The problem with using an analogy to prove a point is that the analogy is nearly always a poor fit for the actual situation.

In this case, I think a better (still flawed) analogy would be where "TeacherSoft" records only everything the students say - so if they're having a lesson about "Joe the Novelist", TeacherSoft sees some words and snippets of information from Joe's work, including when the students quote back a sentence of two, but not all of Joe's work.

This is closer, IMHO, to Bing seeing the URLs of Google's search pages when a user visits them, and which links the user chooses to click, but not the search pages themselves.[1] But it's still a poor analogy, IMHO.

[1] If you don't know what precisely Bing gets sent, it appears to be this: http://projectgus.com/2011/02/bing-google-finding-some-facts...

Especially since copyright law has exemptions specifically for teacher's reading copyrighted works to their class.

His entire argument seems based on his assertion that Google has copyright to search results. I think they probably don't, for two reasons: first, there's no copyright notice on search result pages. Second, a search result is probably a database under copyright law, and databases (like the phone book) can't be copyrighted.

> first, there's no copyright notice on search result pages

Irrelevant. There is no legal requirement in most countries to put a copyright notice on a copyrighted work.

> Second, a search result is probably a database under copyright law, and databases (like the phone book) can't be copyrighted.

Databases can be copyrighted as compilations.

The typical phone book is not copyrightable because it is insufficiently creative, not because it is essentially a database.

Except that first, you don't have to explicitly show a copyright notice for something to be copyrighted. By default, a work is copyrighted. You can explicitly give up a copyright, but not putting a notice is not that.

Secondly, and I don't really know the 'state of play' of 'database copyright' across various countries but one thing is for sure: a phone book is a database of facts. Search engine results are a database of 'creativity'. The first link you see when you Google something is not 'the best result' as a fact. If it were a fact then surely Bing, for example, would know, for a fact, that that is the best result for that search. ie. they would also return the same 'fact'.

Note: the results of a search are not a database per say, they are an interpretation of a database. At risks of falling down the analogy hole again, it's somewhat like an artist's work is an interpretation of the material he uses. He doesn't copyright the clay, but he inherently owns copyright of what he produced from it. Like I feel Google does with their search results of a database of websites (as does DuckDuckGo, Bing, etc.)

Also, that there's no transmission of the results page. There's a URL with a search query in it, and any URLs that the user clicks on from that page.

So any copyright would have to apply to the string showing in the user's browser address bar. Not quite as clear-cut.

Seems that under copyright law, the teachers who set up the recording equipment in their rooms and then read aloud copyrighted books would be at fault for creating copies (or at least aiding in the creation of copies) of copyrighted material. The TeacherSoft company would just be the recipient of these copies; they might be required to delete them, but the primary party guilty of copyright infringement here would be the teacher who read it aloud.

one problem with the author's premise:

IN the example he gave the copyright infringement occurs because audio performance is recognized as copyright infringement of a copyright of a text. And it comes about because that audio performance is replayed to many individuals.

We do not know yet if a click-stream of copyrighted work is in fact copyright-able yet..Hence the debate at hand