[Openicc] Drop size calibration

[Robert Krawitz]

   Date: Mon, 4 Feb 2008 14:15:51 +0100
   From: "edmund ronald" <edmundronald at gmail.com>

   As the years go by, more printers are supported by Gutenprint, and
   it gets better. In the same way, I think that if one known stable
   distribution is created where at least some decent binary drivers
   for current instruments are present, the various programmers and
   users of open source graphics software will progress towards a
   color managed state, which will then hopefully advance towards
   total openness.

There won't be too many open source programmers attracted to this kind
of platform.  In any case, since there are free drivers available for
a number of instruments, those will get used, and the proprietary ones
will be ignored (if there are good substitutes) or reverse engineered
(if they otherwise have useful characteristics).  Proprietary vendors
might not like that, but the truth of the matter is that the dynamic
behind the FOSS world is that someone who wants something badly enough
does it, and there are no managers who can put a stop to this kind of
effort to protect a strategic alliance.

		    Frankly, how can I trust the quality of the
   distributed software if the authors cannot run any
   ***instrumented*** tests to check the functionality from version to
   version ? Testing is the responsibility of the author, and
   therefore a reference platform which will support testing is
   needed.

You're using a particularly restrictive definition of "functionality"
here: no changes in output between releases.  There's nothing
fundamentally wrong with that definition, but there are plenty of
other definitions that authors can elect to use.  This kind of
definition might also be overly restrictive if it prevents us from
fixing egregious problems (like the drop sizes on the Claria printers,
or the light inks on the 2400).

This also doesn't require a reference platform or instrumented tests;
another way to accomplish the same thing would be to simply use an MD5
checksum (i. e. reference output).  We have that infrastructure in
Gutenprint right now and it's something I would consider using.  In
fact, relying on instrumented tests of this kind has problems of its
own: variations in external factors such as ink and paper over time,
printer head wear, noise in measurements, and so forth.

-- 
Robert Krawitz                                     <rlk at alum.mit.edu>

Tall Clubs International  --  http://www.tall.org/ or 1-888-IM-TALL-2
Member of the League for Programming Freedom -- mail lpf at uunet.uu.net
Project lead for Gutenprint   --    http://gimp-print.sourceforge.net

"Linux doesn't dictate how I work, I dictate how Linux works."
--Eric Crampton