[Openicc] X-Rite to aquire GreytagMacbeth

[Hal V. Engel]

On Wednesday 01 February 2006 04:54 pm, Graeme Gill wrote:

snip

> Interesting. I actually have access to a SpectroCam (Serial, not USB),
> but haven't felt it possible/worthwhile to support it in my software.
> If the drivers really were available as GPL (even if there was some work
> to get it going), and anyone was interested in actually using it
> (ie. that it wouldn't be an academic exercise), then it might be
> worth supporting.

I asked the author of the driver a few questions and he got back to me early 
this morning.

He said that the driver is in fact GPL.  He wrote:

"I thought I marked everything.  In any case, all of it (including
documentation)  is licensed under the GNU GPL (version 2 or later).

I know I didn't yet mark all the driver code in CVS as such, but for that it
is also true."

I should add that it appears that the author of this driver is a third party 
and that Avantes is hosting his package.   I also asked if it supports the 
AvaMouse/SpectroCam and he said:

"I don't think so.  However, you can always try. :-)

It supports 3 interfaces, serial port, usb, and emulation.  The latter is for
testing the program without a device, and due to the protocol's inability to
handle packet corruption, I'd advise to use the usb interface.

I don't know what interface the AvaMouse and SpectroCam have, but if it's usb
or serial port, it may very well work.  As far as I know, many of the
spectrometers have the same protocol."

He also stated that what is in CVS is much newer than what Avantes has on 
their web site.  CVS is now autotoolized so getting it to build should now be 
easier.  To access CVS:

If you want to get the latest version, you can use cvs.  For the library:
$ cvs -d:pserver:anonymous at pcbcn10.phys.rug.nl/var/lib/cvs co libshevek
and for the drivers (it's a bit more than just the specrometer)
$ cvs -d:pserver:anonymous at pcbcn10.phys.rug.nl/var/lib/cvs co drivers

It may very well be an academic exercise to add support for this device.  I do 
not know anyone who uses it (does anyone on this list?) and Avantes has 
increased the prices for these by almost 20% just a few weeks ago.   
Considering that these are competing against the i1 (which has more 
capability if you are in the US) and the Pulse either of which can be had for 
about half of what the AvaMouse/SpectroCam cost I think it is unlikely that 
anyone will be using one of these unless they already have the device or 
happen to find a used one cheap.  I saw one on ebay a few months ago that was 
in "like new condition" and had both Windows and Mac profiling software with 
a $500 reserve and no one bid on it.

snip

> In Argyll I have a kind of abstract instrument layer, and the drivers
> are implemented underneath that. Every instrument usually causes some
> expansion of the abstract layer, but it's generally not too hard
> to add support, if the information is available.
> I'd love to have support for the i1 and i0, but I'm certainly
> not going to fork out the A$6K (or whatever it is) for an
> i0 and i1, without firm knowledge that the technical information,
> or compatibly licences driver code, is available.

I have seen new i1s on ebay for as little as $700 US.   But again the real 
issue is with the interface information and the fact that they have not made 
the library available for any platforms other than Windows and Mac.  In 
addition, the SDK license is such that GMB will allow you to release your 
stuff as source code but you can not release anything that interfaces to the 
i1 library as GPL since the code MUST have the following included either in 
the source code itself or in the license:

"This source code may incorporate intellectual property owned by GretagMacbeth 
AG, GretagMacbeth LLC or one of its subsidiaries.  Our provision of this 
source code code does not include any licenses or any other rights to you 
under any GreytagMacbeth intellectual property.  If you would like a license 
from GretagMacbeth (e.g. to rebrand, redistribute), you need to contact 
GretagMacbeth directly (send mail to prorcollicensing at gretagmacbeth.com)."

The above is not onerous but it means that at least some of the code can not 
be released as GPL.  I had intended to get an OEM agreement with GMB to make 
the devices (i1 pro and display) available to LPROF users or anyone running 
Linux/Unix/BSD without having to pay for GMB software (which is Windows/Mac 
only) and then build the device support into an LPROF plugin.  That way I 
could distribute the GMB specific software (the plugin) from the same web 
site as I used to sell the devices as free but closed source since the GMB 
license precludes it being GPL or in any way open source.  Not an optimal 
solution but one that I think would work.  I would however prefer to be able 
to open source the whole thing as this would in fact make things simpler for 
everyone.

The merger with X-Rite creates a high level of uncertainly about all of this 
as well.  Will the i1 be available in a year or two?  How about the Pulse?  
At some point the merged company is going to consolidate their product 
offerings and I expect this to happen sooner rather than later since they are 
expecting to have cut operating costs by $20 million/year in the second year 
after the merger.  One logical and very likely approach is to eliminate 
redundant product lines - both hardware and software.   With software they 
can merge the development/support teams and consolidate the best features of 
each line into one set of product offerings and then cut back the size of the 
development and support teams after the software is consolidated.  On the 
hardware side it will not be that easy to do that kind of thing.  I expect 
that in the short run they will have to choose one product line and eliminate 
the other perhaps keeping one or two devices from the eliminated product 
line.   I expect it will be at least another 6 months before we will have any 
clues what the hardware product line of the merged company will look like.

>
> > I am in talks right now with X-Rite.  At this point I have no indications
> > of how receptive they are but they have not told me to go away and the
> > form that I filled out for them did list Unix as a platform and I checked
> > other and then wrote in POSIX as my platform.   So I don't know if it
> > will go anywhere but this is in the very early stages.
>
> Xrite were pretty good in the past, with their serial port based
> instruments, but I'm not so sure that things like the "Pulse" or
> "MonarcoOPTIX" will have accessible interfaces. Presumably the DTP45 might
> be a variation on the DTP41, and the DTP70 might have a high level
> interface, but it hasn't been worth pursuing, since I don't know anyone
> who's
> got one, and would let me work with it.

Their OEM form lists all of these as possible targets for an OEM agreement.    
When I contacted X-Rite they seemed to be more responsive than GMB.   Even 
though I am not very optimistic about the short term prospects (because of 
the merger) and I am hopeful that longer term I will be able to work with 
them.  But this will likely result in a significant delay in getting 
instrument support in LPROF since I am at a point where I would like to start 
working on this now.  If I had the necessary agreements/assurnaces in place 
now I would have hardware monitor profiling in place (at least beta code) in 
two or three months and I would follow up with printer profiling shortly 
thereafter.

snip

> If you put yourself in their shoes, you can understand that they will
> do the cost/benefit equation (as they see it). Who pays for the
> development instruments becomes an issue. Without the instrument
> company seeing enough benefit to provide some loaner/free
> instruments, or a moneyed supporter of OSS willing to buy them,
> things aren't going to go anywhere.

I am setup to get donations on the LPROF SourceForge web pages.  These 
donations are specifically to help defray the cost of buying these devices 
for doing this work. I have already gotten some donations which is somethat 
surprising considering that LPROF has only been a SourceForge project since 
August of last year.  But not nearly enough to purchase a device.   In 
addition I believe that the CGI company that wants monitor calibration 
support would also help out with this cost.  I had expected that when I got 
to the point where I had secured an OEM agreement that I would use what ever 
was available from these donations plus money from my own pocket to get the 
devices needed to develop the software.  Of course I was also thinking that 
over time the sale of the devices to LPROF/OSS users would allow me to recoup 
this initial cash outlay and help support ongoing LPROF development. 

So in my case I had a plan that made development hardware costs a non-issue 
for the hardware vendor and I never asked any vendor for free or loaner 
hardware.   The only remaining issue for the vendor is the costs of creating 
a support structure for the non-windows/mac drivers.  If the vendor wants to 
do binary only drivers this makes it significantly more costly for them to 
support the drivers since they then have to make these available in a wide 
range of configurations.   There are over 1 dozen hardware architectures that 
need to be supported (I know that LPROF has been built for at least 13 
architectures) and then on top of that the driver binary would have to be 
available for each architecture compiled with various versions of GCC to 
allow users to select a version that had binary compatibility with the other 
software on their system.   This makes binary only drivers a fairly ugly 
thing to support.  On the other hand a source code driver requires none of 
this and makes things fairly simple.  Unfortunately this is a huge paradigm 
shift for these companies since this is completely the opposite from the way 
things are for Windows/Mac and it may take them a while to "get it".

>
> > At this point I am not particularly optimistic.  The Avantes device is a
> > reflection only device for those of us in the US because of patents held
> > by GMB.
>
> Hmm. The SpectroCam I've played with did emission as well.

Models sold in Europe do also.   I don't know about the UK.   Current US 
models are crippled and do not support emission readings.  They say this is 
because of patent issues.   I suspect that it is actually the US drivers that 
are crippled rather than the hardware.   If you go to their web site you will 
see this documented in several places.  They also have a US, UK and every 
where else models.  Since GMB is the only vendor that currently has meters 
that support both emission and reflection in the same device in the US (the 
i1) I suspect that they are either the patent holder or have an exclusive 
license with the patent holder.  

Hal