Paul,

On 2/23/2016 10:18 AM, Paul Jakma wrote:
> On Tue, 23 Feb 2016, Lou Berger wrote:
>
>> Paul,
>>    The link you provide is quite informative -- although I'm a bit
>> worried about encroaching on armchair lawyering .
> Ack, indeed. Best rely on best-practice docs and advice as much as 
> possible.
yup.
>> From that link:
>>> Not every contribution to a free software project is copyrightable
>>> Since copyright notices are not mandatory, there is generally no harm 
>>> in under-using them, particularly if authorship is recorded in VCS 
>>> logs. Over-use, however, can cause problems.
> Sure, but the very next bit is:
>
>   "It’s a good idea to have developers add their own copyright notices,
>   since a copyright notice is a legal representation being made regarding
>   the developer’s rights, but the project should also discourage
>   contributors from indiscriminately adding notices."
>
> And the document does suggest that adding a Copyright string at the top, 
> having made a copyrightable modification, is not indiscriminate (by 
> 'indiscriminate' they seem to mean automated / without regard to whether 
> any kind of copyrightable modification was actually made to a file).
yes, but a qualified statement not an unqualified one.  They make this
point explicitly and in their example.

> With my Quagga hat on, if someone contributes something and feels a need 
> to add "their" copyright claim, I usually'd have no reason to not put it 
> through.
Sure, but I'd hope you'd block someone adding a copyright statement on
others work.

> I'm going to hazard a guess that many corporates would tend to prefer 
> that employees making contributions in the course of their work add 
> Copyright strings, if there's any copyrightable material at all.

>
>>> the project should also discourage contributors from indiscriminately
>> adding notices.
> We havn't though had much of a problem with people adding notices when 
> they shouldn't though. The problem is more the reverse.

Why is that a problem?  As you say above, it's their choice.

> Also, if the wording is discouraging, I think that would 
> disproportionately affect non-large-corporate contributors. Contributors 
> working at large-corporates are more likely to have legal departments 
> who produce policy documents, checklists, even training courses 
> sometimes - all to encourage the appropriate adding of copyright 
> strings.

I was careful to select non-discouraging language in my proposal as it's
the author's choice.

> Non-large-corp-contributors won't have that, to balance out any 
> potentially-leading language from us. So they may as a consequence be 
> more shy about documenting claims.
>
>> Based on that, how about, instead add to current HACKING.MD:
>> +
>> + Care should be taken when adding copyright statements to existing
>> files as
>> + not every contribution to a free software project is copyrightable. See
>> + http://softwarefreedom.org/resources/2007/originality-requirements.html.
> Ah, interesting, though not very accessible? Seems more a legal text 
> than "best practices for engineers" thing? :(

right, so if someone has a question -- they need to consult their
lawyer.  Again, the language above is lifted straight from your reference.

>
>> +
>> + When adding a copyright based on a modification, you must also include a
>> + description of the modification along with the notice.  For example:
>> +           Copyright 2012 Jane Hacker <[email protected]> Fixed
>> random number
>> +           generator to output numbers other than 9.
>> +
> Hmm, adding a changelog type thing seems high-overhead given the SCM 
> data (plus, on the same line as the copyright string makes for a very 
> long line potentially). I'd possibly rather have that in the commit 
> message, and then have some git or automake magic to generate a file 
> from that when we 'make dist' to create a release tarball, and we only 
> need to police the git commit message.
>
> But, even that smells a bit high-overhead too. To really guarantee the 
> information, you'd have to put a file with all the diffs in distributed 
> tarballs. You'd have to replicate the contents of the SCM into the 
> tarball?

Hereto this language is lifted from your original reference.  And as you
say above  it's best rely on best-practice docs and advice as much as 
possible.  I think using their language is pretty safe while if we
really want to craft new language, it's time to ask advice from
corporate counsel -- which will make tis a very high overhead discussion.

>
> Hmmm...
>
> I don't really want to over-complicate this.
>
> - Adding such strings on new files: Everyone is fine on that, no doubt
>    (unless it's a copy? :) ).
>
> - Adding such strings on modifications: Encourage or discourage, but
>    some contributors will have their own 'encouragement' on that. Either
>    we push-back on all, inc. those, or we encourage all, to be fair.
Or we remain neutral on the topic, which is basically what the lifted
language does.

>
> - Replicating change data in the files, less keen on.
>
> I'm just not keen on discouraging people from adding these claims, just 
> on a 'feeling' basis - other than any vlagrantly "wrong" claims (e.g. 
> adding a string for a clearly unrelated entity; or adding one with no 
> other change of any real substance).
>
> Is there a need to discourage?

I see no reason to take a position one way or another.  If someone
doesn't want to add a copyright, it's their choice.  If they do, it's
also their choice as long as they don't copyright someone else's work.

Lou

>
> regards,



_______________________________________________
Quagga-dev mailing list
[email protected]
https://lists.quagga.net/mailman/listinfo/quagga-dev

Reply via email to