Re: [Printing-user-general] Why has nothing changed?

[email protected] (Roger Morgan) writes:

> Today, in 2006, it is obvious to anyone who looks at the forums that
> CUPS is *still* a nightmare to configure. Practically no progress
> has been made in 2 years.

Incorrect.  No progress has been made in *10* years.  Or even 20...

Before CUPS was a pain to configure, LPD was painful instead.  The
diagnostic features weren't there in LPD systems, either.

Both have the same basic architecture, that of a queue-oriented
actively managed network service.  Although I don't think that this is
the correct architecture for many -- or most -- installations, it is
also not the problem.  Rather, CUPS and LPD based print systems are
simply hard to use.  It is a pure implementation issue.

No doubt 3/4 of the issue is that no one person or group implemented
the entire system.  With LPD, one person implemented a basic spooler;
dozens of different people implemented filters, configuration, and
other integration logic; and still other people implemented drivers.
Much the same has happened with CUPS, although the functionality is
distributed differently, with more present in CUPS proper and less
done by the integrator.  In both cases the result is a nightmare where
most problems are caused by some aspect of the integration rather than
by the individual bits of actual software.

When one group takes it upon itself to produce an integrated and
functional system, the result is not bad.  Apple is the shining
example of such a system built from free software components; OS X
bundles CUPS with Apple-written configuration and other logic and a
block of proprietary drivers, and it basically just works.  Certainly
better than what unthinking Linux distributors provide.


> Most people who try to configure CUPS eventually come to Kurt
> Pfeifle's page on CUPS Troubleshooting:

Yes, of course; most of the prose on linuxprinting.org is hopelessly
verbose and obsolete.

For a time I made it my project to highlight the CUPS quickstart page,
and even changed some of the software to make the page smaller.  Alas,
that page has grown ever since (originally it fit in a browser
window!), and no one has ever had time to prune all the other prose
floating around.

In any case, Till has a brilliant scheme up his sleeve -- he intends
to replace the bulk of the website with a Wiki.  The prose will thus
become SEP and the Foomatic database will be easier for people to
contribute to.

Alas, no amount of shuffling documentation around addresses the real
problem, which is software that nobody wants to think about yet
everybody *has* to think about because it fails to work too often.
Perhaps the new FSG efforts will solve the problem; the standards
folks might be able to create enough goodness to replace some of the
ad-hoc interfaces causing the most trouble.


-- 
Grant Taylor
Embedded Linux Consultant
http://www.picante.com/