[sane-devel] CanoScan LiDE20 - broken driver and so

Aldo blinuxman at rack66.com
Sat Aug 27 12:09:34 UTC 2005


On Fri, Aug 26, 2005 at 07:06:35PM +0200, Henning Meier-Geinitz wrote:
> In my opinion, the problem is that too few users test our snapshots
> before the release. There are at least three weeks of feature/code
> freeze where everyone should really test his scanner and tell us any
> problems. If everybody just waits for the new release and doesn't test
> problems are found too late.
> 

I accept and understand this point of vue, but 
a) this supposes that someone has an official system, (who may not be 
dammaged by tests)
b) and another, (tryout or whatever) system who contains the same distro 
(for emulating the same conditions):
depending on the space of your hd it's not always possible to have such 
chance to do this;
on the other hand IMHO i think testing beta applications on an 
"official" system is not a good idea sicne I discovered that under 
Gnu/Linux its sometimes much more difficult to return to a previous 
situation.

But in general I thing you're right ! no test = no feedbak = bug found 
too late.

Now regarding the Canoscan LiDE20, I hadn't read here taht the driver 
was fixed in the latest release, isn't it?
BTW, for my girlfriend who is new to Sarge Debian, she prefers of course 
installing stable packages using Synaptic under Gnome or so.

Maybe CUPS can be a good inspiration for the way SANE should reorganize 
the way it is developped, maintained, and structured? (just a brain 
storm, I'm not developere).

Okay, these are two *absolutely different* products, but maybe there is 
a "way of working" who is fundamental different and that explains the 
reasoon why eahc adding of a new printer doesn't cause broken things on 
other parts of the package.

Hope this helps (cf. Mounty's reflections).

Aldo.




More information about the sane-devel mailing list