[sword-devel] Neural Networks and Optical Character Recognition

[Chris Little]

DM Smith wrote:
> My 2 cents (personal opinion) as one who is not a lawyer:
> For a work that is not under copyright, the text is public domain. But  
> anyone can fairly claim that their markup of the text is copyrighted.  
> For example, in the KJV work that we did, we marked it up. We can  
> legitimately claim the copyright for the osis elements, but not for  
> the actual KJV content. The Strong's numbering of the NT was the  
> result of volunteer effort on behalf of CrossWire. And legitimately,  
> CrossWire can claim copyright for that too.

As Jerry mentions, US copyright is only concerned with creative 
expression. The relevant publication that explains what can be covered 
in derivative works is Copyright Circular 14: 
http://www.copyright.gov/circs/circ14.html.

So whether markup is covered by copyright depends on whether it is 
creative. If you're converting a printed text to HTML and you use tags 
like <p>, <b>, <i>, etc. to approximate the printed presentation, those 
tags are in no way creative or copyrightable. With our KJV text, we had 
two components: automatic tagging by scripts and human 
tagging/correction. The automatic tagging is definitely not creative. 
And I think it was our intent that the human tagging accurately reflect 
the history of the text rather than be some sort of creative expression. 
The last point is certainly debatable, but I am of the opinion that 
nothing about our KJV is copyrightable (except under Crown Copyright in 
the UK).

The same goes for ALL of those Bibles from Importantia where they claim 
copyright on the Strong's numbers (but not the text). You can't 
copyright Strong's numbers tagging if you're being faithful--only if 
you're making it up.

> If one were to "enhance" a work that is no longer copyrighted, one  
> could legitimately claim a copyright for the changes. And without  
> identifying the changes one could claim a copyright for the work as a  
> whole. I have seen copyright claims on the KJV (just visit a book  
> store and look for copyright info on the publications) and have heard  
> that publishers deliberately introduce changes that they are able to  
> identify as their own. I don't know if it is true or just hearsay.

Watermarking can help you know where a text came from. You could use it 
to demonstrate a violation of license terms, but it doesn't make a 
public domain work into a copyrightable derivative work. The 
watermarking itself isn't even copyrightable because (again) it's not 
creative.

--Chris