[sane-devel] discussion: Future of SANE-project

[Henning Meier-Geinitz]

Hi,

On Wed, Jul 07, 2004 at 11:45:50AM -0400, David N. Paules wrote:
> Perhaps I misunderstood SANE's goals. From what I've seen, the
> backends do not dictate what the front-ends look like,

Right.

> only their capability.

The capability of the scanners/backends. They don't dictate what the
frontends do with the data (e.g. color correction...) or which options
they do not display either.

> I don't understand why the bluster about multiple
> front-ends and more control over the gui. I was under the impression
> that one sane backend for a device will support multiple applications
> (front-ends) such as image capture, OCR, digital camera photo image
> acquisition, video acquisition from a digital camcorder, etc.

True. But one of the ideas of SANE is to avoid writing one frontend
per scanner as it is the case with TWAIN. 

> Besides, the SANE project is open for review/suggestions for
> enhancement, so why not get some feedback from scanner makers on what
> makes SANE a non-option for the future? What 'minimum requirements'
> are missing from the current architecture?

Go ahead and ask them!

> I can see your point about GPL'd backends. I agree it is easier to
> debug and extend them. But here's my point: Do you want to be the one
> always repairing and fixing scanners to work with SANE? Or do you
> want all the manufacturers to feel responsible for ensuring SANE
> compliance? If SANE is architected intelligently (i.e., backwards
> compliance with older backends), open source backends may not be
> necessary.

Binaries are alway a problem. Think of new hardware platforms, changes
in the kernel etc. And experience shows that manufacturers don't
really care about their binaries once the devices are older than a few
years. I constantly get questions from Windows users if I can send
them a driver for their scanner for Windows 2000, XP whatever new
Windows system. The reason is that the manufacturer doesn't provide
the drivers for these systems and the old drivers don't work.

> The manufacturers will currently see the SANE org as a free resource
> for reaching other markets they didn't officially target (i.e. the
> *nix markets). There is no urgency or requirement for them to support
> SANE just because the project is open source. In fact, that is the
> number one reason they will NOT be concerned. 'Those hackers at
> SANE-project.org will make my device compliant with their software.
> And I don't have to pay for it.'

If one of those hackers has already written a backend that assumption
is true. But as there many unsupported devices that will probably not
supported in future (e.g. many Canon and Microtek devices)
manufacturers should have found out years ago that it doesn't work
automatically.

In many cases the backends that are part of SANE at the moment are
based on some sort of data provided by the manufacturers or even based
on their code. Some of the manufacturers that don't provide backend
nor data should have found out that backends don't mystically appear :-)

> But, if the ultimate goal of SANE is to supplant TWAIN as the
> standard for accessing image devices,

I'm not sure if SANE has a goal apart from providing a standardized
interface to raster image hardware.

My personal view is that I don't mind if SANE also runs on Windows (it
does) but I won't put any effort into this. I just don't care.

> shouldn't more effort be put on
> making SANE the future standard. This route will cause scanner makers
> and their support network to conciously choose to either support or
> not support the new standard. The SANE project's developers should
> then concentrate on enhancing the architecture, ensure backwards
> compatibility, create and market compliance badges,

There have already been some discussions about "SANE compliant" logos:
https://alioth.debian.org/tracker/index.php?func=detail&aid=300269&group_id=1308&atid=410366

> create plugins
> for Adobe products (like existing TWAIN plugins), etc. and not write
> new backend modules for each new device that comes on the market.

Free software developers usually do stuff that's interesting for them
in some way. For me it looks like that Windows stuff is rather
uninteresting for most of us.

> Perhaps working examples demonstrating how scanner makers can
> leverage their investment in TWAIN support while migrating to SANE
> might be useful for getting scanner makers to take notice of SANE. If
> an upgrade path was readily shown, techno-geeks at the manufacturer
> might prefer and sell the SANE idea to managers within the company.

Please provide facts/information/data. I'll add it to our website.

> Formal propoganda of industry heavyweight support (a consortium of
> scanner makers and front-end application development houses) on the
> web-site might make other makers feel the need to join in.

I don't like propaganda at all. I don't mind pointing to facts.

> Without this kind of direction, I simply see SANE as being a
> hobbyist-level effort which is why I questioned the future of it.

SANE is (mostly) "hobbyist-level" since 7 years. I think it works
quite well. 

As already mentioned above I don't mind helping hands from
manufacturers. If have good contacts, just ask them, discuss with them
and point them to the facts.

Bye,
  Henning