I'm an Italian Linux user trying to become a kernel hacker.
I'm now working with the writer of the USB kernel driver for Linux. I'm
writing some parts of the guide, studing how the subsystem works and want
to write a driver for it when it will be finished.
It will works like SCSI : a subsystem with drivers for any avaible
hardware controller (only two are avaible, no more are in plan) that
provides high level services for drivers for hardware USB devices.
As USB architecture is planned, writing drivers will be very easy !
Due to exams and thesis we both are stopped (Inaky coding, me writing).
I'm happy that a project to standardize access to devices has been
started, so drivers writer have a starting point.
Lot of USB hardware of this kind is coming !!! And support for FireWire
is coming from .at domain as well !!!!
My questions are :
1) SANE is not planned to have mapped access of devices in the filing
system. Everything is accessed from functions, like networking in Linux,
not via devices mapped in the filing system, like sound. Am I wrong ? Is
mapped access planned for the future ?
2) What SANE wants to be ? Like TWAIN or like TWAIN + VIDEO for Windows ?
I think the access to networked resouces is an excellent idea, but I want
to make a proposal.
I'm reading the SANE code. As far as I've understood links to remote
resources are made via TCP to provide images without any error. Animations
from Quickcams are sequences of images one after the other. This is
perfect for scanned images (= TWAIN)
If you want to develope something bigger, I mean SANE = TWAIN + VIDEO,
things must be different.
The technology you use now is perfect for scanners and LOCAL streaming
video or over an UNUSED network due to with bandwidth.
If we want to make SANE the standard also for moving images (Quickcams via
USB and Digital Video via FireWire are coming !) and video-conferences, I
suggest to use also UDP (frames lost are not so important in video
conferences, you must adapt to avaible bandwidth).
Every video source should provide data in three formats :
1) Best quality=zero compression, for local playback, recording or video
editing.
2) Good quality=small compression, for playing across fast networks like
ethernets.
3) Minimal quality=big compression, for playing across slow lines like
analogical or 64K ISDN lines.
Do you think SANE should be split in two different projects ? One for
sanners and one for Quickcams ?
Maybe VANE (Video Access Made Easy) could define a standard so SANE will
have a single device for every *cam to obtain a still image.
They both (SANE and VAME) could be part of a bigger project (maybe
UDA=Uniform Device Access) that will document and collect APIs to access
deices of the same family (Sound devices, joysticks and so on).
What do you think about it ?
I think that Windows if too strong for office apps, and so Linux has lost
before beginning any competition for desktop operating system.
But I think that Linux has a great future in more than one field, if we
(all linux users and all kernel writers) play well our cards before year
2K. Linux can be the right operating system as :
* Number cruncher in scientific applications, where SMP, source
avaibility, Real Time and a light operating system is necessary.
* Network machine, as router, bridge, file server, network server,
spooler, fast digital network point, fast multimedia conferencing system,
fast, Fast, FAST ! Look at net pc developed from Corel. It's a StrongARM
processor running Linux.
Win NT is full of apps, well documented, solid, but SLOW ! This is the key
of success for Linux in the future. Expecially for the last point I've
written.
I'm waiting for comments about this letter.
CIAO CIAO
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com