Re: Boost.Text Unicode licence issues

From: Andrey Semashev (andrey.semashev_at_[hidden])
Date: 2020-08-28 17:18:17

On 2020-08-28 19:29, Janek Kozicki via Boost wrote:
> Mike via Boost said:     (by the date of Fri, 28 Aug 2020 15:12:50 +0200)
> 
>>> The actual information that UTF *has* is which natural number
>>> corresponds to which glyph. Ask about this.
>>
>> Whom
> 
> My apologies, clearly it was another person who wanted to ask (or
> already asked) the boost lawyers and the UTF consortium lawyers.
> 
>> should I ask what?
> 
> If this "number" ↔ "glyph" correspondence can be used by boost on the
> terms of boost license.

The license applies not to the abstract coding of glyphs or characters 
but to the data files published by Unicode Inc. that describe this. It 
is these files Zach translates into bits of Boost.Text code, which is 
then claimed to be not covered by Unicode Inc. license.

The important difference is that (a) Zach is using a material document 
that is licenseable (as opposed to abstract facts or numbers, which are 
not) as a source and (b) the result of translation preserves the 
original information in some meaningful amount (as opposed to e.g. 
applying a hash function or `wc`, which does not).