A blog looking at business communication, knowledge management, scripting tools, OS technology news and other things of interest to mobile tech workers. As I find interesting news this will also contain pointers to thoughts related to configuration managment, change management and general software development.
Here is more on the "tagging good ... taxonomy bad" meme that is making its way around the internet. Here is what got my attention:
There are strong similarities to how wikis and tagging works. Tagging lowers transaction costs for contributions and fixing mistakes. This increases participation and the probability of the right data actually existing in the first place. It also enables a dedicated community to self-govern (and note that as in the case of Wikipedia, the enthusiasm hasn't worn off)
A single tag can be applied in error, and be fixed locally, but that matters less when viewed in the aggregate. Larger patterns arise that are statistically significant.
This is what I am trying to do with TWiki, namely fostering a tagging practice amoung the users instead of a rigid taxonomy. I have implemented a simple tagging system in the wiki with the goal of transfering the organization / categorization work to the whole group of users. The goal is to end up with a better and more useful organization system for the content of a single wiki ... and to have this emerge from The Wisdom of Crowds er ... wiki users.
Just spent 30 minutes cleaning up a huge litter of spam that hit a ton of old posts ... weird thing is that it seemed to cluster around mentions of the Wall Street Journal ... strange. Anyway, is there anything we can do so that PyCS isn't so vulnerable to this?
Social Tagging and its close cousins are a meta data killer app according to Clay Shirky [via Cory Doctorow]:
... users pollute controlled vocabularies, either because they misapply the words, or stretch them to uses the designers never imagined, or because the designers say "Oh, let's throw in an 'Other' category, as a fail-safe" which then balloons so far out of control that most of what gets filed gets filed in the junk drawer. Usenet blew up in exactly this fashion, where the 7 top-level controlled categories were extended to include an 8th, the 'alt.' hierarchy, which exploded and came to dwarf the entire, sanctioned corpus of groups.
This is something the 'well-designed metadata' crowd has never understood -- just because it's better to have well-designed metadata along one axis does not mean that it is better along all axes, and the axis of cost, in particular, will trump any other advantage as it grows larger. And the cost of tagging large systems rigorously is crippling, so fantasies of using controlled metadata in environments like Flickr are really fantasies of users suddenly deciding to become disciples of information architecture.
So group developed meta data trumps the professional developed top down taxonomies. Cool.
Now, a wiki is like a little version of the internet ... lots of topics, links, resources etc. In response to structure issues in Wikis, I have always wanted to let the community develop its own method, as I mentioned in a prior post Wikis are Not Unstructured. But, sometimes a wiki can be setup in ways that does not help the community develop its own classification system. So, what can I do to help a folksonomy emerge in a wiki?
One approach that I think helps is what we see in Wikis like MoinMoin. It lets users classify content by adding tags directly to the content. This allows content to be classified in many different ways... and this is helpful since it is easy to then build queries to find and present the content that has been tagged .. and it is bottom up and so could, if actively used, turn into something like a folksonomy. The tags that are added to the topics themselves link to a master topic for that tag which displays all the tagged topics. So it is kind of like a social linking tool.
Now, my wiki work lately has been focused on TWiki. In this particular case it has an inflexible categorization system based on forms and fields. Someone needs to decide in advance what the tags will be ... so a folksonomy never really emerges. The choices are too structured and all decided in advance and you end up with a lot of questions like how do I fit x topic into your category scheme? ...
What I think TWiki needs is a convention/technique that allows users to add 'tags' to wiki pages as simple text and then have ways of presenting the folksonomy that emerges. What I think I want to do is something like what MoinMoin does which is to just embedd the tags as a piece of text in the topic. It is visible to everyone ... and so would encourage a less structured approach to classification .. especially since we could apply as many tags as necessary to any topic ... something that the form approach can't do.
I thought through the idea of using a tag as a topic and having that topic link to the topics that are 'tagged' ... but that seems to be too much effort for a wiki. I think a better solution is to have tag pages that query for all pages that have that tag on them ... This approach is also more group friendly (having your single category choice changed by a Wiki Gardener isn't nice. Having them add a few more tags to your page on the other hand is a great compliment. Refactoring tag names is also easier since all you would need to do to rename a tag is rename the tag topic that have TWiki fix all the links ... and voila you're done.
Am I implementing MoinMoin's category stuff in TWiki .. yes I think so ... I wonder if anyone else has done anything similar in TWiki.
I have been looking for a great summary of brainstorming meeting techniques and facilitation tricks for a while. Here is a good one. I really like the tricks outlined to get the most creativity out of the session:
What is the opposite of what we want? Get the group to describe, in detail, the opposite of what you want to have happen. What is the worst web navigation design possible? This never fails to get people to smile, and dig in. It is just so much fun to work the other way for once.
Random theme generators: (This works only for design related brainstorming) Before the meeting, make a big list of adjectives, colors, verbs, and attributes that might or might not apply to the project (20 or 30 of each). Write them on index cards, but keep them in stacks. Shuffle each stack. Then pick one from each, and try to design something for it.
Rotate: Anyone, at any time, can call out rotate and everyone in the room has to get up from where they are sitting, and move to the chair to their left. I have no evidence to prove it, but I am convinced moving around physically helps people to move around mentally.
Roadblock removal: Eliminate assumed constraints. Tell the group there is no limit on costs, or time, or other resources. They can think as big or expensive as their minds allow. Think of other constraints that the group is assuming.
Roadblock introduction: Come up with some ridiculous new constraint to the problem. The goal is not to make it harder, just to force peoples creative muscles to work differently.
I see that PyDS now has templates for Rendering Archives by Year, Quarter, Month ... but I have no idea how to turn this on. The underlying python code seems to imply that it should be a target available for the navigation tool so I can a link to the archive page... but no joy in my case. Maybe my templates are too customized for this to work ...