[Foomatic] Using usb IDs for autodetection

Joe Shaw joe at ximian.com
Thu Apr 8 11:59:56 PDT 2004


Hi Grant,

On Wed, 2004-04-07 at 18:13 -0400, Grant Taylor wrote:
> Joe Shaw <joe at ximian.com> writes:
> 
> > I noticed that the foomatic printer XML files have an <autodetect> flag,
> > but virtually none of the printers have anything in it.  
> 
> Yes, this data is intended to fill your very purpose.
> 
> It should be entirely straightforward to write a /sbin/hotplug
> subscript, or similar, which invokes foomatic-configure, or similar,
> to make printers "just work" when you plug them in.  Or similar. ;)

Great!

> > I'm proposing that we add <vendor_id> and <product_id> tags to the
> > <usb> section of <autodetect>, and populate the printer database
> > with those IDs.  This will make autodetecting a lot more reliable,
> > at least for USB printers.  I'd also ask that this information (and
> > probably also the IEEE-1284 data) be required for any new additions
> > to the database, so that we can keep autodetection reliable.
> 
> We can require, but we won't "get".  Data entry is a problem, data
> verification more so.  It's basically Till putting in all the data (on
> Mandrake's dime), with the odd commit from other folks.  

Well, for the USB IDs, there's got to be some streamlining happening
here, if even from the distributors who do the work for a distro release
and then submit the stuff back upstream to you.  The IDs are almost
certainly already in usb.ids.

>From where is Till getting the info?  Does he have the hardware?  Is he
getting the info from Mandrake users or vendors?  If a program were
shipped with foomatic which made it simple to get this information
(similar to what's included with XFree86 or sane), then I think that'd
go a long way to helping solve the issue.  If the user is saavy enough
to want to help out and send printer information, it wouldn't be much
burden to run this program and send the output.

> We also have the longstanding issue that foomatic data is slightly
> insufficient - nobody has had a chance to add useful vendor-ppd
> support to the core data, so postscript printers aren't doable
> properly.

Yeah, this seems like an issue.  To my novice mind, it seems like it'd
make sense for the foomatic DB to be the canonical DB for this info.
Vendor PPDs would be fine, but they'd also need to provide an XML file
to tell the rest of the infrastructure about it.  If Linux distributors
were willing to standardize on this approach, it'd probably go a long
way to enlightening the vendors.

> Hmm.  Ximian and predecessor projects have started two or three of
> these efforts over the years, I do hope yours succeeds.

I think that the difference between the previous projects and this one
is that there wasn't an attempt at a holistic (or user-centric!)
approach to the problem.  We're not just trying to make printers easier
to use, but all kinds of hardware.  Printing is just one of the problems
that I'm focused on right now. :)

> I strongly advise that you coordinate with the kprinter folks (Michael
> Goefful et al, really nice guy), plus Mike Sweet and Patrick Powell.
> Printer configuration should be *totally* independent of the gui du
> user.  This is a key foomatic design goal, and should be for you, too,
> even if you're Ximian.

100% agreed.  In fact, I'd like to leave the printer configuration to
YaST, redhat-config-printer, iPrint, the XD2 printer wizard, whatever.
They're already very good at that.  If they could be taught about HAL
and the hardware universe, then all the better.  My interest at this
point is automatically configuring the printer without user intervention
(paricularly choosing the "driver") whenever possible.

So I guess my next question is, would you accept a patch to the XML
files which added the tags I previously mentioned with the data?

Joe





More information about the Printing-foomatic mailing list