Joho the Blog
An Entry from the Archives

You weren't too clear about exactly what messiness means. If you are talking about structure alone, I'm not sure that has any relevance.

If you're talking about processes and various stages of development etc., that makes more sense to me from the 'messiness' aspect of things.

Have you ever considered the possibility of attaching tags to people, in an organization ... the documentation to which the tags would pertain typically exists, in terms of competency profiles, skills, performance and learning *contracts*.

Marry this with listings of current projects (or roles and accountabilities) and some algorithm-based availability ( or scheduling optimization) software, and you'd have the possibility of pulling together available skills, competencies and roles for addressing the next initiatives, or pulling together potential teams. I also think this might be useful in making *dynamic* learning or skill acquisition more realistic, or more practical in terms of the onging changes that many people experience.

Each person's role could (vaguely at least, in my mind) become a series of overlapping link-based *venn diagrams*, mapped onto organizational objectives.

This may mirror to some extent what actually happens more and more frequently .. and no doubt the resulting configurations would look fairly different to classical org charts. I know that this kind of thing is, to some degree, close to the work of Valdis krebs or Orgnet and Karen Stephenson of NetForm ... bu tthe tags would be useful, no ? For whenever more *formal* depictions of authority or politically-defined decision-making may be necessary, the org chart could still be available for reference. It's probably a both / and thing.

And yeah, this would probably mess up current compensation schemes, or at least cause some unrest.

to Jon Husband-- we have a project like that in my own company, but we're not going out of our way to call them "tags." They are merely attributes and they have been around since the creation of databases.

I finally went to software conference after following the blog crowd for too long, and was curious to learn what practitioners had to say about taxonomies. What I heard was a vary balanced presentation by Theresa Regli. Afterwards, I thanked her, and promised that I'd visit the folksonomist blogs and put in some good words for her. It's tough to get heard.

The "folksonomy" ideology is still an interesting little diversion, but from here on out the major tagging tools (Flickr, Del.icio.us) are just continuing the work towards implementing taxonomic strategies (hierarchical categories, clustering, etc.)

They are merely attributes and they have been around since the creation of databases.

Agreed .. I was only trying to make the point that there are now available other ways to look at the structure and work design of organizations .. that might better reflect the work being done as opposed to the organizational politics, GADM (generally-accepted decision-making) and power dynamics ... alll of which will continue on anyway.

So, as a generalization, why not use both "methods" ? As you point out by example, no doubt it's within reach for most organizations that have databases with employee and role attributes.

I disagree. I do not think tags are "merely" attributes. That reduction misses what is distinctive about tagging. Tags are generated by users and can be generated outside of databases (e.g., the tags at the end of posts that refer to a namespace).

That's not to say that tags are unique and unprecedented. But if all we knew were that tags are attributes, we wouldn't understand what tags are. Hence, I think we should resist the reduction.

As for using tags AND taxonomies: We should use what works, and in many (and probably most) cases, both are useful.

And thanks for the pointer to Theresa Regli.

I found your clarification useful .. thanks, David. How could I be so sloppy as to not recognize (in this conversation) that essential difference ? Attributes are pretty likely to reflect the assignment by external sources of meaning-making, especially in an organizational setting. There is (or can be) a world of difference.

Splitting hairs. To be able to quickly search by tags, one would put them in normalized form.

Depends what you mean by "normalized." Does searching via synonyms and clusters count as normalizing?

The following seem to me not to be split hairs at all when it comes to attributes vs. tags: Who creates them? Where do they reside? What sense do we make of them (e.g., normalizing)? Who gets to make sense of them? These seem to me to be substantial differences between database attributes and tags.