[Openicc] ColorVision Open Source Policy

[Hal V. Engel]

On Monday 12 November 2007 19:02:34 Robert Krawitz wrote:
>    From: "Hal V. Engel" <hvengel at astound.net>
>    Date: Mon, 12 Nov 2007 17:21:07 -0800
>
>    In a follow up from C. David Tobie of ColorVision he wrote:
>
>    "We have supported Linux for years, for specialty clients such as
>    Disney. We continue to consider the possibility of a display
>    calibration solution for end users on Linux. But no one on
>    Linux/Unix,BSD ever contacts us and asks about products they can
>    purchase from us for those platforms, they always ask if we have an
>    SDK, so they can write something themselves. Its a very different
>    type of market.."
>
> The fact that he says people are asking for an SDK suggests to me just
> how wide the gulf is.  I think it's much more likely that people are
> asking for the programming specs for the hardware, not an SDK.  A
> proprietary interface to a binary-only library with functionality
> restricted to what ColorVision thinks is important isn't likely to be
> too interesting to the kind of hard core *IX people who want this kind
> of device.

In fact in my request to them I stated that I suspected that their SDK was of 
little use to me since I thought that it was very likely that it did not 
support the platforms I needed and that what I needed were device interface 
specs.  As you can see he ignored that part of the request like it never 
existed.

On the other hand I suspect that the functionality supported in the vendors 
driver/library would be enough to do about anything that is needed in this 
case.  The real issue is that the driver is likely never going to be 
available for the platforms that we support not that the driver/library is 
crippled or has limited functionality.  After all they use this same 
driver/library to support their own software and at least the higher end 
versions of their software are very feature rich.

snip
> Some proprietary vendors understandably latch on 
> to this, but in the long run it doesn't help anyone -- it just means a
> proprietary Linux solution that's usually inferior to the proprietary
> Windows or Macintosh solution.

I can only think of one proprietary driver where the Linux/X11 version comes 
close to the Windows and OS/X version in performance and that is the Nvidia 
graphics driver which has very good OpenGL performance.  Even at that it is 
missing many of the features that are available on Windows and OS/X.  In 
addition,  it does have lots of problems such as being limited to x86 and 
x86_64 machines and stability issues that would normally not exist in an open 
source video driver.  Every other proprietary device driver that I can think 
of for X11 or Linux has a broad range of significant issues and falls far 
short of it's Windows and OS/X counterpart in every way.

In general I stay away from proprietary drivers and on my current machine the 
only exception is that I am currently using the nvidia driver.  But I am 
rooting for the Nouveau project and perhaps I can remove the Nvidia driver at 
some point.  I also think the recent move by AMD/ATI will have an impact in 
this area and may cause Nvidia to start opening up at some point since they 
are the only remaining video card vendor that is not supporting an open 
solution.

>
> The same thing happens with printers (my area of expertise -- I'm the
> project lead for Gutenprint).  Epson does provide programming manuals
> for many of their printers.  These manuals aren't great, and they're
> missing some things, but they do make it possible to write drivers.
> Unfortunately, they've stopped providing manuals for the Stylus Pro
> printers, and I'd really like to support those printers in Gutenprint.
> HP provides open source drivers for a lot of printers, but some of
> them are limited in functionality (particularly in the color arena)
> and they offer a lot fewer options than Gutenprint.  Both of these are
> light years better than any of the other inkjet or dye sub vendors,
> who provide nothing at all.

There is someone on this list who might be able to help with one of the "light 
years" behind printer vendors.  Perhaps that person could contact Robert off 
list?

snip
>
> Again, there are enough self-proclaimed leaders in the community who
> are willing to live with this situation that it's easy for a company
> with proprietary leanings to miss the other side of it.

But in this case the vendor is not even providing closed drivers of any sort 
(good or bad) for anything other than Windows and OS/X and they will not even 
provide the API documentation for the drivers if this is going to be used by 
an open source project even on one of the supported platforms.  So this is 
one case where even those "self-proclaimed leaders in the community" you 
refere to have nothing to "live with".

snip

> These are selling on eBay for around $50 plus shipping.
>
> Not to mention that if we could get it working we could almost surely
> write better software for it.

That is already the case.  The ArgyllCMS meter code does have preliminary 
support for the Huey but I have not been able to get it to work on my machine 
on either Windows or Linux.   But the code is so new that I am not surprised 
that there are glitches at this point.  In any case once these issues are 
ironed out it will be supported by both ArgyllCMS and LProf and both apps 
will allow you to do many things with it that can not be done with the 
vendors software such as calibration and profiling of more than one display 
and making "pre-calibration" adjustments.  And of course you can do these 
things on Linux, BSD and Unix X11 systems as well as (the vendor supported) 
Windows and OS/X.  

The only feature of the vendor software that is not supported is the ambient 
light adjustment feature and I would rather that they removed the extra 
sensor from the device and charged less money since I think this feature is a 
marketing ploy rather than something that is really useful.

Hal