[sf-lug] Xsane can't see an HP Laserjet 1536dnf MFP scanner / printer

Rick Moen rick at linuxmafia.com
Tue Jul 31 22:16:06 PDT 2018


Quoting Akkana Peck (akkana at shallowsky.com):

> I agree in general, though CUPS might be a special case.

There's some danger, here, of attributing to me a view I don't hold and
didn't articulate.  Nowhere did I claim open-source drivers are
necessarily better than proprietary alternative (or even good, come to
think of it).

All I said was that drivers from manufacturers tend strongly towards the
problems I listed.  In cases where they're your best option, then, hey,
they're your best option.

I personally try to avoid being in that situation, though.


Regarding the Brother HL-3170CDW, Openprinting.org has a troublingly vague
page.  Manufacturer specs say it's an LED (not technically laser)
printer, which is fine.  It has only 128MB RAM, which for reasons
mentioned upthread is troubling and pretty ominous, suggesting it may be
a printer that was deliberately crippled by omitting key 'engine'
electronics and instead emulating that with proprietary software
offloaded onto the attached workstation.  Specs _claim_ it emulates PCL6
as its print language, and also an offbrand imitation of PostScript3
('BR-Script3').  I sense a certain amount of sleight-of-tongue, in that
nowhere is it claimed you can use a generic PCL6 or PostScript print
driver -- and also that seems totally incompatible with having only
128MB of RAM.

I suspect that, instead, this is yet another printer that relies on
offloading of page rasterising to secret-sauce software running in the
workstation, and that _that_ software can be addressed using PCL6 or PS 
printing commands.

Seems like a unit with definite problems.  Me, I'd have avoided it.

Regarding the Dell E310dw:  Oops, no Openprinting.org page at all.
Ominous.  Also, I note that this is yet another multifunction device,
which category as I've mentioned has had an unfortunate history.

(I see your page about all this, at
http://shallowsky.com/blog/linux/cups-printers-urls.html .)

Someone on an Anandtech forum claims the Dell D310dw is a 'rebadged
Brother', which would follow the pattern because Dell does that widely
across its product line (e.g., RAID controllers).  Someone on Amazon.com
claims it it specifically a rebadged Brother HL-L2360DW.  The
Openprinting.org page about that has contested information:
http://www.openprinting.org/printer/Brother/Brother-HL-L2360DW

One person claims that you can use a generic PCL6 CUPS driver with it.  
Others contest this -- but I'm going to guesstimate that the former
person was using ethernet  networking using the built-in wired Ethernet
RJ-45 jack, and the latter were attempting some strange wireless
printing possible only using proprietary drivers.

So, if I owned that printer, I'd try PCL6 (and PCL5e) generic printing
over plain ol' ethernet.  Personally, if I worked away at that sort of
thing for an hour and it was still not a go, I'd sell it.

> I tried to use the CUPS web interface (on Debian testing) to
> configure them. In both cases, there were several versions of the
> printer available in the CUPS menus, but after spending a couple of
> hours trying every combination of driver and port, I couldn't get
> any of the available versions to work.
> 
> In both cases, I ended up going to the manufacturer's website,
> downloading a couple of .deb files, installing them, and then
> going through CUPS configuration again, and was able to get
> them working after some additional CUPS fiddling.
> 
> I think this probably says more about bit-rot and lack of testing
> in CUPS, rather than anything that actually contradicts Rick's
> point. 

I don't think it indicates any problem with CUPS, on the basis of what
you said.  If you examine the downloadable files from the manufacturer,
I believe you'll find that they have _no_ grant of permission to anyone
else to distribute them.  If so (as I suspect), then Linux distributions
cannot lawfully include them in CUPS.

You _visiting_ the manufacturer's Web sites, the only places by default
with legal permission to distribute those copyrighted works, of course
got you copies.

I can't see why this situation would be seen to show any problem with
CUPS -- unless you think it's the job of CUPS to inform you about
software nowhere on your system that you _could_ fetch and install.




More information about the sf-lug mailing list