View Single Post
Old 01-10-2013, 08:45 PM   #20
roger64
Wizard
roger64 ought to be getting tired of karma fortunes by now.roger64 ought to be getting tired of karma fortunes by now.roger64 ought to be getting tired of karma fortunes by now.roger64 ought to be getting tired of karma fortunes by now.roger64 ought to be getting tired of karma fortunes by now.roger64 ought to be getting tired of karma fortunes by now.roger64 ought to be getting tired of karma fortunes by now.roger64 ought to be getting tired of karma fortunes by now.roger64 ought to be getting tired of karma fortunes by now.roger64 ought to be getting tired of karma fortunes by now.roger64 ought to be getting tired of karma fortunes by now.
 
Posts: 1,396
Karma: 846401
Join Date: Jan 2009
Device: KoboGlo
Quote:
Originally Posted by mrmikel View Post
That is why I requested the spellcheck to be able to ignore capitalized words.
This is an OpenOffice method I did try two days ago and which maybe can be adapted to EPUB spell-checking.

I just did a successful try with OpenOffice on a 1000 page book including a lot of proper names I had never met. The spell-checking was excruciatingly slow because of these proper names. OO proposals at times were quite funny, but frequently way off the mark.

The OO spell-checker has an option: "ignore all words fully in uppercase".

To perform a quicker spell-checking, I did this:

1. Transform all the words beginning with a capital into words fully in uppercase.
2. Spell-check the book with the option "ignore words fully in uppercase".
3. Convert back the words fully in uppercase into words beginning with a capital.

You take a small risk: you exclude words beginning with a capital. Among them are also some common words (the one which begin a sentence). This is your choice. Also, you may have to write down existing acronyms (CIA, KGB) to restore them later in their former glory. ;-)

Method

This method works with OO. I did not try to adapt it on EPUB.

1. Regex. Option to select: take the case into account.
Look for:
Code:
\<[:upper:]{1}[a-z]{1,}\>
All the concerned words are selected.
Then: Format: change the case: Word fully in capitals.

2. Spellcheck: Like usual but much much quicker.

3. Regex. Option to select: take the case into account.
Look for :
Code:
\<[:upper:]{2,}\>
All the concerned words are selected.
Then: Format: change the case : Capital with each word

Result

Most of the words beginning with a capital were changed into words fully in uppercase. The ones which were not included accented lowercase (like Psamnétique), hyphens (SAINT-Césaire), and apostrophes (T'ang, L'Encyclopedia).
I did not try to perfect the first regex to get a 100% transform result because I was afraid I would lose some of them on the return process. Doing this I got back all of them. But of course, it's safer to save your file before...

This way, I was able to peruse the book much more quickly and managed to correct some common words which I would surely have left over not using this method. I believe it can be an acceptable solution for big books containing a lot of new proper names.

Last edited by roger64; 01-10-2013 at 11:32 PM.
roger64 is offline   Reply With Quote