Section 104 could have been… and is, worse.
“Mmm, not really. Look things aren’t good with the Nonsense Filter. Some of the results are coming back from those journalists who installed a copy from that CD. It seems Émile hooked up some kind of Bayesian algorithm which he stripped out of an open source spam filtering programme and linked up with google via gmail. He tied it in a network of search engines and online dictionaries, and fed in the notes you provided regarding the Jabberwocky. So far one journalist’s complete digital archive has been erased and his computer only accepts rhyming nonsense, another has sent millions of garbled versions of the Jabberwocky to all his contacts.”
Even though he was still wakening it seemed to Vernon that it could have been worse. He said so. “Could have been worse Jean Luc.”
“It is.” was the terse reply.
Vernon scrabbled around for his glasses and then scratched his beard with puzzlement, “It worked for the others though didn’t it? You know… it recognised the data keyed in as sense or not sense? They could say ‘this is meaningful if I research it’ and ‘this isn’t worth researching’? Tell me he got that much right.”
Jean Luc sighed mournfully. “All I can tell you Vernon is that I am worried about whether Émile stayed this side of the law; I think he went too far this time. You can’t recognise sarcasm and emotion with statistics, you can’t discern what something really means to each user’s context and the Nonsense Filter’s never going to give meaning to something newly encountered the way that woman wanted? I think Émile knew that but had his own agenda.”
Vernon deduced that Jean Luc must have been up a lot longer than he had by the extensive vocabulary at his command at such an unsociable hour. In fact the technical jargon continued to flow jarringly down the line and over Vernon’s head.
“Émile had joked about strapping up some spam filter derived Bayesian-inspired classifier to a search engine. He reckoned nonsense words could be sieved out by employing a switch option in the parser that recognises and sorts bigrams and trigrams. Something fancy like identifying contextual polarity in phrase-level sentiment and idiosyncratic term analysis.” Jean Luc was beginning to sound like an Asimovian robot with a serious language malfunction.
“Hold it Jean Luc, stop, stop. Tell me, was it only three journalists who found it didn’t work?”
Suddenly Vernon realised he had never heard Jean Luc sound so distressed. Clearly the rush of techno-speak was the result of an unconscious pressure release mechanism opening to disperse a rising tide of panic.
“Nn-no Vernon, that’s not all by a long way. Most of the journalists have begun to confirm a suspicion that their computers have been hacked and their financial details phished. Some have confirmed withdrawals from their accounts. The Nonsense Filter is a Trojan horse. Don’t ask me how that came to light. But it has.”
Vernon sunk to the floor in dismay. “This is a disaster. What does Émile have to say about these allegations?”
“Jean Luc struggled to make a reply. “I don’t know Vernon; he’s gone missing.”
Check out the build up to this crisis on The Novel page.