tags:

views:

106

answers:

4

I have rough ideas - like dealing with singular/plural, two or more words/phrases that mean the same thing, misspellings, etc. But I'm not sure of any patterns or rules of thumb for dealing with these, either programatically and automatically or by presenting them to administrators or even users to clean up.

Any thoughts or suggestions?

+1  A: 

As SO does, suggesting existing tags as you type is a very good thing. It will (hopefully, almost) take care of the plural / singular thing and misspellings, as people will re-use existing tags much more.

Greg
That's good. We have something similar in place now, but we should work on improving it.
Thomas Owens
That said, however, we really can't improve it until we fix the existing tags which have problems. Still a good idea, though, as we want to keep clean tags clean.
Thomas Owens
Cleaning the tags shouldn't be too hard... Assuming your site is in English then start by listing all the tags ending with "s". To catch misspellings a wordlist should help you get most of them.
Greg
Depending on your site's scale and purpose, you could moderate new tags - this would give you a very clean data set but also give you a lot of work to do on a larger site!
Greg
A: 

Use an ajax-driven suggestion form, like StackOverflow :)

friol
A: 

Assuming a setup not dissimiliar to SO: how about moderators being allowed to merge a smaller voted tag into a more common one, e.g. VS9 could be merged into VisualStudio2008 but not letting the larger used tag to be merged into a smaller tag grouping. Adding a badge incentive or similiar to this.

dove
+1  A: 

You should have a policy for the format of the tags (e.g. tags should be singular). Depending on how diverse the tags are, it might be useful not only to auto-complete while you are typing in a tag, but also to suggest similar tags, so that it is easy for people to use the tag system. Additionally, a cleanup process could correct common spelling mistakes and substitue deprecated tags according to a translation table.

sdfx