[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [off topic] Re: Licensing issues

>>>>> "T" == Terry Dawson <terry@albert.animats.net> writes:

    T> It was a move to more freedom, the only change was that it
    T> allowed a use that wasn't allowed before. I don't see how any
    T> reasonable description of freedom, or more or less suggests
    T> that is a restriction of freedom.

The LGPL removed all those 'derived' works from being bound by the
GPL.  Thus, it had the result of encouraging developers to work on
non-free software.  So in this sense, no, it did not restrict
freedoms, it just encouraged less free software development. At least,
that was the arguement at the time.  In retrospect, GCC did not
become the industrial standard C compiler until _after_ the LGPL.

Similarly, Linux distros who include non-free software are giving the
message that non-free is required to complete the picture, and also
removing the need to develop good free software.  I don't entirely
agree with this scenario, but this is my reading of the RMS papers on
gnu.org and our correspondence.  His argument has merit, but I don't
think it is practical as dogma even when there are no viable free
alternatives.  The same is true for docs.  Non or partially free
licences are a transitional stage.

    T> The GPL will and does work for documents in exactly the same
    T> way. If you want to take portions of any of my documents and
    T> combine them with portions of other documents the other
    T> documents have to have licenses that are GPL compatible. 

If software is derived from GPL software, that software must be GPL
and while it can be more free, it cannot be less free.  

If, on the other hand, research for a book is based on LDP documents
and makes copious quotes, correcting commentary and grammar, and then
making logical references to it to support a new point not found in
the original (ie function call equivalents, corrected functional
definitions, and new higher-level functions), then, by the GPL, 
can you sell that book?  

If a study course (an application) is based on a document, and the
professor likewise corrects and extends the material, can you charge a fee
for that course under the GPL?  

If Eric Raymond speaks to a room full of engineers using modified
ideas of the GPL licenced documents, can he charge for the appearance?
Can he charge for a book containing his talk?  The whole notion of
"derived/extended" and mere use (ie LGPL) in natural language is very,
very murky.

What if someone were to take an LDP document and transform it into a
politically slanted work in conflict with the aims of the LDP, say,
for example, to change it to highlight the shortcomings of some core
Linux system such as KPPP? It could become a Microsoft sales tool.
Because software has no 'implied' meanings or grades of truth, such a
derived work could be attributed to the LDP under the GPL.  

For the same reason Salvador Dali threw a bathtub through Macy's
window, authors seek control over changes of content, and some control
over uses, and these are issues which do not occur in software (where
the only measure of correctness is "bug-free")

    >> But only the restriction on book-publishing, while opening up
    >> the free distribution of all material for all other avenues.

    T> They're already open. There is no gain there.

I see this as the first problem of the LDP: How do we stay current?

Docs are not open if there are none.  Unlike software, we don't have a
pressing need for correct docs; an engineer is not stopped in their
tracks if the SMP-Howto is not correct, but they are stopped if the
SMP implementation is buggy.  There is an itch to drive the creation
of software because the use of that software can be translated
directly into a business model to pay for the investment.  There is no
such carrot on a stick for documents: Far more than software patches,
LDP docs are a total selfless gift from the author.

Some of my own (non-Linux) documents (now published under the OPL)
were written under contract for other purposes (teaching of courses,
whitepapers for clients) and thus, because I had recouped the cost of
producting these papers, I had no problems donating them for any use
whatsoever.  I was tickled pink when I learned two universities had
printed them as handouts for webdesign and internet courses.  The LDP
does get some docs this way, from people like RedHat and other
engineers who had to document certain systems for their design teams
(thus recouping costs) but the vast majority of our Howtos are just
friendly gifts and anecdotal personal experiences.

What I want is a model whereby the people who do have an itch to
scratch by having good docs (the publishers) can fold all or parts of
that effort back into the LDP for the benefit of those who cannot
afford or otherwise have no access to the books (or even just so they
can properly evaluate a book)

    T> The donating whole works thing is a whole different subject. If
    T> a publisher, as copyright holder of a work, wants to provide
    T> their work to the LDP with a license that allows free online
    T> publishing but restricted paper publishing then that is
    T> entirely their perogative. I would welcome that, but I would
    T> oppose any move to attempt to make any other LDP documents
    T> conform to that style of license.

Yes, exactly. Perhaps we have been talking about the same thing from
the same side but using different words.  I am not advocating slapping
an OPL Option-B on _all_ LDP documents. When I mentioned corporate
adoption of documents, I meant *big* documents, those of whole book

If, say, ORA were to adopt the SAG and make it into a first class
book, cool diagrams, accomodating all the major distros, and their
rule was they would let us ship it out on the web, ftp and CDROM but
we are prevented from printing, binding and marketing it (all three
together; just printing for person or course-study use is still
allowed), then is there a problem?

    T> The LDP has to make a decision as to what constitutes an
    T> acceptable document for inclusion within its charter.

Yes, exactly.  Let me be more direct: I am working with a dozen
authors to create what I see as the missing functional specifications
document for the Linux kernel.  I nagged and nagged and nagged MCP
until they agreed to let me publish under OPL Option-B, but now I am
worried this will be for naught, and since the LDP may not accept the
document --- I am also finding many very good authors reluctant to
participate for fear of LDP reprisals.

    T> I think all you're really trying to say is that you believe the
    T> LDP should leave enough flexibility in its minimum licensing
    T> standards to accomodate the model you describe.

Yes.  If Lucent wants to donate a binary kernel module for their
winmodem, do we refuse it?  It's a similar situation only they are not
saying we cannot modify the content to improve it, only that they are
the official maintainer until they say otherwise (which is true of all
LDP submissions) and that they are restricting the document from being
used against them in the marketplace.

I believe the LDP should encourage the free flow of information and
accomodate Linux education.  When a sprout meets a stone, that is only
one of many degrees of freedom lost, and I see the interim solution of
New Riders, ORA, Sams, Coriolis and others using Option-B to protect
their investment as only the loss of one degree out of many.  I
believe this clause only exists to frustrate other publishers, ie,
like the reason Sun and HP &c endorsed open source in X11: To protect
them against their own kind.

It does not seem likely to me that your typical LDP users would *want*
to snarf an entire book and market it in direct competition with an
ORA edition.  I admit it is a loss of a single freedom, but it doesn't
seem to me to be a freedom anyone would miss.

    T> Let me ask you a question. It's fairly obvious what the
    T> advantages are to the publisher in having a document branded an
    T> official LDP work and having exclusive publishing rights, but
    T> what's in it for the LDP or the greater Linux community to have
    T> a document branded an official LDP document but with
    T> restrictions placed on its publishing in physical form?

I see several advantages for the LDP, and huge advantages to the
Linux community.  Of course, I tend to see lots of things and that
is probably due to all those hundreds of acid hits in my youth.

An immediate advantage is having professionally crafted and current
material in our body of work.  When someone searches the LDP, they
find these documents, and hopefully find the answer to their question.
In the Windows world, the 'Knowledge Base' provides a one-stop shop
for end-user information (ie, if the KB doesn't have it, no one will,
even though the latter condition is the most common ;) Down the road,
when all docs are XML and we have the tools to deal with it, one doc
can lead to other related docs and even small fragments can be
instantly updated across the board --- it will become unthinkable that
a question about Linux cannot be answered by the LDP.

Another advantage is in having readable docs.  There are notable
exceptions, but as a rule, our current opus is very badly written and
naively organized ... but we never ever complain because they are
gifts and we can usually figure out what they meant to say.  Not that
books are guaranteed to be better ;) but they do have teams of people
whose full-time profession is to ensure documents are accurate and
readable, ie tech reviewers, language reviewers, translators ... all
of their skills are in producing documentation.  We can leverage those
skills for the benefit of the Linux community.

The long term advantage is relationship building.  If we can bind
together all the publishers such that they lose their fear of each
other and can work together, we can eventually apply their talents to
all of the LDP docs, and without the Option B restriction because they
will see the integrity of our collection as their best resource ---
right now, the LDP is what sells every publisher's books precisely
because of its obtuseness.

    T> The ability to read the document online but not to be able to
    T> print it?

Option-B only removes the right to bind it into a book for resale.
That is very different from banning all printouts.

OPL allows a teacher to print out sections for use in a course, or to
include it on RH, LSL or Infomagic CDs, or to ship it to other mirror
sites.  Option B is only to protect them from having their investment
scooped and leveraged against them in the marketplace by other

    T> It seems like a fairly one-sided arrangement to me.

Let me play Devil's Advocate:

You could just as well say publishers are giving away everything and
getting nothing they couldn't already take for free.  I was greeted
with polite shock when I contacted the LDP to tell them we were
including an entire doc in one of our books; yes, I'd read and
understood the licence, but it seemed to me the polite thing to do.
All the LDP really offers is a brand name, and that name can be
replaced by another (anyone remember Mosaic? Simtel? Decwrl?)

And some, like ORA, already have a very good name. Why should they
bother with the LDP?  Why burn their profits just so the LDP can have
some nice timely doc updates?  Why should Corliolis fund revisions for
your benefit?  Each publisher probably puts out more lines of text per
year than the entire LDP.

I can't speak for any of them, but I can relay the exact words of the
senior publisher for the Waite Group: 

   "I could publish 'cooking with Linux and it would sell', but I
    don't want to do that.  I want to do it right"

It is the cause of righteousness, the infectious community-support
aspect of open source that attracts them --- that they are attracted
to the LDP is just a lucky co-incidence for us (if we choose to seize
it).  They see that they take so much from the community, they see
that this community does not even require them to say 'thanks' and
they want to give something back; they look to the LDP as their
closest brethren because we are all in the same line of work.  Are
they so different from the rest of us?

It all comes back to the mandate of the LDP.  If the mission of the
LDP is to foster only 100% free documentation, then no, there is no
advantage to any such partnership, and publishers and the LDP must
always be in direct competition with each other.

If, on the other hand, the LDP mission is to foster Linux education,
then I believe we have as much to gain as the Kernel had to gain from
corporate support from SGI and Compaq: We gain the mutually beneficial
pooling of our knowledge with the technical publishing and document
management skills we are missing, all working together to help make
the LDP into an even better resource for the Linux community.

Please keep in mind that I am not a publisher, nor do I speak for one.
I don't even pretend to be one on TV.  My vision of LDP/corporate
cooperation makes sense to me only under the banner of serving the
Linux user community interests; you only need to watch a day's worth
of comp.os.linux.setup to see how we fail to provide what the
community needs.  I still may be completely off the wall and the whole
notion of a master Linux documentation project may not be feasible,
practical, or even desirable, to either publishers or the LDP.

Gary Lawrence Murphy <garym@canada.com>  TeleDynamics Communications Inc
Business Telecom Services : Internet Consulting : http://www.teledyn.com
Linux Writers Workshop Archive: http://www.egroups.com/group/linux-hack/
"You don't play what you know; you play what you hear." -- (Miles Davis)

To UNSUBSCRIBE, email to ldp-discuss-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org