Re: Of tags and topics

[Matthew Paul Thomas <mpt myrealbox com>]

On Mar 6, 2006, at 1:35 AM, Peter Harvey wrote:
> ...
> The problem I have is that no structure can be inferred from tags. 
> Tags can be anything at all, important or unimportant. Topics are 
> those particular things that are important about the object and that 
> the user wants to be presented with to find an object again.

It takes effort to type a keyword. Therefore keywords *are* "those 
particular things that are important about the object and that [I want] 
to be presented with to find [it] again".

> A more detailed blurb on topics is on my page describing the 
> hierarchical bookmark menu. What interface would you propose instead 
> if we just had tags/keywords?
> ...

If by "interface" you mean "method of creating submenus", the same as 
we have now.

> Note that I don't like the term "keywords" to replace "tags" as they
> need not be "key" at all.

If they weren't, people wouldn't use them. (There will always be people 
who don't understand their purpose and enter useless keywords/tags -- 
see the spurious tags on 43things.com, amazon.com, and occasionally 
www.metafilter.com for examples -- but that problem can be lessened by 
using the more familiar and precise "keywords" term.)

> In real life we use the term "tag" to refer to all sorts of labels for 
> objects (prices, warnings, sizes, etc). Perhaps "labels" would be a 
> better term? But not keen on "keywords".
> ...

In the ontological sense, "tag" is merely a Web-2.0-ese synonym for 
"keyword". That's why I prefer the latter: it means the same, is more 
familiar, and is more obvious about expecting words rather than phrases 
(which is necessary for the interface to be simple enough for most 
people to bother with, i.e. a single text field).

--
Matthew Paul Thomas
http://mpt.net.nz/