[sane-devel] Hardware compatibility ongoing problem and what can be done about it
simon.matter at invoca.ch
Fri Jun 18 15:26:00 UTC 2010
> On Fri, Jun 18, 2010 at 1:07 AM, augustin <sane at overshoot.tv> wrote:
>> I believe in wallet advocacy: if we only had a reliable and objective
>> assessment of the linux-(un)friendliness of hardware manufacturers,
>> users across the globe could be encouraged to buy from certain
>> rather than from others. To be effective, this should be a concerted
>> based on facts rather than hearsay (the kind that is spread in forums
>> and user
>> mailing lists by people who don't really know).
> Support from a maker is a bitmap:
> 6: Provides FOSS drivers (or pays FOSS devels to write)
> 5: Provides Unencumbered Documentation
> 4: Gives hardware to outside FOSS developers
> 3: Loans hardware to outside FOSS developers
> 2: Provides Encumbered (NDA) Documentation
> 1: Provides proprietary drivers for one or more platforms
> Different makers actually have different divisions, which might do
> different things: Buying an expensive scanner will get you loaner
> hardware and docs. Buying a cheap scanner will get you ignored, cause
> you are talking to two different groups.
> I will only speak to the companies I have worked with:
> 1. Fujitsu USA paid me to add some enhancements to the sane-fujitsu
> driver, and gave me hardware and docs. I even spoke with a couple of
> engineers over the years. However, they were unable to help when I
> asked for docs about the Epson-based (epjitsu) small machines. Fujitsu
> bitmap: 111110, epjitsu bitmap: 000000. Stark difference, no?
> 2. Canon USA loaned me hardware so I could reverse engineer their DR
> series machines, even though they were holding the docs I needed.
> After I developed the driver, they released a competing closed driver
> that covers fewer models and runs on fewer platforms :( DR series
> bitmap: 000101.
> Canon of India has recently provided funds and documentation for us to
> improve support for their least expensive model: 110000. However,
> IIRC, 'sane-pixma' driver was completely reverse engineered: 000000.
> See- Three completely different responses from the same 'company'.
> 3. Kodak loaned me a very, very expensive scanner for over a year to
> write the sane-kodak driver. They gave me docs, access to engineers,
> and worked one of my suggestions back into their firmware! bitmap:
> But then, they released a closed driver for a few of their smaller
> machines, which only runs on a few platforms, and tried to make it
> sound open by word-smithing. 000001.
> The take-home here is that companies are people too. The cannot be
> rated or measured to an objective standard, and they cannot be
> expected to react logically, even when presented with a problem with a
> single obvious solution :) Perhaps I have become jaded, but there is
> not enough buying power in the Linux community to make a blip in the
> average scanner maker's sales numbers. It might be a fine way to
> reinforce existing behaviour of good makers. If you want to change the
> bad ones, better luck will be had with convincing a human with some
> power. People make the decisions, not spreadsheets.
What the community could to is recommend more to others what works well
and what does not. I remember the time we were looking for decent scanners
for our inhouse document management system and how difficult it seemed to
make a choice. Fortunately I've got some recommendations from you and
others and I learned about the "bitmap" you wrote above and once that was
known the decision was easy. We got some of the Fujitsu devices and it
looks just perfect. So now whenever we need a new device we don't think a
second and all other vendors are out of the game.
More information about the sane-devel