3C Document License? It does not allow
>Indeed. I've submitted a detailed critique of the Document License, and 
>we'll be discussing these issues soon.
>Kirill Gavrylyuk wrote:
>>Hi Joseph,
>>Rereading the new W3C Document License [1] from December 31^st , 2002,
>>I find that it allows only copying and distribution:
>>& Permission to copy, and distribute the contents of this document, or
>>the W3C document from which this statement is linked, in any medium 
>>for any purpose and without fee or royalty is hereby granted, provided
>>that you include the following on /ALL/ copies of the document, or 
>>portions thereof, that you use:
>>Where as the previous W3C Document License [2] from April 5^th , 1999,
>>that we always used in our submissions in 2001 and 2002 allows use, 
>>copying and distribution:
>>& Permission to use, copy, and distribute the contents of this 
>>document, or the W3C document from which this statement is linked, in 
>>any medium for any purpose and without fee or royalty is hereby 
>>granted, provided that you include the following on /ALL/ copies of 
>>the document, or portions thereof, that you use:
>>So essentially the new Document License does not allow to use the 
>>materials published under it. Which means one cannot even read the 
>>materials, let alone use them to design implementations.
>>Is it a typo or intended? I believe this was brought up on XML Schema 
>>WG as a blocker for using the license for test materials.
Although I fully understand the desire to limit the 'scope' of
derivative works, and specifically, to prevent the modification of the
tests, I have some serious concerns about the applicability of the
Document license to software. 

Not surprisingly, since the license was devised for specifications, the
terms "software" or "code" are not even mentioned. Even if we extend the
interpretation of "document" to include "code", the license does not
seem to grant the rights to "use" the software that I believe were
intended. (It grants the rights to "copy and distribute" but explicitly
does not grant the additional "copyright rights" of "perform and
display", which could perhaps be stretched to imply "use" in the
software sense.)

The situation is made more problematic by the fact that many W3C "test
suites" are not in fact complete and self-sufficient test suites, but
rather collections of tests and some documentation from which complete
test suites must be constructed (by adding a test harness, additional
documentation, etc.) 

This implies that the kind of "use" we will need to grant will often go
beyond the "take it out of the box and run it" kind of use that applies
to a packaged software application. Sun's legal counsel is concerned
that the Document license does not definitively grant licensees the kind
of rights that I think we want and need to grant them. For example, the
right to incorporate tests (unchanged) into a larger test suite, the
right to add extra documentation, the right to change directory
structures, and so on. Moreover, our counsel is concerned that these
rights would not be granted even if the rights to "perform and display"
were added to the Document license.

Bottom line: the Document license was intended for documents, and seems
inadequate for sofware, even if some of that "software" is more
analogous to data files. It would probably be easier to start with a
software license that explicitly grants use rights, modifying it if
necessary to prohibit modifications. 

A suitable model might be a license for some kind of compiler product
that grants licensees the right to dynamically link against library
files produced by the compiler, and to redistribute these libraries with
their applications, but that prohibits "disassembly" or "reverse-

Tests, like libraries, have - or should have - a well-defined and
documented interface that allows people to 'link with' them. This
interface is the test 'metadata' that provides information about how to
invoke the test, any inputs required, and the expected output. Such an
interface supports the creation of a test harness.

So - I think we need a license that explicitly permits licensees to
copy, redistribute, and "link with", or execute, the tests, but that
prohibits their modification. "Execution" is a somewhat difficult term
to define. Where actual executable code is provided, the meaning is
clear, I think. However, many "tests" distributed by W3C would perhaps
more accurately be called "data files", since they are fragments of XML,
for example. These are fed to a parser or processor, and the output is
compared to what is expected.

Nevertheless, I think this is what we want. A real-world test of whether
the license is usable might be to determine whether we can answer "yes"
to the following questions (assuming, of course, that we want to permit
this action). In each case, we must consider the answer when the "testsuite"
is actually an incomplete set of tests.

Is a licensee permitted to:

(1) create a derivative testsuite from a W3C testsuite?
(1.1) by adding extra tests to it?
(1.2) by removing some tests from it?
(1.3) by removing some other components (data files, documentation) from it?
(1.3) by adding some code (programs) to it?
(1.4) by adding extra documents or files to it?
(1.4.1) by adding extra files to directories in the W3C testsuite?
(1.4.2) by adding extra files to directories outside of the W3C testsuite?

(2) instruct its own licensees to perform the actions described above by
downloading a W3C testsuite from a W3C web site and combining it with
material supplied by the original licensee?

(3) execute only a portion of a W3C test suite (for example, excluding
some tests from the run on the grounds that they have been determined to
be invalid)?

Received on Thursday, 20 February 2003 13:32:14 UTC