BSD and MIT license "compliance" with the MS-PL

Tzeng, Nigel H. Nigel.Tzeng at
Fri Apr 17 18:48:54 UTC 2009

>From: Matthew Flaschen [mailto:matthew.flaschen at]

>Tzeng, Nigel H. wrote:
>> 3.2 requires source availability of modifications to source code.  
>> 1.9 Defines what modifications are.

>Yes, but 3.6 says 3.2 applies for /any/ executable-only distribution 
>(meaning there has to be source offer) of Covered Code.  The Mozilla 
>FAQ corroborates this.  It doesn't say it only applies for executables 
>made from modified code.  See , 
>specifically "I want to distribute Firefox (or other MPL-covered code) 
>that I have compiled myself but not changed. What do I have to do?"

It says you have to tell folks where to find the code.  There's no obligation for you to distribute the source with the binary as you assert.  The notice fulfills my obligations to MPL.

>> It's not a side point if you assert that all copyleft licenses 
>> require source distribution with all binary distributions and one does not.

>If MPL didn't require source distribution, then MPL wouldn't be 
>copyleft.  That doesn't change the definition of a word.  But you have 
>a very heavy burden in trying to prove Mozilla interprets their own 
>license wrong on this fundamental point.

Mozilla is not misinterpreting it's license wrong because the FAQ clearly indicates that your obligations in distributing binary from unmodified source is to add a conspicuous notice stating where to find the original source.  Not that you must also include the source as part of your binary distribution.

You are also not obligated to host the source for the 12 months as stipulated in 3.2 for modified source.

Heck, you can even not supply the entire modified source and just the diffs and point upstream for the unmodified source according to their FAQ.

You can claim that MPL isn't copyleft but I think the burden is on you to provide evidence to the contrary.  Reasserting your claim/definition is not evidence.

>> That is the definition for strong copyleft yes.

>No.  The difference between strong copyleft is whether new files are 
>included in the copyleft.  Trivial example:

>Original program made from A.c and B.c, both under weak copyleft license.

>Derivative program made from A.c, B.c, and X.c.  X.c does not need to 
>be under the copyleft license.

Yes, X.c remains proprietary (or whatever) and if you haven't modified A.c and B.c under the MPL you do not need to distribute the source.  Just point to your upstream source repository in your about:

Source for A.c v1.2.2 and B.c v4.3 can be found at and

>> Perhaps the MS-PL license was not
>> accidently named and appropriately categorized as a permissive 
>> weak-copyleft open source license.

>Where is it categorized as copyleft (besides the FSF list)? many lists are there?  The OSI doesn't have that kind of list so the FSF one is about as canonical as you get.

More information about the License-discuss mailing list