[TxMt] Re: Writing Bundle [was: Lookup current word in Dictionary.app]

[Jacob Rus]

Mark Eli Kalderon wrote:
>> 5. Some commands for getting better statistics than the current 
>> document statistics count.  It would be nice to have a word count 
>> which knows how to ignore stuff in html/latex/markdown tags, etc. etc. 
>> (or maybe bundles can provide overrides to this command, and all call 
>> out to a single script), but it would also be nice to be able to get 
>> some readability statistics, such as counts of average word length, 
>> average sentence length, and maybe metrics like Flesch-Kincaid, etc.
> 
>> 6. It might even be nice to add some tools for checking grammar 
>> (flagging things like wordy sentences, etc.).  There are some decent 
>> open-source programs for this, I believe.
> 
> Jacob, you might have a look at diction, a descendant of the Writer's 
> Workbench:
> 
> http://directory.fsf.org/GNU/diction.html

Yes indeed, I was looking at that page when I wrote the post :).  When I 
actually have some time to make this thing, I think I will indeed use 
style/diction.

> and `diction` checks for grammar (although it perpetuates some of the 
> unfortunate aspects of Strunk and White's prescriptions--- useful 
> nonetheless).

I didn't look at exactly what it did, but I assume it can't be worse 
than MS Office's grammar check.

> As for filtering LaTeX commands, I run the generated pdf through 
> ps2ascii before running these. Won't work for html documents, but you 
> could run them through a text browser like lynx or links with the dump 
> option instead.

Well, I'm still thinking about the best way to get things to work for 
multiple document types.  It has to do a few things: a) strip out 
extraneous junk, and b) figure out how to get back to the specific 
places in the document where there are points of interest.