[Openicc] GoSoC 2011: CPD and target printing

Chris Murphy lists at colorremedies.com
Fri May 13 18:57:22 PDT 2011

On May 13, 2011, at 5:23 PM, Michael Sweet wrote:

> On May 13, 2011, at 4:13 PM, Chris Murphy wrote:
>> I've never seen one single application or RIP I've ever used replace cgpdftoraster. I can't imagine Adobe would touch this with a 10 foot pole.
> Nope, because then they couldn't charge per-seat (printer sharing).

I think it has a lot more with going into the store and possibly breaking a bunch of stuff they don't want to buy. If they bust core parts of the OS, either directly or it ends up broken with an OS update - too risky. If it really were a piece of cake, risk free and fixed this eye sore of a problem that only exists on Mac OS X, I think they'd consider this like a sure thing lottery ticket.

>>>> ...
>>>> b.) There is no user or application access to the pdf print spool file.
>>> PDF workflows (which were added in 10.4 I think) give you access, and CUPS allows the owner of a job to access the spooled jobs since 10.6.
>> /var/spool/cups is owned by system, everyone else has no access, on 10.6.7 and since forever. I've always had to change their owner or add an ACL to access the spooled jobs.
> There is API to access the spooled documents. No, you can't access them directly, and for good reason...

I'd sooner say follow the PDF/X-3 or 4 spec and fix the APIs if there is a problem, rather than ban a portion of a functional specification that's known to work. Rather that defend the ban on /DevicRGB in a very specific kind of PDF spool file, a ban which itself has caused more problems than it has solved.

>> Anyway, there is no meanginful or practical user access that would pose a problem in normal function of the system. That spool file lives for a VERY short amount of time. Interception and interaction is non-obvious.
> PDF workflows resolve this particular issue and are accessed through the print dialog...

To this day I have never heard a good explanation why /DeviceRGB was banned in what is otherwise nearly a PDF/X-3 based PDF print spool file. The requirement to double tag every object isn't consistent with the spec, rationality, or even best practices because it adds obscurity, complexity - and here we are 7 years after this was predicted to be highly problematic and all of those predictions have come true.


More information about the openicc mailing list