Word count discrepancy between Nisus Writer Pro and Word

[samwa]

I have just finished writing a book in Nisus Writer Pro, and unfortunately had to convert it to Word. After I did, I noted something surprising; the word count, which according to Text Analysis is something over 137,00 words, is just over 133,00 according to Word. I don't think anything was lost in the conversion. Any idea why these are so different, and which is more likely to be right ?

[Roger]

I just did a quick test and discovered at least a partial answer:

1. Nisus counts hyphenated-words as two words.

2. MS Word counts hyphenated-words as one word.

[samwa]

Thanks ! That looks like a good chunk of it. I doubt 2% of all the words are hyphenated, but there's a hyphenated name that appears frequently, so that no doubt contributes.

[martin]

How words are counted isn't an exact science, so different software can produce different results. As Roger mentioned, hyphenation can affect this. We have had requests that we add an option which allows users to choose whether hyphenated words are counted as one or two words. Even then it's tricky, because something like "high-risk" might best be considered 2 words, but "re-edit" just 1. Languages are messy beasts.

If you don't mind sending your document to us privately, we could check and see what else can be done to harmonize the word count between Nisus Writer and Word. Although perfect agreement may never be attainable (or desired), we prefer that our users not have any unwanted surprises.

[samwa]

Thanks, but it isn't that important. There's even a certain advantage to having two values to choose from. Thanks for your help !

[Þorvarður]

Martin wrote:

We have had requests that we add an option which allows users to choose whether hyphenated words are counted as one or two words.

If I’m not completely mistaken, this option is already included in Nisus, namely in the Find/Replace box.

Samwa wrote:

Any idea why these are so different, and which is more likely to be right ?

You can use the Find/Replace box to custom tailor the word count, i.e. tell Nisus exactly how a word should look like and where word boundaries should be. Try the macro below and see if there is still much difference.

A few things worth having in mind when counting words:
The Statistics palette and the Find/Replace box count differently. The Statistics palette counts an hyphenated word as 2 words, but delivers otherwise correct results. — Now, in case one wants an hyphenated word to be counted as 1 word, one must create a search criterion in the Find/Replace box that looks explicitly for hyphenated words; but this is not enough, because for some reason the Find/Replace box interprets an apostrophe as the end of a word. The phrase "Man's best friend" is therefore counted as 4 words if we use the Find/Replace box without any extra criteria.

1.png (57.96 KiB) Viewed 9055 times

The following search criteria should resolve the problem.

2.png (94.98 KiB) Viewed 9055 times

The third apostrophe in the search box captures (additionally) smart punctuation, in case the user has turned on "smart punctuation" in Preferences > QuickFix. The first apostrophe is what we may sometimes get when we copy text from the web.

[martin]

If I’m not completely mistaken, this option is already included in Nisus, namely in the Find/Replace box.

Ahh that's a clever way to get yet another word count! Thanks for sharing Þorvarður