Today, I fully realized the potential of programming in q.Up until now, I had been thinking in parentheses, because for years, Ihave modeled time-series in spreadsheets, like Gnumeric and Excel.When I woke up this morning, things were different.I suddenly realized that I was now thinking in vectors.So far, I have managed to build both a real-time and a bulk-data time-series model using parentheses thinking.Under that framework, my bulk-data model was dreaming-up 10,000 time-series prices in around 110 milliseconds.Re-organizing the bulk-data engine along vector thinking lines, it nowspitsout 10,000 time-series prices in 10 milliseconds.That's a 90.09% reduction in processing time.Now I can't stop laughing... mostly because the vector model uses justone line of production code.As a consequence, I am going to pull out all stops. I am thinking big,I am thinking that I might even use two lines of production code.However, that seems to be going in the wrong direction, because Irecently found a quote in a 1995 style-paper on Stevan Apter's sitewhich said: 'The more K you know, the less code you will need towrite.'That is certainly the truth. One month ago, I knew nothing. Today,Iknow very much more.In his book titled 'Art of Assembly', Randall Hyde, ...when discussingthe use of NAND gates and state-machines,.... makes the observationthat 'there is a one to one relationship between boolean functions andelectronic circuits'.He then goes on to say...'you can implement all hardware functions insoftware'...It seems that Arthur W. has realized this prophecy, by turning a tinysoftware code base into something as powerful as a vector-processingcomputer.As a result of that effort, q is so blazingly fast, it's like having aCray in a box... underneath the desk.Regards,Q.
