I know for one thing that when I scan into PDF format it scans twice over. 1 to scan all the text and 1 to scan all the graphics seperately.
If you can give a way to quickly add scanning ability to an alternative format which can make use of compression which is good for the graphics commonly found in manuals (black/white) it'd be great.
Problem is that scanning it as JPGs takes up way too much space because they are so damn large.
The PDF though slow don't need to be cropped when you specify what to scan initially so it doesn't scan too much. Thus the flaw should be circumvented. And PDF is one of the better ways to faithfully recreate a scanned document without losing something. JPG has artifacting which would be negative though. HTML though a good alternative would require all too much work and not work anyway because of border problems. You could set it as background to make a border but it'd take up a lot of work time to get it aligned right. Something I for one wouldn't want to do with every single page of a manual.
You might have a manual of some sort ECA. How about you giving a demonstration of the miracle method you want? Scan a page into JPG and then scan a page into what you want it and convert it to HTML. That'd make for easy comparison between the original manual and the recreation you're able to do. And do remember to time yourself. A PDF scanning takes roughly 1 minute per page.