Jun. 12, 2020: initial commit
* gpl-lgpl: read/worked up to I.1.2.3;
a few changes in rewriting some language for clarity (larger
changes made as comments to come back to), with sporadic, minor, unmarked
edits to language on first read-through

Additional connecting text for irrevocability discussion.

A forward reference is added to connect to the irrevocability section, and
one transition sentence added in the irrevocability section itself, since
it's another "digression" from the walk-through of GPLv2 in these sections.

More detail on Irrevocability of GPLv2

This section, fit in just after the detailed discussion of GPLv2 Section 6,
explains in futher detail various arguments for why the GPLv2 is irrevocable.

More words that don't appear in some dictionaries.

Add to the LocalWords list for Emacs spell checker.

Clarify & expound on issue of warranty disclaimer

The existing paragraph on this issue was inadequate, since it punted
entirely to GPLv2§11 for dealing with critics' claims of
unenforceability. That left a mistaken impression of validity of such
claims.

The commit herein adds reference to CIGS, which likely permits GPL's
sort of warranty disclaimer in most jurisdictions, and also bolsters the
reference to the UCC earlier in the section.

However, given academic debate about the applicability of CIGS to
software licenses, this commit includes a footnote referencing the two
sides of that debate.

Tony Sebro and I co-drafted these changes together.

Signed-Off-By: Tony Sebro <tony@sfconservancy.org>
Signed-Off-By: Bradley M. Kuhn <bkuhn@ebb.org>

Note that most USA states/commonwealths adopt UCC.

While most USA lawyers will know this as simple fact, the tutorial wants
to welcome an international audience and non-lawyers too. As such, this
dependent clause is worth adding.

Clarify who "most people" meant, drop UCITA ref.

The intent of this text was to point out that most users don't actually
believe they get warranties, which is still surely correct, given that
GPL disclaims warranties in the same manner nearly every software
license -- proprietary or free -- does anyway.

Also, the forward-reference to the later section's discussion of UCC
should be hinted at here. There is no explicit reference to UCC made
here, but it is encompassed in "many local laws", since the later
section mentions the specific section of UCC involved.

Meanwhile, the reference to UCITA is dropped, but perhaps it should be
reintroduced in other text in the main warranty section. UCITA has had
much less policy impact than was expected when the original version of
this text was written. It might be useful to ask policy folks and
attorneys from Maryland and Virginia who might be able to help explain
what impact UCITA has had being on the books only there.

Suggestions for in depth discussion of GPLv2§2(a)

GPLv2§2(a) is mentioned quite cursorily. Olly Betts suggests more
coverage here might useful on IRC. I noted some ideas after he
suggested it.

Rework text to introduce definition of CCS.

CCS ultimately wasn't mentioned until much later in the GPLv3 sections,
where, ironically, we have to point out that GPLv3 defined the term as
"Corresponding Source" [0], not CCS, and explain why GPL enforcement
wonks still say CCS.

This rework now introduces the acronym at the natural moment: while
describing GPLv2§3's use of the words "complete" and "corresponding".

Adding that made the section even more disjoint than it already was. I
put in some \subsection's to make it slightly less so, and did some
wordsmith work on surrounding text.

[0] I wish some GPLv3 drafter had asked me what to call the defined term
so that I could point out what fit standard parlance. :)

"by definition" is clearer here than "by default"

There's probably another phrase we could use that's even better.

FIXME re: "dynamic linking delayed" to runtime.

The existing text of the Guide hints at this point but doesn't discuss
it directly. This FIXME is merely a reminder note to investigate this
issue in further detail and perhaps add text here on the question.

Consolidate copyright notices into single place.

Historically, this project used (more-or-less) a file-by-file copyright
inventory. This commit ends that practice. The project now has a
single toplevel copyright inventory, stored exclusively in
comprehensive-gpl-guide.tex (so that it appears also in compiled
versions of the Guide as well).

The side-effect of this commit is that the parts may no longer be easily
publishable separably without (at least) the additional work of
copyright notice reconstruction. This may in particular create a
challenge for the FSF, who has historically selectively published
sections of this Guide as materials for its CLE classes.

However, without this change, this Guide will eventually suffer from the
inherent problems in maintaining file-by-file copyright inventory.
Circumstances simply dictate a single, top-level copyright and license
notice for the entire Guide.

In addition to consolidation of copyright notices, I've also herein
updated my historical copyright notices to properly credit me for my own
work done in 2003 through 2005.

I've also updated the license notice to reflect the changes made by the
previous commit and related issues.

Citations of third-party works in appendix only.

As alluded to in 2ea19b71d4a917babb29024f06acabfe73309f40 's commit
message on 2014-12-17 19:52:15 -0500, keeping any information on a
part-by-part basis is difficult and error-prone, since there exists no
reliable way to auto-generate such information accurately.

Therefore, citations to third-party works, in addition to remaining
fully documented in the commit log as they always have been, are now
placed in specifically one location in the body of the text itself: a
single appendix specifically designed for that purpose.

In this manner, contributors have no house-keeping work regarding
citations. Contributors need only list third party works and links in
one place: third-party-citations.tex.

Documentation in CONTRIBUTING.md for making contributions of third-party
works is left as a TODO.

Wordsmith sentence after phrase "such software".

I think Fontana's change a few commits back which s/it/such software/ in
this sentence is a useful one. However, the entire sentence reads even
worse because "such software Free Software" looks so strange as a string
of words in the middle of the sentence.

This change is the best rework I could find that resolved the problem.
It's probably not the best option but is certainly an improvement.

Replace editorial commentary w/ FSF's link.

My personal comment here, which I wrote on 2003-05-26 (see
f05ce6c657e07a5e6c6def3f7ff8cb2b2bcf6246 ), is probably not particularly
useful. I still tend to use the phrasing as original stated in the
removed text herein; however, I'm admittedly the only one. I don't deny
that I hope to coin some terminology usage through my work on this
Guide, but this particular use of "nonfree software" to mean
"noncommercially proprietary" is not so important IMO that this Guide
must coin it.

The FSF's page on this doesn't make that distinction, and has much more
detail on this issue than this section does. Therefore, the
personal statement is removed, and the organizational statement on the
FSF's site is instead linked to.

Some editorial changes.

Commenting on one: the initial-caps stylistic preference for "Free
Software" (though it contradicts prevailing usage, including that of RMS
and the FSF) ought to be respected, but I think it is confusing to
capitalize the 'S' when referring to nonfree software as "non-Free
Software". So I changed this to "non-Free software" and also implicitly
acknowledged that the preference for "non-Free" over "nonfree" is the
editor-in-chief's stylistic idiosyncrasy.

move long parenthetical text out of short sentence

it's easier to read this way.

correct literal meaning of phrase

GPL protected code-bases don't demand industry trust, modern industry
trusts GPL protected code-bases.

Clarify sentence.

Make this sentence a bit clearer.

Remove spurious line.

This line was carried over as a paste from a previous version of the
paragraph above it, and is now no longer needed.

Additional published work.

This publish work was also referenced in the old commit logs already,
but now that we have a list in the work itself, it can also be listed
here.

Add hyperlinks to remixed published works.

Now that a list of published works is part of the document proper, it is
reasonably practicable to add hyperlinks to those works.

Remove incomplete lists of authors entirely.

The lists of authors in each part has been continually out of date and
incomplete. There are multiple examples, here are a few:

* In September 2005, John Sullivan made improvements and was not placed on
the Authors lists until I did so in a March 2014 commit.

* In March 2014, Martin Michlmayr submitted many patches, but was not
placed on the Authors lists until I did so in an April 2014 commit.

There is no easy way to keep these Authors lists current, and they aren't
necessary under CC-BY-SA-4.0 anyway, so I herein remove the Authors lists.
Additionally, previous commit added "published sources" in each part, which
is more static and easier to keep up to date and provides similar
information.

List published works & link to Git logs.

References and details regarding these published works from which some
text was incorporated already appeared in the commit logs in great
detail. The information, already fully available in the Guide's Git
logs in full compliance with CC-BY-SA-4.0 §3(a)(1-2), now appears in
summary form additionally in the compiled PDF/HTML/Postscript output.

not all cc sharealike licenses are copyleft, so be specific
automatic relicensing is an exaggeration
cc by-sa case is even more nuanced, briefly mention in fn

Spell check and correction.

Seems there were a lot of correctly spelled words missing from the
LocalWords lists as well.

Grammar and wording fix.

I think this sentence was improperly merged together with another one in
a previous commit and therefore needed correction.

Replace ’ with '.

The unicode ’ was introduced by the pasted text mention in the previous
commits. While I believe LaTeX can be configured to accept Unicode
quote equivalents, it seems simpler to me merely to replace the
character with an appropriate version that LaTeX expects in this
situation by default.

Integrate pasted text regarding LGPL.

Much of the pasted text here was useful. However, some of the claims
were broad reaching, I've reigned those in. (e.g., saying "Taken
together, these provisions mean:" was a bit strong).

Also, in that specific spot, the conclusions made in the text were
described as applying to LGPLv2.1, but are clearly conclusions about
LGPLv3. I've corrected that herein.

Finally, I had to write a bunch of next text to make the pasted text
work, and also added one FIXME for later of where things could be
improved further.

Pasted sentence usable as is.

I simply combined this pasted sentence with the previous paragraph; I
think it probably works as-is.

Incorporate pasted text re: AGPLv3/GPLv3 combos.

This text was mostly useful as is. However, it failed to make a key
point I've often made: that the combinations created by comingling
AGPLv3'd code with GPLv3'd code may be difficult to disentangle, and
thus in practice, it may turn out that such a combination effectively
must be licensed as a whole under AGPLv3, even if technically some
copyrights included therein are GPLv3'd.

In practice, this nuance is only a technical barrier, since complying
fully with AGPLv3 automatically complies with GPLv3.

Comment out text that may or may not be useful.

I'm not sure this text is actually useful, so I've commented it out.

Integrate pasted introductory text for GPLv3§11.

This text was (on the whole) useful as introductory text to this
tutorial's existing extensive section on GPLv3§11.

The example, however, belonged further down in the section, so I've
placed it there.

Integrate pasted text on GPLv3§10.

Most of this text was useful, particularly since there was a previous
FIXME here that GPLv3§10 was not extensively discussed.

However, the same footnote regarding Jaeger's opinion under German
copyright law applies to this text, so a reference back to it has herein
been added.

Integrate pasted & commented out text for GPLv3§7.

The pasted text, most of which was useful, is now integrated as the
desired laundry-list of GPLv3§7 subsection explanations.

This also allowed for easy integration of some of the older
commented-out text that originally came from GPLv3 rationale documents.

Meanwhile, however, I discovered, upon more careful examination of the
pasted text, a serious and grave error regarding GPLv3§7(d).
Specifically, GPLv3§7(d) makes the modern "third clause" of the 3-Clause
BSD compatible with the GPL, *not* the problematic old-school BSD
advertising clause (from the 4-Clause BSD).

I'm amazed that anyone versed in licensing could make this error,
frankly, and readers should be told, since other materials are now
disseminated by others, that the point is incorrect. Therefore, I've
not only noted the correct compatibility conclusion, but also
affirmatively identified the incorrectness of the wrong conclusion that
was previously added via the pasted text from SFLC's "Guide".

Finally, on a LaTeX formatting note, the enumitem package is now needed
since I'm using that for the list of GPLv3§7 subsections.

Rewrite pasted paragraph with historical context.

This change perhaps is somewhat controversial, but reflects honest
reality of this history of additional requirements on GPL. The
additional requirements that GPLv3 permits mostly represent historically
known situations where GPLv2 permitted license compatibility with Free
Software licenses containing such requirements.

Orthodox compatibility theory demands that such additional requirements
have explicit codification in a copyleft license, which hints at why
GPLv3 needed to include this section.

However, historical copyright holder toleration of these sorts of
requirements placed on GPLv2 works is well-documented, and failure to
mention it here is a disservice to the reader.

Integrate introductory paragraph from pasted text

Most of the pasted text served as a useful introductory paragraph.

I also made some wordsmith changes to the following few paragraphs.

Once sentence here was not redundant.

Most of the pasted text was redundant to the existing guide text, but
one sentence reiterates a useful point in a creative way, so that
sentence is herein integrated into the preceding paragraph.

Incorporate pasted text into intro paragraph.

Some of this text was useful as part of the introduction paragraph.

Simply remove this pasted text.

Upon second reading of this text, I don't know why I saved it. The
existing guide already covers these issues adequately, and there is no
specific place I could find where this new text worked well as an
introductory paragraph (as has been done with others of these pasted
texts).

Wordsmith last sentence of the footnote.

The word "nearly" appeared far too much in the previous version of this
text.

Qualify "additional copies == new license" claim.

The last commit brought in text that categorically claims: "automatic
termination cannot be cured by obtaining additional copies from an
alternate supplier". While this position is by far the overwhelming
majority position among copyleft advocates, theorists, and legal
experts, the small minority dissenting opinion is simply too strongly
sourced to ignore.

Specifically, Till Jaeger's position was central to Harald Welte's
gpl-violations.org community-oriented GPL enforcement efforts.
Therefore, this tutorial must include his position when covering the
issue of automatic license reinstatement in this tutorial.

I have told Till that I can't believe his position is possibly correct.
(I understand that many other copyleft theorists and legal experts have
done so as well.) However, Till remains steadfast that this position is
correct, at least under German copyright law. Speaking for myself, I
have never met a legal expert as well-versed in both copyleft and German
copyright law as Till Jaeger is, and therefore I cannot in good
conscience allow this tutorial to remain silent regarding Till's
position, lest the tutorial propagate an inappropriate bias for the
majority belief.

That said, I still feel that a footnote is the right place for the
argument. It *is* a tiny minority position [0] among an overwhelming
consensus to the contrary, and therefore adding the point to the main
text would only serve to distract the tutorial reader.

[0] In particular, I am convinced Jaeger's argument, if true, is a
peculiarity of German law exclusively. For example, French lawyers
I've spoken with believe that the standard USA legal position on
this issue is also accurate under French copyright law. I therefore
conclude the minority position (if accurate) is unrelated to
differences between civil law and common law copyright regimes, and
is instead a unique peculiarity to German copyright law.

Edit pasted automatic downstream license text.

This pasted text was quite useful and well-written. I've incorporated
it fully, although it needed some copy edit work and additional
material.

Add subsections for GPLv2§2 discussion.

The GPLv2§2 has gotten long enough that it really needed subsections.

Integrate pasted text on "separate & independent".

This pasted text was pretty useful, and is now integrated fully, with
additional text I wrote to improve and expand the point.

Integrate pasted explanation of GPLv2§2(c).

The tutorial did not previously have an adequate description of
GPLv2§2(c), so this pasted text is as good as any that I could think of,
although I reworded it slightly.

I chose to integrate this text *before* the discussion of GPLv2§2(b)
because the latter section requires so much attention, it seems
appropriate to get the less complex parts of GPLv2§2 out of the way
before diving into that essential discussion of GPLv2§2(b).