If you do the bar code development for them feel free to drop me a note if you run into any issues. They're actually not that difficult to resolve but it would have been helpful (for me) to know going into the project that the "logical" approach wasn't going to be adequate to compensate for speed variation.

The only way I was able to get good scans out of the original method I use (calculating a single average bad/space width) was if I got a "running start" on the scan. That is, if you start scanning a couple inches to the left of the barcode to get up to speed and then just carry on through the bar code and continue about an inch or two to the right before even thinking about stopping. But inevitably the user (myself included) puts the scanner just to the left of the barcode so is still accelerating as he moves it over the first portion of the barcode. The middle of the barcode is the closest there is to "constant speed" and usually the user is subconciously decelerating before he even finishes scanning the far right side of the bar code.

We humans can be very frustrating to software development! :)

BTW, if your application is for the blind then I suspect they'd be using a gun-type scanner and not the type of scanner that you have to manually run a "pen" over the bar code. So the speed variation issue might not even apply to you.

Regards,
Craig Steiner