[uf-discuss] hCa* CATEGORIES as TAGS

Mark Rickerby maetl at mcs.vuw.ac.nz
Sat Nov 26 18:18:21 PST 2005

>> On Thu, 2005-11-24 at 21:27 -0800, Chris Messina wrote:
>> My first reaction is that categories !== tags

> On 11/27/05, Luke Arno <luke.arno at gmail.com> wrote:
> In fact, tags == categories == labels == etc...

Another way of looking at it is that tags are a superset of
categories, labels, etc...

If something is a category, then it can definitely be represented as a
tag. But if something is a tag, it may not necessarily have a category
representation. This of course, is totally dependent on the IA of the
site/tools in question.

One potential benefit of relTags is that they can be decoupled from an
actual metadata representation, and just exist as a
hyperlink->uri->relationship binding.

For example, the following makes sense to me - I don't know if it will
to anyone else:

<span class="tags">
 <em>Filed under:</em>
 <a href="subject/physics/" rel="tag" class="subject">physics</a>
 <a href="subject/philosophy/" rel="tag" class="subject">philosophy</a>
 <a href="topic/information/" rel="tag" class="topic">information</a>

The class attributes could well be superfluous, but serve to
illustrate the point here, that tags can represent well defined
"types" of metadata for a single item.

To me, the issue seems to hinge less on the categorical or semantic
"differences" between categories and tags - more on the actual tag
syntax itself, in terms of whitespace rearing it's ugly head. To
(slightly) contradict what I said earlier, it seems that only tag
schemes that support space separated, comma delimited phrases, would
be directly compatible with the conventional notion of categories. Eg:

Categories: Philosophy of Science, Information Theory

Not all tagging apps do support this phrase structure, but obviously,
any that did would be able to produce metadata where categories<->tags
are interchangeable.

So categories would be a subset of a subset of the set of all possible
tag schemes. (feel free to laugh me off the list for such
meta-stupidity - I'm just throwing the idea out there :)

>>> On 11/23/05, Brian Suda <brian.suda at gmail.com> wrote:
>>> Is this an abuse? because you could always just mark-up the hCard with:
>>> <span class="category"><a href="../tag/Conference"
>>> rel="tag">Conference</a></span>
>>> But this is asserting a relation between this HTML page an the TAG,
>>> not the iCalendar file and the tag?

In the vCa* example, the use of class="category" certainly seems less
ambiguous, but I would think it's ok in some situations to just assume
that tags are categories when the structure of the content is well
defined but the metadata labelling isn't well defined.

I wouldn't think of this as an abuse unless the IA of a site
specifically defines a difference between "tags" and "categories", or
where the relationship between different items of content to the same
set of tags is much more ambiguous.

Very few writers or publishers will care about whether a keyword is
defined as a category, tag, topic, etc... The actual keyword itself is
much more important.


More information about the microformats-discuss mailing list