Re: Starting a discussion about SANE and TWAIN...

Oliver Rauch (oliver.rauch@Wolfsburg.DE)
Tue, 17 Aug 1999 18:38:31 +0200

252353N@knotes.kodak.com wrote:

Hi Mark,

> TWAIN doesn't really have any layers. At it's most
> abstract description TWAIN is nothing more than an
> elaborate conversion API between a standard name
> space to a set of arbitrary native commands for some
> given device.

Yes, but which parts are bonded together?

If I understand it right there is one source manager
on a system. If we have two different scanners with
two different drivers (sources), this source manager
gets contact to the one or the other driver.

I am able to select one source from one application
and the second source from another application.

Or is the source manager included into the application
and each application has its own sourcemanager?

How is the contact done, is the driver a library that is
loaded on runtime?

> This is why the notion of a sane.ds file is so appealing
> to me.

sane.DS? Did I miss a mail?

> In this model TWAIN becomes an application
> using SANE, and should therefore receive all the
> benefits that any application using SANE would expect
> to receive.

Yes, and I don`t think it is a problem to add a
TWAIN interface to an existing sane frontend.

> However, if the sane.ds has the intelligence
> to discover SANE devices on the network, then we gain
> remote scanning control with no change to the TWAIN
> spec (and more importantly, no changes to any TWAIN
> applications).

That is no problem, there is no difference for the frontend if
it connects to a local backend or to the network backend
that connects to a remote scanner.

> I guess that's fundimental to my current notion of the best
> implementation of a TWAIN using SANE and SANE
> using TWAIN system. How much can we do to enable
> applications access to more image capture devices
> across more platforms, and do it with the applications
> changing little or no code.

Are you interested in programming code for unix,
especally to implement the source manager and
a test application? Of course we will work together
but we need somone who is familiar with TWAIN.

Andreas Beck and me also need help for creating
the interface between the source selection manager
and the sane frontend.

Do you like to do this and do you have the time for it?

Andreas and I talked how the communication by
events between application and source can be
implemented on unix.

It looks like there are already differences between
Windows and Mac.

I have to read the TWAIN standard at first, but it will
take a while (>500 pages and less time).

Do you already have the sane standard and source code
example (you can download it from the sane homepage).
If you have problems to read the formats of the files/archieves,
let me now and I will convert them for you!

Bye
Oliver

--
EMAIL: Oliver.Rauch@Wolfsburg.DE
WWW: http://www.wolfsburg.de/~rauch

--
Source code, list archive, and docs: http://www.mostang.com/sane/
To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com