Neville Hobson (who appears to have had a change of address) pointed me towards a company that seemed to believe that enterprise tagging was somehow different from tagging. You can find the full text of his post here.
That started me thinking. Which is probably a bad thing, as most people who know me will aver. But so be it.
I like tags because they’re simple. Because they don’t have to be predefined pieces of some gigantic reference data model for the universe. Or even the enterprise. Reference models are to information what Gantt charts are to projects; ways to make the intermediate product (the database or the plan) more important than the goal. Okay, rant over.
I like tags because they allow one person to say tomahto and another to say tomayto and still figure out they mean the same thing. There is value in letting people describe things exactly as they see them, because that’s probably how they would intuitively look for them.
And if we land up with lots of synonyms, even misspellings, so be it. Use heuristics and collaborative filtering techniques to weed out or let atrophy those things that need to.
I can see “free, unfettered” tags helping with:
- crossing language and geography barriers, translating between cultures
- cross-referencing between systems, translating between data formats
- bridging individual perceptions, translating between perspectives
My gut feel is that we should avoid being prescriptive about these things. Otherwise we will land up with modern versions of the e-mail folder systems I love to hate.
Which brings me to what started me thinking about this in the first place. What is the difference between tagging and enterprise tagging, and why would I need specialist software to help me do it? I can understand privacy and data protection and secrecy and all of that ilk, and if that’s the reason and all we are doing is a behind-the-firewall implementation of the same thing, then I need to understand why I need something separate for it. I’d be interested in other opinions.