For the hundredth time I contemplated the challenges of managing personal data with general purpose database management tools (FileMaker Pro for OS X and XP, Microsoft Access for XP) and special purpose database management tools (Address Book and iCal for OS X, Google Contacts and Calendar, Outlook for XP, 1Password for iPhone and OS X).
It’s really an intractable mess. There’s no practical way to move data between these stores (tab delimited files were never enough), much less synchronize data. Everything is in proprietary data stores – fundamentally hostile to data freedom.
Then a little bulb went off. Lately I’ve been slogging through Allemang and Handler’s “Semantic web for the working ontologist”. It’s heavy going, not least because the further you get from the early chapters the greater the underlying uncertainty. Will we really use OWL and SPARQL ten years from now?
On the other hand, the early stuff is persuasive, especially the descriptions of the RDF Triple Store and the RDFS data definition and transformation language. The above wikipedia link, as of today, doesn’t do justice to the real value of this work. To me the fundamental appeal is that there’s a solid and increasingly accepted specification for a universal data representation store that not only encodes data, but also metadata – the equivalent of column names, data types (strings, numbers, etc), table names, relationships and so on.
This stuff was created over the past 20+ years to integrate “say anything, anywhere” data sources across the Net . The movement has been popularized by Tim (web ground zero) Berner-Lee in lots of semantic web presentations.
So we now have a standard way to represent data and data relationships, to manipulate both things across all platforms, the RDF Triple Store + RDFS combination. Heck, it can even manage graphs!
That’s how I want my data to live, whether it’s my Contacts or my Calendar or my Passwords everything else that’s not a document. Finally I’ll have the tools I need to, you know, manage my contacts and passwords and so on.
Sure, performance is an issue – but I’m not dealing with a trillion records. I’m dealing with a few thousand records. Stick the damned structure in RAM and build indices galore – I don’t care. The key thing is having a standard way to store and manipulate my data.
So all you niche vendors out there, all you guys that want to take the market away from FileMaker and especially Microsoft – give me a consumer-grade (think Bento – but less stupid) Triple Store based solution that will let me work with data across platforms and applications.
I’d appreciate it.