# [OS X TeX] counting words in 2010

cfrees at imapmail.org cfrees at imapmail.org
Sat Oct 30 23:16:13 CEST 2010

On Sat 30th Oct, 2010 at 13:34, Michael Sharpe seems to have written:

>
> On Oct 30, 2010, at 12:46 PM, <cfrees at imapmail.org> <cfrees at imapmail.org> wrote:
>
>> Strange. detex is definitely taking out the contents of \emph{} here.
>> If I run detex without piping through wc, I can see the relevant words
>> are missing. (This is without passing detex any options.) Maybe we are
>> using different versions of detex? Earlier versions definitely did not
>> behave in this way.
>
> I'm using the intel 64-bit version of detex dated 7/13/10,
> though I get the same results if I use the universal 32-bit version of detex dated 6/16/10.

My version is dated 17/6/2010. I'm definitely using the universal
32-bit version or it wouldn't be working at all.

> In both cases, the detex output from the line
>
> Test\footnote{ a footnote}  \emph{it} quickly.
>
> is
>
> Test  it quickly.

I can't now reproduce the disappearance of \emph text. Not even using
the same document. (I've edited it but not any of the TeX stuff.)

However, when I put your sample sentence into a test file and run
detex, I get:
Test  a footnote  it quickly.
Although footnotes are still being deleted from my paper when run
through detex.

So now I'm just confused and have no idea what's going on. I get the
following results:
detex 2010 + wc: 208    3088   19174
detex 2009 + wc: 192    3215   20003
detex 2008 + wc: 192    3215   20003
texcount:
Words in text: 3115
Words in float captions: 121
Number of floats: 0
Number of math inlines: 8
Number of math displayed: 0
All of these are being run without any customisation. I'm not surprised
to get a different result from texcount, of course. (Though I'd like to
know which way of counting is most accurate!) But I'm curious about the
different results from different versions of detex. The documentation
doesn't seem to be any different. I downloaded and compiled the source
and get the same results as those for 2010 above. That version is 2.8
but the documentation still refers to 2.6. So maybe changes made to 2.7
are responsible for the different results? (Versions 2.7 and 2.8 are
identical aside from the licence, I believe.)

Thanks,
Clea

> Michael