[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: SaneWidget
From:       Kåre_Särs <kare.sars () kolumbus ! fi>
Date:       2007-03-27 11:25:35
Message-ID: 200703271425.36388.kare.sars () kolumbus ! fi
[Download RAW message or body]

On Tuesday 27 March 2007, Aaron J. Seigo wrote:
> can you rework the code without changing the public API?
If I would only rework libkscan I guess it could work, but the SaneWidget code 
does not have a kpart and I would like the API to be very simple:

- Create the Scan object (*parent_widget)

- Function to get available devices

- Create the Interface for a device (vendor, model, devicename) 
  (return false if device does not exist) 

- signals for scan_start, progress and image_ready

- function to get the scanned image

Would there be a need for 16bits/color support? 

> the only problems we have are:
>
> a) you've got one week to introduce new subsystems into kdelibs for kde4
I guess it might be too late for "SaneWidget", or whatever it would be called, 
to be included right now, but is it possible to include it later? (kde4.1?)

> b) we'd have to port the apps currently using libkscan to your new API (or
> is it the same as libkscan?)
Do you know how many they are?

> c) libkscan is at least somewhat proven, if needing love =)
Yes that is true. I only have an Epson scanner, so that is the only one I can 
test on right now.

>
> > > kooka will probably not be making it into kde4, and really the only
> > > thing that needs to go into libkscan from kooka is the OCR stuff. that
> > > job consists essentially of ripping code out of kooka, cleaning it up a
> > > bit and putting it in kscan.
> >
> > I haven't checked the OCR code, but it is using external OCR engines
> > (gocr, ocrad and kadmos).
>
> that's correct; it's just a convenient bridge to these engines which are
> all command line driven (sort of what we're trying to avoid by creating a
> gui ;)

Would it be good to make a separate library for the OCR stuff? Scanning and 
OCR don't have too much in common code wise.


Would there be a need for 16bits/color support? 


/Kåre
 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<

[prev in list] [next in list] [prev in thread] [next in thread]