Concerning datasheets

Wolfgang Spraul wolfgang at sharism.cc
Sun Dec 6 02:46:16 EST 2009


Kristoffer,

> Oki, so its generally a "please dont hand out these documents unless someone requests them".

Yes.

> I have no intention on giving qi-hardware/ingenic any problems. I will
> put a notice instead and say that datasheets are available on request.
> 
> Something like "Datasheets are available if you send a request to email bla at bla.com"?

Perfect. You can give my email address if you like.

> Well, I can certainly accept a "silent" agreement, in the sense that
> its okey to give away the datasheets but they/you dont activly push
> them onto people (people have to ask first). 

Yes. (phew, same thing 100 times I feel :-))

> I also hope that they will be completely open with the datasheets eventually (published on site).

Unfortunately I doubt that.
Now let me give you some background information on this whole subject.
I personally have been doing endless lobbying towards Creative Commons
licensed datasheets, before at my work at Openmoko, and now at Qi.
At Ingenic, I had at least 2 or 3 substantial meetings, with prepared slides,
lots of people attending, etc. etc.
The problem are lawsuits, especially patent litigation. Lawyers are most used
to dealing with documents. If they have a nice PDF file, with a company
address on the header page, and then some statements (text) on following
pages, they are in business.
It is much harder to base patent or liability litigation on source codes,
maybe even developed by multiple authors in a 'chaotic' (for the lawyers)
way.

So unfortunately companies are becoming MORE concerned about publishing
official documents. I heard about a case where nvidia had to pay a 7-digit
USD amount to a patent troll for something they only _claimed_ in their
documentation, even though the actual chip didn't even support that feature :-)
Think about this for a moment and you understand what I mean.

So in the industry, datasheets become either exchanged via NDA (the standard
way), or just 'under the table without NDA' (the Chinese way).
One example about Samsung: The 2 GB NAND chip in our Ben NanoNote has an open
datasheet, but the datasheet for the 16 GB NAND chip, which is easily available
and already being tested by us, is closed right now - we got it 'under the
table' after some searching :-)
Should we not work with this chip until we get an open datasheet? If we start
now, can we trust Samsung to eventually publish the datasheet?
It's a mess!

In the long run, it's clear to me that we need to make our own copyleft chips,
fully GPL'ed, with all tools needed for edits of complex data also free, etc.
But for a number of reasons this makes no sense for the next few years at least.

In Ingenic's case, my task is to avoid sending the PDF files to lawyers.
Coders are OK. You go figure ;-)
And believe it or not, I did already get some suspicious mails from people I
had never seen before, and after asking the slightest tech question back
their need for Ingenic datasheets suddenly disappeared.
So if possible I would hand over that responsibility to you as well, if you
pass those datasheets on. If we fail, and feed a patent troll, we will not
get further documentation.
In parallel I suggest we document source codes lavishly. Feel free to copy
and paste from the documentation.

Until the day we have our own deeply GPL'ed chips, manufactured by multiple
independent foundries, I'm afraid we will have to live with this unfortunate
blurred situation, datasheets 'upon request', sometimes 'under the table',
sometimes only under NDA, etc.

> Thanks for clearing it up.

At least I told you my thinking. Others may disagree...
Any more questions please let me know. This is a very important subject IMHO.
Wolfgang

On Sat, Dec 05, 2009 at 05:24:54PM +0100, Kristoffer Ericson wrote:
> On Sat, 5 Dec 2009 16:55:15 +0100
> Bas Wijnen <wijnen at debian.org> wrote:
> 
> > Hi Kristoffer,
> > 
> > Not that I'm in a position to answer your question, but I'll do it
> > anyway.  Wolfgang can correct me if I misunderstood. :-)
> 
> Shoot :)
> 
> > 
> > On Sat, Dec 05, 2009 at 03:25:38PM +0100, Kristoffer Ericson wrote:
> > > From what I understand you (qi-hardware) are not permitted to hand out
> > > datasheets unless asked?
> > 
> > There is no agreement.  Legally, they got the datasheets, nothing was
> > signed, and they cannot be expected to treat them in any special way.
> > They could publish them on their website.
> > 
> > However, Ingenic asked them not to do that.  I have no idea why.  But in
> > order to keep a good relationship with Ingenic, they don't do it.  And
> > they ask anyone they send them to to do the same.  There is no agreement
> > there either.  Wolfgang sent them to me, and I would be allowed to
> > publish them anywhere.  However, this would likely damage QI's
> > relationship with Ingenic, which would be a pity.  So I don't publish
> > them.
> > 
> 
> Oki, so its generally a "please dont hand out these documents unless someone requests them".
> 
> > > So does this mean that Im unable to host the same datasheets (supplied
> > > by qi-hardware) in any public environment?
> > 
> > If you're a nice person, then yes. ;-)  However, I would not have a
> > problem with a semi-public environment, such as the intranet of an
> > organization.  I would add a statement about their use though,
> > explaining that it would be appriciated if they weren't published.
> 
> I have no intention on giving qi-hardware/ingenic any problems. I will
> put a notice instead and say that datasheets are available on request.
> 
> Something like "Datasheets are available if you send a request to email bla at bla.com"?
> 
> > 
> > > The reason Im asking is because the jlime machines that we support we
> > > also have sections in our wiki where hardware documentation resides.
> > > Would be nice to keep copies there for people interested (a few at the
> > > most).
> > 
> > You can let them ask and send them on request.  You may certainly
> > redistribute them, it is only asked that you don't make them available
> > for random people who are clicking random links, I think.
> > 
> 
> Most of the datasheets that jlime post are old enough to be of little interest
> of companies, So I dont hesitate to post datasheets of those hardwares. Since
> the nanonote is new though, I thought it best to ask so I have the picture clear.
> 
> > > Would that be permitted or does the agreement you have also affect the
> > > receivers of the datasheets.
> > 
> > If there would be an agreement, it would probably forbid this, and that
> > would make the agreement unacceptable to people who like free hardware.
> > Now it's so vague that we (or I at least) don't know what to think about
> > it. ;-)
> 
> Well, I can certainly accept a "silent" agreement, in the sense that
> its okey to give away the datasheets but they/you dont activly push
> them onto people (people have to ask first). 
> > 
> > However, looking at how things are going, I would say that if we show
> > that having those datasheets available actually generates things that
> > Ingenic wants to have (such as better support for Linux), they are
> > likely to continue further on the path to openness (making the
> > datasheets really open, and opening more).  In other words, send the
> > datasheets to anyone interested and encourage them to do some work. :-)
> 
> I also hope that they will be completely open with the datasheets eventually (published on site).
> 
> Thanks for clearing it up.
> 
> 
> > 
> > Thanks,
> > Bas
> 
> 
> -- 
> Kristoffer Ericson <kristoffer.ericson at gmail.com>
> 
> _______________________________________________
> Qi Developer Mailing List
> Mail to list (members only): developer at lists.qi-hardware.com
> Subscribe or Unsubscribe: http://lists.qi-hardware.com/cgi-bin/mailman/listinfo/developer




More information about the discussion mailing list


interactive