HI !
> PR >and keep woundering if it would be a good idea to adapt the TWAIN API
> PR >as the high level API for scanners on UNIX.
Hmm - I am not too convinced about using the TWAIN API for that purpose.
It is interesting for using SANE on Windows and for quickly porting existing
applications, but in that case the main porting effort will not be the
scanning part, but the GUI stuff that makes up the rest of the application.
Basically I would rather suggest to have a standard way for "remote-
controlling" SANE frontends, so that an application would be able to
spawn off any SANE frontend the user likes, tell it to set some options
to preset values and then get back an image from the frontend - either after
user interaction or directly.
This is more or less what TWAIN does, but much cleaner for a Unixish
environment. TWAIN had to use quite some ugly clutches (single entrypoint,
piture handed back either in a large memory buffer or in a file, ...) that
originated form the system environment it was designed to run on.
Such a standard remote control API could then as well be exploited by a
TWAIN layer to export a true TWAIN source on Win.
> PR >We would only need to implement one TWAIN driver - the SANE driver,
> PR >and a TWAIN API library to use this.
Please all interested parties: have a look at the TWAIN specs first, then
continue discussing it. The old discussion suffered of a lack of knowledge
about some TWAIN specifics in some points. The specs are openly available.
Oliver Rauch, with whom I had discussed this topic at great length before
I stopped active participation in the SANE project, should have pointers
and copies of old conversations.
CU, ANdy
-- Andreas Beck | Email : <Andreas.Beck@ggi-project.org>-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
This archive was generated by hypermail 2b29 : Thu Mar 16 2000 - 06:56:22 PST