Koma-Script doc/source files installed with doc/source disabled

[Max Chernoff]

Hi all, 

It appears like this may have slightly spiralled out of control. I had
just figured that there was a typo in a script or something, and I
couldn't find this exception documented anywhere. I'm not upset or
offended or angry or anything about this; the only feelings that I had
were curiosity or surprise. 

(replying to the emails in order)

> > I recently installed TL without any documentation or source, but I've
> > noticed that the Koma-Script doc/source files were still installed.
> 
> Providing KOMA-Script for multiple users without documentation is illegal. If
> you cannot accept this, just don't install KOMA-Script.

I really have no problem with this; I just found it unusual that I had
the source and docs installed for exactly one package.

> See the license for more information.

You should probably make this requirement more prominent. The CTAN page
just says "LPPL 1.3c", scrguide-en.pdf just says "see the file
lppl.txt", and "README.md" just says "see lppl.txt". The only file where
this is mentioned is "MANIFEST.md", which *is* quite clear:

> ### User manual main files
> 
> These files are typeset using LaTeX to produce the user manuals of the
> KOMA-Script bundle of LaTeX classes and packages. All these files are also
> include in the TDS and the CTAN files. All these files are mandatory for legal
> distributions of the KOMA-Script bundle of LaTeX classes and packages.
> 
> * `scrguide-en.tex` (English user manual main file)
> * `scrguide-de.tex` (German user manual main file)
> 

The LPPL does tell you to look in the manifest file, so this seems
legally sufficient, but it might be a little confusing for the users.
~3500 packages on CTAN use the LPPL, yet only Koma-Script requires
distribution of docs+source. I'd recommend that you also mention this in
the readme/manual just to make sure that people understand this
requirement.

To be clear, I think that the requirement is perfectly reasonable, but I
also think that you would have a better chance at compliance if you
mention it somewhere in the main documentation files.

> I can see nothing in Max Chernoff's message (extract below) to suggest
> that he is installing KOMA-script for multiple users,

Not that it's particularly relevant, but this installation is just for
me, on a personal laptop. In a technical sense, it is a "multi-user
installation" though since I installed it into /usr/local instead of
/home/max.

> And after thinking several times, I still reached the 
> conclusion that Max Chernoff /*cannot*/ and /*should not */be required
> to conform with the requirements of a multi-user licence if he is 
> installing on a single-user machine.  The fact that the TeX Live 
> infrastructure does not make it possible for KOMA-script to identify 
> unambiguously whether or not it is being installed on a multi-user 
> system is clearly a defect in the TeX Live infrastructure which should
> be addressed.

So I see this in a slightly different way. The LPPL is a free software
licence (I'm pretty sure that no one wants to get into *that* argument),
which generally means that it can put no restrictions on use or
modification, and very few restrictions on distribution. So I think that
we're all in agreement that on my own, single-user computer, I can do
whatever I want with Koma-Script, like changing all of the fonts to
Comic Sans, putting "\immediate\write18{rm -rf /}" at the top of all of
the files, or deleting the documentation. But what I can't do is make
those modifications, still call it Koma-Script, and distribute it to a
bunch of people.

The biggest question here is what counts as distributing an incomplete
copy of the work. If I put every single one of the files listed in the
Koma-Script manifest loose on Website A, then this should count as a
complete copy. This is what CTAN does, so it's certainly okay. On the
other hand, if I download a zip of Koma-Script, delete half of the files
at random, then put that zip on Website B with no other files, then I'm
definitely distributing an incomplete copy, which isn't okay. 

What if someone comes to the Website A, and only downloads half of the
files? I'm distributing a complete copy, and I have no way to force the
user to download everything at once. This is possible on CTAN right now,
so I think that this is okay. The LPPL lets you use modified copies so
long as you don't distribute them, so it would be fine to use these
incomplete files as long as you don't share them.

Now what if I want to give my friend a copy of Koma-Script, but his
computer can only use floppy disks? I'm using a really good compression
algorithm, so I only need 3 disks. I randomly divide up the files onto
three separate disks and give all three to him. I gave him a complete
copy, so this seems okay. And if he later loses one of the disks, that
cannot possibly be my fault, so that seems okay too.

Now lets say that I carefully divide up every in Koma-Script into three
separate zip files, one with all of the runtime files, one with all of
the doc files, and one with all of the source files. I put all of these
zips in a folder right beside each other on my Website C. Is this okay?
I'm distributing a complete work in a way that seems equivalent to both
Website A and the floppy disk scenario, so I think that this should be
fine.

If I come to Website C and download all 3 zips, I have received a
complete copy, so this seems fine. If I go to Website C and only
download one of the files, Website C has done nothing wrong, so this
still seems fine. But if I go to Website C, download only one file, then
put just that file on my website, we're equivalent to Website B, which
isn't okay.

My point here is that it would probably be *legally* okay for TeX Live
to divide Koma-Script into three separate zips, just like it does with
every other package.

One major problem is that partial mirrors of TeX Live would be generally
impossible. There would be certain subsets of packages that would be
okay to mirror, but other subsets that would be illegal. This could be
documented, and tooling could be developed around this, but it would
certainly complicate things.

But...

None of this matters at all since legality and morality are completely
different. Markus has politely asked that we always distribute the
documentation with Koma-Script. This may be perhaps eccentric, but it is
both reasonable and understandable. The real downside that I can see is
that this adds less than 15MB to a 3.5GB TL distribution. I see no
reason why TL should *not* follow an author's wishes.

> Hi Philip,
> 
> are you a volunteer whou would develop a multiplatform code which will
> delete KOMA script if installed without docs and a second user is
> added to the computer? 

I think that the multiuser questions are mostly a red herring here. It
may not be legal for someone else to distribute Koma-Script without
docs, but it also isn't legal to use TeX to produce fraudulent
documents. This isn't to say that we should encourage or otherwise
support users in breaking their licences, but I don't think that we
should feel obligated to prevent them from doing so. The only way to
ensure that the users follow the licences would be to add DRM to TeX
Live (which to be clear, no one is proposing).

> No, Zdeněk — I am a user of TeX Live who is defending the right of a 
> fellow user to install software for his sole use without being required 
> to conform to the requirements of a multi-user licence.  I find it very 
> odd that in the world of so-called "free" software (free-as-in-/libre/, 
> that is), the freedom of a user to install software for his sole use 
> without being required to conform to the requirements of a multi-user 
> licence can be such a contentious matter when it is clearly his 
> inalienable right.

Nothing at all is preventing me from downloading and installing Koma-
Script without any of the documentation or source. I just need to go to
any CTAN mirror and download a half dozen files one-by-one. Nothing is
preventing me from doing so. It's not my right to force the authors of
free software to add specific features. 

TeX Live bundling the source with the runtime files of Koma-Script is no
different than Debian bundling Koma-Script with Memoir or Microtype or a
few dozen other packages

   https://packages.debian.org/bullseye/texlive-latex-recommended
   
and the issue of Debian+LPPL has already been discussed **extensively**.

> Perhaps Max should take his case to the European Court of Human
> Rights.

Aside from being unnecessary, I'd need to move to Europe first (I'm
Canadian).

> The user can install TeX Live without KOMA script and then install
> KOMA script from CTAN without docs. [...] The author has the right to
> choose a license and the distributor MUST follow it.

Exactly.

> So on the whole I think even offering a switch to only receive barebone 
> dropping both source + documentation is bad and should not be provided.

To be fair, the "no docs" switch is quite hidden. Someone could perhaps
add additional warnings to the switch to dissuade users from using it,
but it's still necessary for things like CI/CD systems or Raspberry Pis.
Many users also have metered data, so the extra download could get very
expensive. Where I'm from, mobile data is $5/GB and is often the only
choice in rural areas, so installing docs could cost someone upwards of
$15.

> I want to be able to tell user, see
> 
>    texdoc usrguide
> 
> or
> 
>    texdoc array

Yeah, that seems understandable. As a package author, I'm well aware of
how many user questions are answered directly in the documentation.
Since most users have internet these days, the following will provide a
partial solution:

   https://github.com/TeX-Live/texdoc/pull/79
   
> In my opinion too many users (these days) never 
> look at documentation but go and and bother others with questions 
> without first researching their problem and I don't want to see this 
> being even institutionalized by TL offering them to make *sure* they 
> have no documentation.

Agreed, but how many of those users would even know *how* to find local
documentation? To access local documentation, you either need to use
"texdoc" on the command-line, or browse through a very deep and somewhat
confusing TDS tree. The easiest way for less-experienced users to find
documentation is to Google "PACKAGE-NAME manual", go to the CTAN page
and click on the manual, or use the texdoc.org.

> Until this time, we are honoring the license requests by authors.

This seems like the best decision to me. I see no reason to go against
the author's wishes for something so minor. The only real issue that I
could potentially see here would be "what if everyone did this", but TL
has existed for nearly 30 years and this has happened precisely once, so
I wouldn't even be concerned about that.

> The author has the right to select a license. Users can dislike it but
> that's all what they can do. Distributors has the right to decide how
> they will cope with the distribution in order to honor the licenses.
> The question whether the computer is used by a single user is not
> sufficient. The user will have to be clearly informed about all
> consequences. The user must know that since that time he is not
> allowed to grant access to another user because he/she will violate
> the law. The distributors (her TL maintainers) have their rihgt to
> choose how they will prevent the risk of legal problems. Thus the
> current installation procedure does not have a bug, this is just a
> feature.

This seems correct to me. Although it _may_ be legal for tlmgr to not
download the docs/source, it would definitely not be legal for the
receiving user to redistribute the package. It would be a bad idea for
TL to make it so easy for a user to violate a licence like this.

I'm perfectly happy with how Koma-Script is currently packaged; I was
just curious about why it was different. Thank you everyone for all of
the replies, and sorry for stirring up such a kerfuffle. 

-- Max