Monday, February 15, 2010

Internet authenticity

Ever feel frustrated when reading a page on the internet and not knowing if it's authentic?  I certainly do.  Google's "Page rank" (named after Larry rather than the more obvious web page) helps somewhat with this because the more other sites that connect to your site, the higher your rank, but in and of itself that proves nothing.  For example, a creationist web might have a relatively high page rank because it is referenced by many other sites (there are lots of them, by the way).  But that's not the same thing as being peer reviewed.  Even Wikipedia can be in error, especially in the short term.  I could go in right now and change the date of William the Conqueror's invasion of England to be 1067.  It would be reverted, and if I was silly enough to do it using my Wikipedia account, I'd earn a black mark in some electronic book somewhere.  But some number of visitors might actually believe the 1067 date while it was up.

Indeed, this very ploy was famously perpetrated a year ago today regarding Titian (see article in the Daily Telegraph).

Yet, this problem could be remedied so simply.  Here's how it would work.

An author called Archie publishes a web page about some topic, let's say the Soca (Isonzo) front of World War I.  Archie calls Bert and asks him to peer-review it.  Bert reads it and publishes on his own web site a review of the article, with a link to the article.  Archie adds a link to the review on his web site.  But here's the neat part: the link from Bert's site to the article is qualified by a long string of characters: the message digest of the article as it was when Bert read it.  If Archie goes in and updates his article, the link is no longer valid.  Thus anyone, say Charlie, reading the (original) article can visit Bert's site and read the review with confidence that it was the same written about the same article.  Of course, the authenticity of the review is now based on Bert's reputation.  Charlie may want to see if respected people (especially if they are not Archie) have written favorable reviews of Bert's works.  That's the essence of the peer review system.  The authenticity of an article rests on the reputation of not only the author, but also those of the reviewers.

All we need is a browser which (or browser plug-in) which will inform us whether or not the link is was intended for the current version of the article.  Simple!  I think I'd better write one.

No comments:

Post a Comment