A way to deal with hyphenated words’ counts

[valente.mac]

I’ve been trying to find a way to work around the way Scrivener stats deal with hyphenated words (most of the times counting them as 2 or more words, according to the number of hyphens). When writing in English this doesn’t pose such a big problem, but in Portuguese many of our verb tenses have hyphens and the way Scrivener counts them became problematic (specially when dealing with scientific articles, that many times stipulate a maximum of words).

On this matter, Keith said:

Scrivener’s word count uses Apple’s code for determining what a word is. When you double-click in the text of a program using Apple’s text engine, the word gets selected. That is what the Apple text engine determines as a word - Scrivener uses that for counting words. It is fairly accurate - at least as accurate as similar programs. I did a check with a document containing hundreds of thousands of words against Word, Nisus, Mellel and others some time ago. The word count in Scrivener came out very close to Nisus and Word, with Mellel being a little different to all, I believe. (No application had exactly the same word count, because of the subtleties involved in determining what a word is.)

Ideally, I would love to have a preference somewhere (in Scrivener?) where I could choose to count hyphenated words as just one. But, since that doesn’t seem to be possible, I had to find a way to work around it.

So far that has been WordService from DEVONtechnologies (the same folks of DEVONthink). I included the Statistics… service on my Services menu and gave it a shortcut (see attached image). What I now do is: when needed to check the word count in a Portuguese text, I’ll select the text and hit the shortcut. A little window will popup showing the Word Services’ stats, which gives the hyphenated words as one. It works quite well and it’s fast.

Click for large view

I am, however, wondering if someone has another — equally or more efficient — way to deal with hyphenated words’ counts.

– MJ

[AmberV]

This place where this should be set is in Apple’s Language & Text system preference pane, under the Text tab. There is a drop-down for specifying the language model to use for word breaks, but only Japanese, Greek, and “Computer” are listed. The latter makes it so some strange constructions like “one:word” are counted as one word, and is mostly useful for programmers and web developers. Unfortunately it still differentiates hyphenated words into two counts.