This post emerges from a conversation on twitter arising from a comment I made about categories after uploading a document to edocr.com

To get straight to the point, I usually hate being forced to choose a category from a drop down menu. It smacks of technocracy where the system designer is unnecessarily imposing, subtly or not, a narrow view of how data should be described. This clumsily reasserts the unequal relationship between the developer and user. Instead of being empowered to input your own stuff and personalise your profile, a poor uploading experience can leave you feeling like a patronised data entry clerk. I’m also convinced that it’s not the cleverest way to organise content

I happened to mention that I was writing about railways and ManojRanaweera tweeted me “we now have a whole category for railways”

I took this as encouragement to submit an upload in document format ( .pdf) to the excellent edocr.com service, of which I am an occasional but enthusiastic user. But why would the existence of a new category code make any difference? As if there were a whole load of content waiting around which cannot be submitted for want of more categories? Categories are only descriptions, they are not empty containers compelling somebody to fill them. If there isn’t a perfect category available, which is nearly all of the time, you just have to choose the nearest. It’s annoying, but common experience. Yes, but since we now live in the new twitter enabled webosphere, I don’t have to keep these thoughts to myself. The CEO, owner, promoter and whatnot of edocr.com is an active participant in the conversation and a long term twitter contact of mine so I can ask directly to Manoj:

@ManojRanaweera I’ll upload my post with far too much detail about Yorkshire railways, but do categories mean more than tags?

M replies:

@aroberts #edocr tags totally rely on publisher – that means no discipline. Categories force you 2 choose 1 category that is most relevant. http://lin.cr/fac helps structure the library of docs and make it easy to find within edocr.com

me:

@ManojRanaweera successful social object sites have used used folksonomy tagging only, eg flickr whereas fixed categories always problematic. On the other hand youTube insists on a broad category, but at least it’s not multi-level #edocr

Actually, that’s not even strictly true about flickr. They do have a top level categorisation but it is subtle and defaults sensibly so you might not even notice that all flickr pictures have to be designated either as

Safe, Moderate or Restricted AND as either a Photo, Screenshot or Art/Illus

so immediately I can think of an exceptional case. How should scanned documents fit within these three? I’d probably choose photo but only after hesitating.

YouTube categories are silly, mixing subject matter and genre indiscriminately:

But back to document sharing:

@aroberts – depends on who your audience is. I do not think we are too far off other players providing similar product

The only similar product I know of is scribd so I went through the uploading process there as well and soon discovered:

“Scribd forces you to choose from some odd categorisations too”

Twitter must be really catching on now because Jason at Scribd clearly had an alert set and replied

jasonatscribd @aroberts Categories are a work in progress, and welcome suggestions, but they’re largely to help people browsing Scribd. The real power is in tags. The more rich and descriptive your tags, the better your chances of being found on Google, etc.

which is probably closer to my own view than Manoj on this topic, but there’s more to come on this from him:

@aroberts #edocr – The importance of this level of categorisation will become clearer as we continue to build edocr functionality. Will blog my thinking.

Manoj correctly suggests taking the discussion further on the blogs so this is my contribution.

Web2.0 may be old hat by now, but that doesn’t mean we have to bring back the old printed matter, library and directory based metaphors of web 1.

Search and tags are the contemporary means by which content is discovered and browsed. Explore has some value, but is secondary and library classification, indexing and directories are hangovers from last century because they always come up against the top down design problem, ignoring the more powerful emergent patterns that come from bottom up self organisation and collaborative meta data.