OT: printer support [was: Re: Should I add this Repo?]

[Mario Vukelic]

On Tue, 2008-10-28 at 15:40 -0400, charlie derr wrote:
> Or is the worry that discussing it on the list as a potential fix will
> lead someone else to possibly try the same command (and break their
> own machine)? 

The main concern, as I understand and have it, is that Brother tells
users to execute potentially damaging commands (at least
--force-architecture, but as the link shows that I just posted to
ubuntu-users, also the the infinitely more dangerous and stupid
--force-all).

This is not only bad because the install of the brother driver might
break things for some people but, more importantly, it conditions users
to accept that dangerous commands are not all that bad ("I used some
force option last time, so what can possibly go wrong?").

If GNU/Linux ever gains significant mainstream success, such
conditioning is a sure vector of attack by malware. Windows experienced
how this goes. A number of software providers who engage in such broken
behavior can neuter all of Ubuntu's attempts to protect the use. It's
not a long way from mindlessly executing --force-all at package
installation to answering every OS prompt with "yes" and provide the
password to whomever may ask.
Protecting naive users on the internet is hard enough without legitimate
software providers making it needlessly harder.

Brother shows its inexperience with Linux distros and is not being
malicious. Nevertheless, rather than accepting such nonsense, users of
their products should educate them that harmful antics learned in the
Windows world won't fly in the free software world (hopefully).
Their driver works in 64 bit systems anyway, apparently, so why not
spend the few minutes and create a dedicated package that needs no
--force-all,architecture?

Plus, this broken package installation advice prevents the package from
being installed easily and nicely by simply doubleclicking it, as it
should. Installation-on-doubleclick has been demanded for years, along
with being touted by some folks as a requirement to match what is
perceived as the ease of use of Windows. Why break this without good
reason?

It is of a minor concern to me that such an installation method is
defended on a support list and the providers of said broken method
celebrated as having "the greatest Linux support", but I don't find it
particularly helpful, either (as probably was apparent :)

> Even should we accept that as a possibility, I'm not sure how anyone
> could possibly blame Leonard.

Nobody did. I understood Derek to mean "if you cannot conceive how an
option that is marked as dangerous could be dangerous, then you should
not execute it". This is sound advice, AFAICT