h a l f b a k e r yThere's no money in it.
add, search, annotate, link, view, overview, recent, by name, best, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Tuple Tags
Revise the notion of 'tags' and 'tag clouds' by turning them into tuples. | |
Tagging is common in the folksonomy
structure we are all familiar with these
days. However, they fall short of semantic
usefulness.
I propose that tags become tuples. A tag
is no longer a single word. A tag is now a
name:value pair. Simple, but so much
more useful.
So, instead of simply
tagging a
photograph (for example) with the usual
assortment of words, I tag them with for
example place:London place:Beckton
weather:sunny weather:'bloody cold' and
so on.
This could even extend to other media,
such as music, where you could have
organised tag tuple clouds instead of the
rigid structure of ID3 tags. For example,
many songs appear in many albums.
Instead of one 'album' field, you could just
keep adding them:
album:'Ziggy Stardust',
album:'Changesone' and so on.
Faviki
http://www.faviki.com/
Use words from Wikipedia (including sometimes the disambiguating type in parentheses) [jutta, Oct 04 2008]
Zigtag
http://www.zigtag.com/
Still in beta. Tags are not prefixed, but internally typed and disambiguated by giving descriptions at tag-time. [jutta, Oct 04 2008]
Fuzzzy.com
http://fuzzzy.com/
Untyped tags with descriptions and relations between them [jutta, Oct 04 2008]
Haiko Hebig goes at it alone (2005)
http://www.hebig.or...ves/main/001793.php
Facet prefix (as here). Never mind that critical mass problem behind the curtain. [jutta, Oct 04 2008]
Semantic tagging systems, May 2008
http://i9606.blogsp...gging-projects.html
Quick overview of commercial and academic systems from Benjamin Good [jutta, Oct 04 2008]
DBpedia - Rethinking Wikipedia infobox extraction
http://blog.georgik...infobox-extraction/
Demonstrating some of the challenges extracting structured information from Wikipedia's (attribute, value) system where scientists have name, birth_date and birth_place, and tennis players have playername, datebirth and placebirth. [hippo, Oct 20 2008]
[link]
|
| |
...one of the basics of database design which has dropped off the cart a few times over the years. |
|
| |
(-) It's an idea that's bound to exist, but I don't think it'll be successful. There's something about the language-based simplicity of tags that you're taking away here without actually adding reliability and specificity - that'll take a lot more. |
|
| |
+, because with all due respect to [jutta], in my honest, expert opinion, tag clouds are the worst kludge on the web. (besides perhaps IE or NS 4.0) |
|
| |
This just shifts the problem from everyone having to have consistent tagging systems to having to have consistent 'attribute' systems in some sort of attribute/value markup system. See link for the difficulties Wikipedia seems to have being internally consistent with their attribute names. |
|
| |
yeah, next thing you know you'll have to put the information starting in a certain column position.... oh the horrors... |
|
| |
No, you don't have to do it any particular way. If you don't do it this way, then you're not doing tuple tags, you're doing some other sort of tags, probably just ordinary tags. They'll still exist, just like apes and fish still exist despite the evolution of humans. |
|
| |
//apes and fish still exist despite the evolution of humans// |
|
| |
We're working on it... Anyone for a McKong? |
|
| |
I'm not sure the term "tuple" is semantically correct: the term refers to a set of *data* elements; what you have is a descriptor and a datum. |
|
| |
Dictionary is the right word, I think. |
|
| |