discuss: Thread: Dead links; linuxdoc.org (was: Dead link in a HowTo)

On Wed, Nov 12, 2008 at 09:14:38PM +0100, jdd for http://tldp.org wrote:
> Mail Lists a ?crit :
> > Hello,
> > whilst browsing the mirrors I sometimes dip into a howto and came across this 
> > dead link in the Home-Network-Mini-HowTo.  The target is
> > http://www.coastnet.com/~pramsey/linux/w_hub.gif 
> > and the result is 404, file not found.  What is the proper procedure.  Email 
> > the author?
> > Cheers
> > Allister
> 
> I think it's too early to begin such things, we have to concentrate on
> important fixes.
> 
> I openned this page
> 
> http://wiki.tldp.org/NeedAction
> 
> to take note.
> 
> I hope soon we will be able to move to the wiki any such problematic HOWTO
You'll likely need to move most HOWTOs there.
> 
> But we have yet to exactly list all the HOWTOs (some exists only in
> some form, text...). The Page_Status being the central point for this
> aspect

In my own HOWTOs that I formerly updated every several months, I've had
on the average a few broken links per HOWTO.  Thus since other HOWTOs
are not updated very frequently, I suspect that the majority of them
have broken links.

So what is needed is a volunteer to run wget (likely via a shell
script) on all HOWTO's periodically and get a list of broken links for
each HOWTO and then email them to the appropriate author.  Actually,
authors should do this frequently, but I don't think many do,
including myself.  Fixing a broken link is sometimes easy if wget
shows a redirect.  It also can be difficult or impossible when the
link has disappeared without any replacement found using Google.

I think the typical HOWTO has broken links (except where the HOWTO has
almost no links).  

One common "broken" link is linuxdoc.org.  Well, it's not really
broken but it points to the wrong place.  Our old domain has been
taken over by someone and made into a commercial site.  We should have
asked Poet to give linuxdoc.org to us, but I guess he sold his
business and whoever got the business may have sold this domain.  So
what we feared actually happened.  However, whoever got it never
offered to sell it back to us.  Thus the task of fixing these
erroneous linuxdoc.org links in HOWTOs is important and one could do
this without getting the authors' permissions, although perhaps not
legally.  Free licenses usually require that statements about
modifications be included in the HOWTO.  But it would be a waste to do
this for such a simple change, especially for HOWTOs that have a
license permitting modification.

If we change them (all or some), it will hurt the traffic to
linuxdoc.org, and affect their business.  But it's justice to have the
link point to where it was intended to point.  Can a third party, the
owner of our old domain, linuxdoc.org, sue for violation of licenses
when this third party is not the copyright owner?

Note that Alta Vista reports almost 5,000 links to linuxdoc.org.  but
I've got them beat at 37,000 links to my email address, ####@####.####
which is why so much spam is sent to me (my ISP filters most of it out
and I try to get the accounts cancelled of those whose spam emails
actually reach me, via use of a form letter).  But due to my email
being widely available, I've also gotten many emails that are helpful,
usually either complementing me, or offering constructive criticism
and often pointing out errors or typos I've made.

I've quickly looked over the linuxdoc.org site and it's quite poor.
It has a section "Download free software".  I think it's the result
of a Google search or the like and it mostly finds free software for
Windows.  It doesn't seem to even have a link to our site.  So I think
that while it may have a lot of hits, most people will just realized
that it's not LDP and exit the site.  It seems to have few links to
Linux documentation.  Actually it's not at linuxdoc.org but
linuxod.org has been set up to redirect to another site that seems to
be partially about Linux.  One ploy is that companies may pay to put a
link on this site since Google, etc. will give the site a high ranking
since there are so many links to it.  Thus links on this site get a
higher ranking.  People have offered my cash to do this (put a link to
them) on my website and I've refused since my site is non-commercial.

			David Lawyer

Quoting David Lawyer ####@####.####

> So what is needed is a volunteer to run wget (likely via a shell
> script) on all HOWTO's periodically and get a list of broken links for
> each HOWTO and then email them to the appropriate author.  Actually,
> authors should do this frequently, but I don't think many do,
> including myself.  Fixing a broken link is sometimes easy if wget
> shows a redirect.  It also can be difficult or impossible when the
> link has disappeared without any replacement found using Google.

This means, in particular, that the sooner a broken link gets spotted,
the more likely you are to find the linked page/image at a replacement
URL, for a number of reasons including people's tendency to leave HTTP
301 redirects in place only for ~6 mos. to a year, if that, so you often
have a limited time window in which to find where things moved to,
subsequent to which knowledge acquisition, you have at least three
places you might find it:  the Internet Archive mirror copy of the
original location, the replacement location, and the Internet Archive
mirror of that second location.

> One common "broken" link is linuxdoc.org.  Well, it's not really
> broken but it points to the wrong place.  Our old domain has been
> taken over by someone and made into a commercial site.  We should have
> asked Poet to give linuxdoc.org to us, but I guess he sold his
> business and whoever got the business may have sold this domain.

No, I strongly suspect that Poet simply let it expire without asking,
for example, LDP if we'd like it back.  At that point, a domain
speculator re-registered it via a domain-acquisition and parking script
that deals in such domain properties in bulk (and in large numbers)
without human involvement at any point whatsoever.

How can I tell that?  First, have a look at the present Web contents:
It's one of those low-grade, low-budget "search engine" setups beloved
of domain squatters (er, "speculators") everywhere.  They don't
seriously expect anyone to actually appreciate that feature; it's
strictly there to fool unwary people who've followed old, obsolete links
into spending a few minutes on the site, which drives up their
ad-impressions count, which generally nets out to more than their US $7
annual renewal cost.  _Or_ it increases the traffic stats that they can 
show to potential domain buyers, driving up the auction price they can
charge in selling it off.

Ergo, if you want that domain to eventually get dropped so LDP can
re-register it, the only winning strategy is to seek out _all_ external
hyperlinks that drive traffic to "linuxdoc.org", and pester the various
site maintainers into either eliminating the hyperlink entirely or
substituting a suitable tldp.org link.  "Parked" domains that cease 
making speculators more annual revenue than they cost to keep "parked"
get dropped.


> However, whoever got it never offered to sell it back to us.

I surmise that it's a domain-speculation subsidiary of Dotster, Inc.,
called "Domainbank".  I doubt that any human being, there, has even
looked into who wants the domain at all:  Domainbank probably has
scripts that analyse expiring domains for revenue prospects, attempt to
grab the ones that look promising, and put up fairly generic "parked
domain" so-called search engines on them without any manual labour
whatsoever.  Imagine a firm that does this with tens of thousands of
domains all that time, and that's very likely the sort of firm we're
talking about:  They make very little on each typical speculation domain
"property", but they make it up in volume.

> Thus the task of fixing these erroneous linuxdoc.org links in HOWTOs
> is important....

Agreed.  Because that (along with hunting down and pestering out of
existence hyperlinks to that site elsewhere) is how we would change the
economics experienced by Domainbank (or whichever squatter, er...
speculator it is), making it earn less than it costs.

Any domain with significant "google juice" will stay locked up forever, 
because the scripts will keep renewing.  Change the economics, and the
script will branch the _other_ way, and let it expire.


> ...and one could do this without getting the authors' permissions,
> although perhaps not legally.

I would favour this.  Again, who on earth would haul you into court and
say "You committed a civil wrong against me by making me look more
competent than I really am.  I demand damages for the... um... loss to
my reputation", hmmm?


> If we change them (all or some), it will hurt the traffic to
> linuxdoc.org, and affect their business.

Right.

> Can a third party, the owner of our old domain, linuxdoc.org, sue for
> violation of licenses when this third party is not the copyright
> owner?

No.  To answer that for yourself, consider whether any obligation was
failed, or any legal right was infringed.

> I've quickly looked over the linuxdoc.org site and it's quite poor.

It's a "parked domain" robot page, doing really low-quality associative
matches of guesstimated fields of interest.  I'll bet you good money
that it's 100% autogenerated by a generic site-parking script, and no
human at the owning firm has ever dealt with it except as a metric item
on a spreadsheet along with many thousands of others.

On Thu, Nov 13, 2008 at 11:44:51PM -0800, Rick Moen wrote:

> Quoting David Lawyer ####@####.####
> > One common "broken" link is linuxdoc.org.  Well, it's not really
> > broken but it points to the wrong place.  Our old domain has been
> > taken over by someone and made into a commercial site.  We should have
> > asked Poet to give linuxdoc.org to us, but I guess he sold his
> > business and whoever got the business may have sold this domain.
> 
> No, I strongly suspect that Poet simply let it expire without asking,
> for example, LDP if we'd like it back.  At that point, a domain
> speculator re-registered it via a domain-acquisition and parking script
> that deals in such domain properties in bulk (and in large numbers)
> without human involvement at any point whatsoever.

One approach would be to try to contact the speculator.  Explain the
situation: that we are non-profit and work for free, that we would
like our old domain back for a nominal fee (like say $50), and that
if he refuses our offer, then we will make an effort to change the
links that currently point to linuxdoc.org.

But the speculator might think that if he responds there might be some
chance that we might sue him and thus he might not respond at all.
linuxdoc.org is printed in a number of older books and it would be
nice if we could get back the domain. But as time goes on, it's
importance decreases as old docs that have it disappear from the
Internet.

			David Lawyer

I would say: automatically replace the old domain in all documents
with the working one
so that the links will be restored seamlessly. If an author has
objections, work them out
with the author case-by-case.

Make a web page listing the changed  files and the details of the change.
Drop a link to that web page at the footer of every document.


So long as the 'fixed' document preserves what the author intended to
link to, I.E. the link
still points to exactly the same thing as when the document was
written, it seems
like no real modification has taken place.

It is still the same link to the same thing.   It is debatable whether
the work has changed at all.

The actual text of the links included as references is not part of the
copyright work,
in fact, they're provided by a third party; it's the choice of which
links to include which
matters.


Otherwise many web browsers break copyright when a person  'saves' a
page and shares
with their friends;  since browsers rewrite image URLs to be the local
saved files (for example)





> If we change them (all or some), it will hurt the traffic to
> linuxdoc.org, and affect their business.  But it's justice to have the
> link point to where it was intended to point.  Can a third party, the
> owner of our old domain, linuxdoc.org, sue for violation of licenses
> when this third party is not the copyright owner?

A license is between the owner of a copyright's exclusive rights and a party
exercising rights authorized by the copyright holder.

Unless the copyright owner transferred some of their exclusive rights or named
the third party their agent:  a third party has no right to sue on the basis of
copyright infringement,  since they don't own any exclusive rights.

They might have some other basis, like unauthorized linking,
anti-competitive practice,
discrimination  (for removing links to them), etc, but it seems unlikely.


-- 
-J

Quoting David Lawyer ####@####.####

> One approach would be to try to contact the speculator.  Explain the
> situation: that we are non-profit and work for free, that we would
> like our old domain back for a nominal fee (like say $50), and that
> if he refuses our offer, then we will make an effort to change the
> links that currently point to linuxdoc.org.

The downside to that idea is that, if they're not generous to professed
non-profit groups (and I think a moment's reflection would make you
realise why they wouldn't be), all you've done is call their interest to
others' desire for a domain that otherwise might be let drop when it
fails to pass the automated earnings tests.  In that scenario, all
you've done is deferred the domain's availability and raised its price.

> But the speculator might think that if he responds there might be some
> chance that we might sue him and thus he might not respond at all.

The only legal threat a professional domain-speculating company is
likely to take seriously is one announced by a credible trademark
complaint.

Quoting James Hess ####@####.####

> I would say: automatically replace the old domain in all documents
> with the working one so that the links will be restored seamlessly. If
> an author has objections, work them out with the author case-by-case.

I concur.

It seems astronomically unlikely that an author would complain; if
he/she does complain, there are no "actual damages" to be recouped, 
and the worst-case scenario, realistically, is that LDP either reverts
the change or withdraws the HOWTO.

> So long as the 'fixed' document preserves what the author intended to
> link to, I.E. the link still points to exactly the same thing as when
> the document was written, it seems like no real modification has taken
> place.

At the risk of pedantry:  It _is_ a real modification.  And it is
technical copyright violation if not permitted by licence.   It's a
particularly harmless variety of violation, though -- in the sense that
even an unreasonably annoyed author would have both literally nothing to
gain by suing and no damages.

> It is still the same link to the same thing.   It is debatable whether
> the work has changed at all.

I hate to sound argumentative, but, sorry, it's changed.

> The actual text of the links included as references is not part of the
> copyright work, in fact, they're provided by a third party; it's the
> choice of which links to include which matters.

Sorry to say, HOWTOs are provided with links; they're not "provided by a
third party".  The fact that the content available when a Web browser
dereferences those hyperlinks is "provided by a third party" is
irrelevant to the question of what comprises the creative work to which
copyright attaches.

See, computerists think they're really clever when they come up with
arguments like yours, e.g., that the URL is not really specifying
anything in particular at all, but rather is merely a pointer that gets
dereferenced by a third party; ergo, the URL can be changed without
changing the copyrighted work.

Absolutely dazzling argumentation -- and, if it were possible to dazzle
judges with bullshit, it might be a winning ploy.  I'd strongly advise
you not to try that, though.  The judge is going to say "The published
document had URL foo.  You changed it to URL bar -- without permission. 
That's copyright violation.  Your argument is wrong." 


> Otherwise many web browsers break copyright when a person  'saves' a
> page and shares with their friends;  since browsers rewrite image URLs
> to be the local saved files (for example)

Again, nice try on the creative rhetoric.  It's possible that you missed
the earlier points about implied licensing.