2004-08-03 07:47:16 +00:00
|
|
|
To do:
|
|
|
|
======
|
|
|
|
- implement coarse calibration to determine settings for analog-front-end
|
|
|
|
All CCDs encountered so far seem to need about the same settings however.
|
|
|
|
|
|
|
|
- Store analog calibration data somewhere.
|
|
|
|
Use currently used values as default.
|
|
|
|
|
|
|
|
- Suggestions by Marcin Cieslak:
|
|
|
|
- clean up structures
|
|
|
|
- clean up namespace
|
|
|
|
- clean up source formatting ('indent --gnu' ?)
|
|
|
|
|
|
|
|
|
|
|
|
- implement real calibration, using 12 bit data at 600 dpi
|
|
|
|
instead of 8 bit data.at 150 dpi.
|
|
|
|
-> is now at 600 dpi, not yet at 12 bit
|
|
|
|
|
|
|
|
- design a mechanism to set the scan and calibration areas
|
|
|
|
a) part of HWProps structure, or
|
|
|
|
b) set by user at first scan
|
|
|
|
|
|
|
|
- further investigate scanner registers
|
|
|
|
a) study backtracking register more
|
|
|
|
b) identify bits in bit-oriented registers.
|
|
|
|
|
|
|
|
- add more error checking ('sanity checking') on parameters
|
|
|
|
|
|
|
|
- figure out what the USB interrupts mean that sometimes occur.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Done
|
|
|
|
====
|
|
|
|
|
2004-08-31 18:43:58 +00:00
|
|
|
- Get rid of mytypes.h / fix types
|
|
|
|
Now an int is implicitly assumed to be 32 bit.
|
|
|
|
|
2004-08-03 07:47:16 +00:00
|
|
|
- make backend comply with tstbackend
|
|
|
|
-> all externs and publics removed when compiling towards sane
|
|
|
|
|
|
|
|
- implement sane_cancel properly
|
|
|
|
-> done by ullsig
|
|
|
|
|
|
|
|
- implement timing mechanism for scanner lamp
|
|
|
|
so scanning can only start if the lamp was already on for 30 seconds
|
|
|
|
Possible implementation:
|
|
|
|
On startup: if lamp is off, turn it on and start timer
|
|
|
|
On scan: if lamp is off, turn it on and start timer. Wait for timer expiry.
|
|
|
|
On exit: if scanner chip is niash00012 then turn off lamp,
|
|
|
|
otherwise let the scanner turn it off automatically
|
|
|
|
-> Implemented by Ullrich Sigwanz
|
|
|
|
|
|
|
|
- try out parallel port scanning on HP3400/4300
|
|
|
|
-> parallel port support will be dropped, no people to test/develop it
|
|
|
|
|
|
|
|
- investigate vertical resolution setting of hp3400/4300.
|
|
|
|
-> experiment with the motor table, formula is already known to generate it
|
|
|
|
-> 600 dpi is possible by doubling most values in the table
|
|
|
|
-> 150 dpi can be done by using register 0x06 to skip every other line
|
|
|
|
|
|
|
|
- Add separate gamma tables for R, G, B. This should be easy (look at how
|
|
|
|
other backends do it) but is not really required yet.
|
|
|
|
-> not required
|