so here i am, home from new orleans, feeling like a little unfiltered stream of consciousness to end the day.

yesterday i woke in chapel hill and flew to new orleans for lunch. did some work and had an amazing dinner, duck appetizer in crepes, salad, and the most wonderful duck for dinner. with a beer before dinner at the bar and a half bottle of wine with, my night was complete. well, almost complete. can’t forget the obligatory stroll down bourbon street, with obligatory beads and obligatory onlookers. we can now say trip #4 to the big easy is complete.

today some work, some lunch, and then some riding in the plane. unlike most trips where i feverishly try to make database improvements up until the last possible moment, during this short trip i pretty much sat in meetings and listened, voicing an occasionally cautious note of “maybe we shouldn’t have me do anything until we figure out what it is we want me to do…” fun though watching the process of telling faculty and researchers things like: we need you to find a way to spend $750,000 quickly.

it does seem there is a pressing need for greater access to and centralization of all manner of information, with the web being the most obvious candidate for extending access to people outside the office. unfortunately microsoft access makes doing that a wee bit tricky (which is good for microsoft, forces most people to purchase SQL Server) but bad for us techy chaps who’d rather roll our own system out of whatever else seems appropriate and free.

more and more i’m questioning the viability of the traditional table data structure and thinking more in terms of ‘documents’ as the starting point. it seems to me that something is lost when we take what start out as documents and chop them up to fit in a database and suddenly have limited powers for reconstituting the original document (as a document). maybe if we left the document as a document, impose some rigorous structure (xml?) and provide some way to query a collection of documents (xquery?) as if a database, we could forgo using a database in some cases altogether.

oh, and my last thought: user interfaces. access is nice, but really it mostly sucks. the web is insufficient, but… mozilla is not. if rather than using $300/per user access as a user-interface client to access data in $0 mysql database with $0 myodbc so that $0 php can spit the data out to web, maybe we just ought to use $0 mozilla and its XUL+CSS+Javascript technology to replace the access user interface. i wonder how feasible/powerful that is? glad i bought the $40 mozilla book by oreilly (which is also available online for $0) to read for fun.