Re: Canoscan FB630U

From: volodya@mindspring.com
Date: Mon Mar 05 2001 - 08:21:08 PST

  • Next message: Petter Reinholdtsen: "Re: OT: Is "fileno()" conform to ANSI or POSIX?"

    I my experience talking with hardware manufacturers is useless.. Unless
    you don't need them. (But don't dismiss PI experience - somehow they got
    them to release the 3d part). I.e. first you write driver that works with
    your card. Then they might give you docs.

                             Vladimir Dergachev

    On Mon, 5 Mar 2001, Marian Eichholz wrote:

    > Simon Heywood wrote:
    >
    > > Today I telephoned Canon UK, and was informed that Canon are not going to
    > > release any technical specs at all, even with a NDA (what a surprise);
    >
    > ...Bummer. They are *really* unwilling to release information. Perhaps
    > they made a *very* bad- design and are really afraid, that people might
    > notice it.
    >
    > > however I was given the Managing Director's address for delivery of any
    > > letters I might care to write explaining the situation. Does anyone have a
    > > recommendation as to starting an online petition (and of course whether
    > > you think this would be appropriate).
    >
    > This is like my situation with Microtek. The engineer, who wrote to me,
    > mentioned an NDA but knew in advance, that I would not accept it.
    >
    > I tried/am trying to convince him, that I will publish a driver anyway,
    > regardless of all possible problems and insane design due to lack of
    > information, and that it will be the best for Microtek and us all, if
    > they support it by specification disclosure, rather than to slow the
    > progress down by obfuscation.
    >
    > I really hope, that Microtek is smart and honest and discloses the
    > scanner details. But I have no idea, if I used the right tone and
    > temperament. But what did we had to lose?
    >
    > NDAs are a real problem with the cheap "no api" solutions. The Microtek
    > scanner - for example - uses a register transfer model over USB, fairly
    > simple and straight. It saves the company lots of money, because they
    > don't have to build a hardware abstraction (in terms of SCSI2 or so),
    > but they have no chance to hide the device internals and I.P. in this
    > model.
    >
    > > Canon claim they are "looking into" producing a driver package for Linux.
    > > However, they were unable to tell me whether or not this would work in
    > > conjunction with SANE, or in fact conform to any established Linux
    > > standards.
    >
    > Drivers by a vendor are good, but specification and/or drivers with
    > source and an open license are really better. Just take NVIDIA as an
    > example. It is nice, that they are building 3D-drivers. How nice it
    > would be, if they worked reliable. :-(((
    >
    > And there are "driver construction kids" for Linux out there, that
    > really breach *any* kernel mode protection. Nobody should give vendors a
    > chance to work with such blundering tools without beeing blamed for it.
    >
    > > In the meantime, I will investigate the possibilities offered by
    > > "sniffing" the comms protocol and/or looking at the
    > > <canonscanner.sourceforge.net> code. Time to read up on USB/SANE methinks.
    >
    > I wish You all the best with it!
    >
    > --
    > Mit freundlichen Gruessen / Yours sincerely
    >
    > Marian Eichholz (0211/53087-521)
    >
    > --
    > freenet.de AG Vorsitzender des Aufsichtsrates: Gerhard Schmid
    > Deelbögenkamp 4c Vorstand: Eckhard Spoerr (Vors.), Axel Krieger
    > 22297 Hamburg Amtsgericht Hamburg, HRB 74048
    >
    > --
    > Source code, list archive, and docs: http://www.mostang.com/sane/
    > To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
    >

    --
    Source code, list archive, and docs: http://www.mostang.com/sane/
    To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
    



    This archive was generated by hypermail 2b29 : Mon Mar 05 2001 - 08:05:53 PST