[huebi]

I#m doing this with

s: ([a-zA-Z])</p>\s*<p>
r: \1##
o: minimal matching

The reason for the two #: its more easy to find out if a word is splitted or the sentence is splitted between two words.

Now i'm looking for a lower case letter, followed by ##, followed by an uppercase letter. This is for sure a sign for two seperate words.

s: ([a-z])##([A-Z])
r: \1_\2
o: minimal matching, match case

(the underscore represents a blank)

Now i'm going thru looking for two separate words:

s: ([a-zA-Z])##([a-zA-Z])
r: \1_\2
o: minimal matching, match

If the ## is between two words, i press alt-r (replace), otherwise its a splitted word and i'm skipping this with alt-f (find next)

At the end only the ## splitting a word are remaining and i'm substituting them with

s: ([a-zA-Z])##([a-zA-Z])
r: \1\2
o: minimal matching

Last step is searching for a comma, question mark or exlamation mark:

s: ([,?!])</p>\s*<p>
r: \1_
o: minimal matching

More steps then yours, but i'm replacing the text without changing it manually which can be really annoying.