Anyone know of a word frequency counter?

[gtodell]

Does anyone know of a program that can count the frequency of a particular word contained within the html (within metatags, etc) of an entire web site (online or offline)?

I have seen several programs that can do this one page at a time but I need something that can perform this task for an entire site.

Many thanks in advance.

Gordon

[tedster]

Homesite does this -- every document in a folder, including or excluding sub-folders as you choose. I mention it because some people may have it handy. But it would be an outlandish expense to purchase it for just this function.

In general, I'd suggest looking for programs that can do "extended search and replace" or "search multple files" -- as part of the core function, they often report on how many instances of the search term have been found.

[gtodell]

Are you referring to Macromedia's Homesite? I thought the feature I need would only work for local web sites, not live sites that are not mine. Let me know if I am mistaken. FrontPage, which I already own, will do the counting I need but, again, only locally, or lives it is my site.

[tedster]

You are correct -- offline only. I didn't understand that you wanted both functionalities, I was hypnotized by the "or".

An online check could be a tough program to write -- unless you can assume that all pages and folder have a link trail from the home page. But assuming that is the case, then why not rip all the html files down to a local copy (essentially that's what must happen anyway) and then run Homesite or FP over the downloaded files.

[DrDoc]

Why do that, even?
Use Homesite or FP (or whatever you have) to count the word frequency locally. Then use Google to check how many times that word appears on live sites. For example, if you want to know how many times the fabulous word DrDoc appears here on WebmasterWorld:

site:webmasterworld.com DrDoc

;)

Or, if you have access to all the sites, just run a nifty shell command ... like

grep

. It will tell you the number.

[henry0]

G search is a good tip.
However, correct me if I am wrong, wouldn't that query only ranked pages?

[tedster]

Yes, the Google idea is theoretically good, but it would only work for fully indexed pages. And I'm not also convinced of the accuracy of that number -- Google does say "about" 58,500 results. In fact, I've seen the number change as you click through to a deeper page of the results, possibly because of a change in data center.

[gtodell]

DrDoc,

Does grep (or other method) search WITHIN the html of the pages? This is critical for my purposes.

Thanks.

[gtodell]

DrDoc,

Got it! Windows Grep 2.3 does what I need - at least locally.

Thanks for the tip.