[WikiEN-l] Non-free images, there has to be a better way

[Andrew Gray]

On 07/07/05, Ray Saintonge <saintonge at telus.net> wrote:

> AFAIK we have not yet received any official take-down orders.  That's
> disturbing in its own peculiar way.  Receiving a few of these gives us
> an idea of which way the wind is blowing.  Without them we are only
> speculating about what might happen.  Until we receive such notices it
> should be enough to be reasonable.  Winnability should not enter into
> the discussion until after the notice.  So far, case reports don't give
> a conclusive picture, but some decisions have been favorable.  I'm
> probably more optimistic than you.

This is quite an interesting comment. Has there been much in the way
of informal contact about such matters? (an email saying "we're a
little concerned about X" and it getting pulled or an agreement
reached, &c)

> >There are more images labeled fair use than I can count --
> >1,000 deep and I am still in the A's. Anybody have a better way to
> >estimate it?
> >
> Find the number of total pages considered and extrapolate.  The "A"s are
> probably better covered than most letters. 

For what it's worth, the English-language distribution for "A" and "B"
is such that the two of them, together, make up about 1/13 of the
characters in use - in other words, they combine to give a good
"average letter" value.

However, B is very over-represented as an initial letter - it's only
the twentieth most common letter overall, but it's the fifth most
common initial letter. A is third in both ranks.

Your best bet, for a fair representation, would be to count those in
"T" and in "K", and then multiply by thirteen.

(T and K together make up about 1/13th of the letters in the language
- but T is the most common initial, and K the least, so that should
hopefully average out)

-- 
- Andrew Gray
  andrew.gray at dunelm.org.uk