The IP angle is interesting.

But consider another angle.

Anyone can crawl the web and create and index. Even a cache. But some questions begin to emerge:

- are they holding the IP (e.g. copyright) of others’ sites as URI’s or even cahed pages?

- are they operating as middlemen, as gateways to others’ content, and letting advertisers intrude into the pathway between user and the content she intends to retrieve?

- will content owners agree to anyone crawling and indexing and even caching their content, or only to certain parties (e.g. Google, Yahoo, MSN, etc.)

- do we need middlemen?

- are “browsers” even necessary, if we are merely searching and selecting known content from an index? consider a tool that can handle regex searches, a tool that can generate an index and a tool that can retrieve a file when given a URI

is the internet of today (cf. to that of say 1993) too big to “browse”? compare the “portal” concept (e.g. Yahoo! etc.) of the early web versus today’s point of entry: the Google search

To the extent that the subject “Google Technology” is broadly conceived, there are several subjects of broad interest that I hope will be covered.

For example, what is (or should be) the intellectual property status of algorithms like PageRank?

According to Wikipedia, Stanford University holds a patent of PageRank algorithm, which they exclusively license to Google … for the princely sum of $350 million dollars.

We can contrast this the charter of UC Berkeley’s CHESS program (Center for Hybrid and Embedded Software Systems) which explicitly states “The Center publicly disseminates all research results as rapidly as possible … software is made available in open-source form … center patents are expected to be rare.”

A wonderful question is: which of these two models is a more plausible/more viable/morally superior model for the future of science?

On a mathematical note, if we take “n” to be the number of vertices of the web, then every aspect of Google’s page-rank algorithms has to be O(n) … even matrix-vector multiplication. So every matrix arising in PageRank has to be structured in such a way that (e.g., sparsity and/or factorization) that practical O(n) computation is feasible.

The property of “O(n)-ness” is becoming ubiquitous in every aspect of modern simulation science (which itself would be a good theme for a lecture). Because nowadays it seems that almost every large-scale system that we humans care about–whether natural or human-made, whether classical or quantum—possesses sufficient algorithmic compressibility that “O(n)-ness” can be found.

In this sense, aren’t Google and CHESS (and many other public and private enterprises) very much in the same line of business … mathematically *and* morally *and* legally?

These are deep waters … and for this reason I look forward immensely to further lectures in this series.

http://benhur.teluq.uqam.ca/SPIP/inf6460/rubrique.php3?id_rubrique=13&id_article=52&sem=Semaine%2011