[d@DCC] Unintended consequences of treating the anonymous part of the Internet as other than BY-NC-ND

[mskala at ansuz.sooke.bc.ca]

On Sat, 21 May 2005, Russell McOrmond wrote:
> mskala at ansuz.sooke.bc.ca wrote:
> > Because in that case they're infringing my right to be identified as the
> > creator of that image.
>
> You are stating this, but not yet clarifying how making non-commercial,
> verbatim, royalty-free copies of your files in any way relates to your
> right to be identified as the creator?

You skipped the phrase "without attribution".  Whether it's
non-commercial, whether it's verbatim, and whether it's royalty-free, are
all irrelevant to whether it violates the right to be identified as the
author; whether it includes attribution or not is relevant.  So I can't
say that a "non-commercial, verbatim, royalty-free" copy does or doesn't
infringe my moral rights unless you tell me whether it includes
attribution.

I think maybe the difference here is that you believe "A verbatim copy
will always satisfy the author's right to attribution, because either the
verbatim copy will include the author's desired attribution, or else it
may be assumed that the author has waived the right to attribution."  Is
that what you believe?  I don't believe it.

I don't believe that the attribution must be included "in" the item being
copied (whatever "in" means - it's NOT obvious) in order for inclusion of
the attribution to be a requirement.  It isn't always possible to include
attribution "in-band" in the technical sense, and even if it's possible,
the fact that I didn't choose to do so should not *by default* constitute
a waiver of my right to have attribution included at all.  Attribution may
need to be included out-of-band, and a "non-commercial, verbatim,
royalty-free copy" might not do that.

It's very difficult, and the very kind of inappropriate meddling in
technical matters you and I object to, to try to characterize exactly what
is or isn't in-band attribution (and what is or isn't a "verbatim" copy).
For instance, suppose I write a document on a Macintosh and assign it an
icon that includes my name.  You copy it to a PC and the icon is lost as a
result of filesystem differences.  Is that a "verbatim copy" or not?  Was
the icon part of the document?  What if I post an image with attribution
information in the ALT attribute of the IMG tag?  Is it a "verbatim" copy
to post that same image file on some other Web site without the ALT text?
Is the ALT text part of the image?  I'd rather not have to code the
answers to those questions into the statutes.

The concept of something being "in" a "file" or not is a technology-
specific concept and shouldn't be in the law; but whether you lied or not
about who wrote a document is a question about social beings, not
about technology, and it's the law's job to deal with such questions.

> If you wanted to be identified as the creator of the image, you would do
> so in the file (metadata, RMI, watermarks, call it what you want).

I don't think that the fact I didn't do so should be an implicit waiver of
my right to attribution.  I think attribution should be required by
default unless I explicitly waive it.  At the very least, I should be
allowed to explicitly state that I require attribution, in a Web page that
includes an in-line image, and have that requirement apply to the image
file even when divorced from the Web page, even if the image file could
include a copyright notice according to its file format, and doesn't.  I
also think that answering the question of whether attribution was included
"in the file" or not is harder than you suppose.  The boundaries for
where you should look for attribution are not always obvious.

Although I have not done so (instead, I explicitly put the file in the
public domain), I think I could have claimed copyright, including an
attribution right, on the minimal-byte version of MoleSter available here:
   http://ansuz.sooke.bc.ca/software/molester/molester-min

There's no way I could have included a licensing statement in that file
without breaking the minimum-byte property which is an important part of
the point of creating that file in the first place.  Instead, I put the
licensing information in the enclosing Web page here:
   http://ansuz.sooke.bc.ca/software/molester/

> The "filename", "URL" or the page it is referenced from is a temporary
> technology-dependent issue that I strongly believe will backfire very
> badly (nasty unintended consequences) if it is allowed to constitute
> part of attribution.

There's more to it than just the URL.  I'm saying that requiring
attribution to be "in the file" is also a technology-dependent issue, and
inappropriate.

> In what way is someone who references an image on their page, or has a
> copy of a file under a different URL/filename/etc, making
> representations as to the authorship of the file?

The act of referencing an image on a Web page may or may not involve
making representations as to the authorship of the file; that description
doesn't provide the most important pieces of information needed to make
the determination.

> ...which makes this a temporary technical issue, where certain important
> features are not yet available in common tools.  The image needs to
> indicate who the copyright holder is (and metadata about the license in
> a standard format, etc), and reader tools should display this
> information to the audience members who care to know.

Unless we develop true AI, and unless society trusts the AI to perform the
functions currently performed by the law, I don't think it will ever be
possible for technology to solve all cases of the attribution question.
I don't think it'll even be possible for technology to solve the "image
downloaded from a Web site and reproduced verbatim, non-commercially, and
royalty-free" limited subset of the attribution question.

> The web page that references the image should be considered irrelevant.

Maybe in some cases it should.  I'd rather not code a universal "yes" or
"no" to that into the law.

> directly into the whole deep linking issue, and the question whether it
> is an infringement for me to link to a document (image, page, whatever)
> on your site without your permission.  The answer should be "of course

Linking to a document on my site *and claiming to be the author of it* may
be a violation of my legal rights.  Merely linking to it isn't.  It's the
making false claims about authorship that is the problem, not the link.
The link isn't itself a claim of authorship, but it may be part of the
claim of authorship, and I might well include it in the evidence in my
lawsuit.  It would be a very strange case where the link would be
sufficient evidence all by itself.

Linking to an image on my site and not claiming to be the author of it,
but without mentioning the source nor referring to the original context,
is not a violation of my legal rights, but it is very rude.  I should not
be allowed to sue you just for that and win; but I should be allowed to
issue redirects and otherwise make it difficult for you by technical
means.  This particular case actually occurs (with people linking to
images on my Web server from postings on phpBB systems, usually) quite
often.

Linking to a document on my site, not claiming to be the author of it, but
in a situation where you know or ought to know that it's likely to cause
problems for me in the nature of a DoS attack, might be illegal.  In that
case it has nothing to do with linking, and it's not under copyright law
that it might be illegal - the link was just the tool you used, and the
issue is that you caused a DoS attack.

> broadcast-era technology.  Any lack of technological neutrality in the
> law fights against the very existence of new-media such as the Internet.

I think requiring the attribution to be "in the file" in order to count,
would be an inappropriate lack of technological neutrality.

> If you are in an airport that has one of those browsers that doesn't
> show the URL location, does that mean that the airport is claiming to be
> the copyright holder for all content you see?

If they posted a notice saying they were the copyright holder for
all content you see, that could violate the rights of whoever actually
were the copyright holders.

> Are you suggesting that it should be considered a copyright violation to
> have a browser display a page differently than the author intended? (IE:
> am I infringing your copyright if I use Lynx rather than a browser that
> displays using your CSS file?)

If you make a printout of an excerpt from my page with Lynx or any other
browser, not including the part of the page that explains who wrote it,
and you publish that without mentioning where you got it, that could
violate my copyright.

> Note: The 12:00 flashing problem was fixed by having the date set via
> the network, not by the average citizen learning how to properly set the
> time ;-)   I think the same thing will happen with files on the network
> where document metadata will expand to properly identify documents in a
> network-location and technology neutral way.

I don't think the attribution which could be required to satisfy moral
rights can be fully captured by any amount of metadata, because an
identical sequence of bits may have different significance depending on
where it is and how it comes into being.  This is the point I'm talking
about here:  http://ansuz.sooke.bc.ca/lawpoli/colour/ .

> Today files are anonymous, and the recipient of a file only has the fact
> that the file is being distributed anonymously to go on as to whether

Who distributed it "anonymously" when it was originally posted on my Web
page with my name clearly indicated?  Whoever did, is the person we should
be looking at.  Just because I distributed it to you without asking about
*your* identity, or on an HTTP server capable, when given the right
request, of handing out the image without including text describing the
authorship, didn't mean I published it "anonymously" in the sense of
giving up my right to attribution.

> when all documents have full metadata, we will likely have filtering

All documents will never have full metadata, if "full" implies "truthful
and known to be truthful".  That's impossible because of the "bits have no
Colour" issue.  I can always say "Here's an example of a document with
false metadata:  blah, blah, blah" and someone looking at the "blah, blah,
blah" part alone wouldn't know that the metadata was false.

>    With a "non-commercial, verbatim, royalty-free" minimum assumed to be
> authorized when a computer is set to send the file to an anonymous
> recipients computer, it solves all the problems IMO other than the need
> to educate people about the technology.  If someone wants something

I think "with attribution" must be assumed as one of the conditions on
copying, unless that is explicitly waived.  I think "with attribution" is
not automatically implied by "verbatim", and I think it's more important
than those other conditions (not all of which I'm convinced are
necessary).
-- 
Matthew Skala
mskala at ansuz.sooke.bc.ca                    Embrace and defend.
http://ansuz.sooke.bc.ca/