Ok, i admit it. i don't "get" labels.
Allow me to clarify that stand a bit. i don't get the overwhelming niftyness behind the latest movement to assign copious meta data to packets of information using interfaces that are simply not designed to provide that sort of thing.
What got me thinking about this was a discussion i had regarding GMail and Longhorn's general push to increase the amount of meta-data assigned to an object. Lots of people were generally shouting that labels are intrinsically right because they provide methods of parallel organization. They allow you to associate groups in predetermined pools for fast referencing. Thing is, i don't understand exactly how this solves the problem.
Let me start off with an example. Although Anne Marie would strongly disagree, as would anyone who's ever looked inside of my cube at work, i actually do have a pretty good method of data organization that i've grown out over the years. Projects are assigned to areas, with active projects closer to where i happen to be sitting. Once a project is done, it's collected into folders, and placed in my archive (pronounced "fī'-ling kăb'-∂-nĭt"). Every year i go through that archive and move less critical items to longer term storage (pronounced "lănd-fĭll"). Although it looks like chaos, i prefer to think of it as random access storage.
i'd also note that this same mechanism works pretty well with most of the other elements of personal data that drift about in my life. In fact, my mail folder system mimics this aspect, with mail automatically routed into various sub folders based off the key items they relate to. Occasionally, i do have to move messages because of a rule misfire, but that's actually pretty rare.
It takes me seconds to find email i've gotten from five years ago. (Well, those that haven't been placed in the electronic long term storage or were eaten by Outlook, but i can even manage to track down those within a few minutes.)
Oh, did i mention that this all uses a single label system? It's called putting mail into folders.
The problem that labelling (and actually meta-data storage) attempts to resolve is that people have zero organizational skills. The reason that i don't think labels are going to solve this is because these organizational tools are adding clutter themselves. In order for labels to be effective, you have to have organizational skill to manage your data and effectively use the provided tools, otherwise all you're doing it dumping post-its, paper-clips, and bits of string all over, and then start to wonder which of the 147 lists actually contains the email you want.
Plus, there's the sensibility issue. Don't get me wrong, i love Thunderbird as my mail client, but frankly, it's a lousy database. i can't even do proper unions with it. For that, i'd rather use a real life, honest to God database like MySQL. Heck, pump some of those mid-term archive files (which are nice, parsable SMTP logs) through a quick and dirty Perl script and bam, i've got my mail loaded into a full fledged database that is way more flexible than Outlook/Exchange could ever be. Plus, i can search on weird ass stuff that fits my needs.
Yeah, it's work. So? Look, it's all front end labor rather than back end cursing. Labels and metadata try to turn things that are loosely defined as databases into something more. Why not simply use a real database and be done with it?
Like i said, the root problem is that people don't have organizational skills. What we need is less organizational "helpfulness" and more real addressing the issue. Create wizards that help people stay organized (e.g. "You've not read anything in this folder in 90 days. Do you wish to archive it?" or more importantly "i see that the following X messages have been reread the most in the previous Y days. Obviously they're pretty important to you. Do you want to move them back into your Inbox so you can get to them faster?"
i know full well that i'm going to lose this whole "Metadata is the Saviour of Mankind" argument. Heck, with any hope, i know that in a few years someone will create PostLang which will allow you to create macros to categorize your mail off of frequently occurring headers or some crap and everyone will jump on that bandwagon.
Me? i'll stick with my current single label system. It's worked great for ten years and i don't see it getting any better soon.
