Randomize

Richard Tallent’s occasional blog

Questions I’ve been asking Wolfram|Alpha for my own amusement…

http://www54.wolframalpha.com/input/?i=what+is+the+answer+to+life%2C+the+universe%2C+and+everything%3F http://www54.wolframalpha.com/input/?i=how+old+are+you http://www54.wolframalpha.com/input/?i=What+is+the+airspeed+velocity+of+an+unladen+swallow%3F http://www54.wolframalpha.com/input/?i=how+are+you%3F My juvenile sense of humor at work, I suppose… Google should just buy these guys and integrate the search results, I can’t imagine going to this search engine to do actual work. Read more →

Rube Goldberg and Scripting Languages

There’s a really cool diff class hosted by Google here: http://code.google.com/p/google-diff-match-patch/ I’ve used it on a few web sites where it is necessary to compare versions of the same text. But now I want to integrate that functionality into Excel 2003. Unfortunately, Excel 2003 only understands VBA, not JavaScript. So, here’s the convoluted path I’m exploring: – Have my users install ExcelDna (http://www.codeplex.com/exceldna), an open-source Excel add-in that allows execution of . Read more →

The Healthcare Data Model

As a systems architect in a non-medical but equally complicated field, I’ve been incredulous about Obama’s plan for a “paperless” healthcare system. I agree there’s a huge *need* for such a change, but throwing $20 billion at this without a cohesive, common, open data model would be a complete waste of money. Turns out I’m not alone in my opinion, and this guy knows a lot more about the problem than I do–he almost died from it: Read more →

Natural Keys are a Fallacy

Originally posted 2004-02-18, part of my slow effort to convert old dasBlog content to WordPress. On the big argument going around about the use of identity() fields as surrogate primary keys, I’ll just make a few comments from my own experience. My overall opinion: in a real-world system, your list of candidate keys is constantly being reduced, and composition required to generate uniqueness is constantly increasing. If you start using a natural candidate key, you might as well plan on redefining the PK at some point in the future, period. Read more →

Keeping Your Sanity When Using Lookup Tables

Note: Originally posted on 2004-02-17 under the permalink 4aa1a488-fa83-4df5-83e8-a2b218e75d75… someone requested that this one be resurrected from the archives. Someday I’ll write a script to convert all that old content to WordPress! *sigh* Marc Shiker has seen the light on how to address the issue of lookup tables. The problem here is not one of identities primary key vs. storing the value, it is one of whether you should have a relation in the first place. Read more →