Organising blog tags

conjoint.biz:1080
conjoint.biz:1080 tags
One thing I notice is that the cross wise linking of tags is easier to create in blogger in that I only have 26 posts so far. But there is still the problem of the tag's relevance. I notice that in the texai project he uses tags in an exemplary fashion. As would be expected from Steve Reed who worked seven years with Cycorp, Inc., the creators of the Cyc Knowledge Base.

Here are the categories Steve has settled on:-
* AGI
* Automatic Programming
* Cluster
* Construction Grammar
* Content
* Design
* Dialog
* Free Software
* Hierarchical Control System
* Java
* KB Implementation
* Lexicon Acquisition
* Linux
* Ontology
* Project
* RDF
* Sesame
* Skill Acquistion
* Spreading activation
* Technical Disclosure
* Web
* Workflow

That is 22 in all.
Notice that they are self consistent and that he has settled on a single style -
Capitalise, use spaces, no sub categories.
Actually he does have
Skill Acquistion
and
Spreading activation
rather than Spreading Activation. I cannot see the reason for that.
He also has a spelling mistake, the bane of any pre-thought tagging system. Though I would think he could correct that?
Steve does not discuss how he arrived at his categories, but it looks to me that he settled on them before he started writing the blog, perhaps possible where the blog has a very specific focus?
But I can say that it makes navigation of the site that much easier for the user.
This means that I am aware of the limitation of using tags in this site.
Further, on my other site (unreferenced here), there is chaos.
As to conjoint.biz:1080, well it is essentially a note book. I am not sure it it worth the effort for the benefit of the site - as opposed to my own benefit - to learn how to alter the url referenced on the fly. I have the problem that I can only create linkable urls that are within my firewall when on the home network, and I cannot do NAT to make them appear external. That means that if they are to be followed externally, I cannot follow them, which in turn means that they are not listed on those pages as referencing links, unless someone else follows them.

This suggests an alternative, which is to have a script crawl my site from my remote account, this should generate the links as it goes along.
This way I write all links as external. Of course I still cannot follow them myself.
I think that changing Nanoki is a bit of a mess as that would mean the code is bound to this web site and this port. It is bad enough that the site links are.
It does seem that at some point I need to dive into lua.
Not now though!

0 comments:

top