[sane-devel] Scanner Database

[Michael Nagel]

Hi everyone,

I'd like to hear your opinion/feedback on the following idea:

Background:
I am currently sorting the bug reports filed against Simple Scan in Launchpad.
Simple Scan is a, well, simple scanning application focusing on casual scanning;
it uses the SANE backends, it is the default scanning application in Ubuntu
and you might actually have heard about it...

...anyways, there are lots of bug reports, many of them from novice users,
that are not really about the scanning application (or "frontend").
Many issues are with the backends, exists because the wrong backend is used,
are in some way hardware-related and/or depend on how all of that is bundled
together by the distribution...
Also, often these reports are created and then receive no attention whatsoever.
Nobody works on them, and the reporter doesn't provide follow-up information.
Asking for information often gives no response at all, or the reporter either
gave up at some point or the problem magically resolved itself somehow in the
meantime for unknown reasons.

Such tickets are frustrating for everyone involved (both the creator as well as
the developer) and in my opinion actually slow down work.

The tickets are not completely useless, however.
They contain some valuable information, BUT that information would be much more
useful if it was somewhat structured and stored in a database.
Besides letting aforementioned users make their statement without cluttering a
bug tracker (and probably leaving them with the feeling that they did their part
and now something is going to happen and their problem will get resolved) such a
database might actually be very useful for different use-cases:
1) users can search it for information on how to get their scanner working
2) users can browse it to decide what scanner to buy
3) developers might analyze it to get an overview and spot regressions

HOWEVER: unfortunately, there is no such database currently.

[
This is not about replacing the list of supported devices available at:
http://www.sane-project.org/sane-supported-devices.html
primarily for the following two reasons:

1) It's about making it easy for as many people as possible to submit data.
2) It is about measuring the real-world, end-user experience,
not about "what should theoretically work".
Unfortunately at least for Simple Scan on Ubuntu
we regularly see regressions with scanners that "should work" and *have worked*.
]

As a basis for discussion I created the following mockup:
http://nailor.devzero.de/tmp/scandb/loop/hardware.htm

This is in a very early brainstorming phase, so everything is possible.
Neither the set of questions, nor the methodology to gather the answers
(or even the fact that it's a website) is final.
I'd like this to be as generic as reasonable
(and e.g. not (only) a plugin for Simple Scan),
making it useful and usable for users of different scanning applications
and users of different distributions.
On the other hand, it should be easy to submit useful (structured) data.

I talked to Robert Ancell from Simple Scan,
and he is definitely interested and supporting this.
I talked to Kåre Särs from Skanlite/libksane,
and he is basically supporting this as well.

So get your hats ready and tell me what you think about it as a
SANE-standards-writer, backend-developer, frontend-developer,
packager for a distribution, debugger, supporter or end-user
and tell me what information would be useful for you...

All and any feedback and help is welcome.

Best Regards
Michael