kmarks at technorati.com
Wed Jan 3 16:16:05 PST 2007
On Jan 3, 2007, at 3:10 PM, Nick Peters wrote:
> Seeing the tag implementation on Operator has made me question the
> existing tagging standard. With wordpress you may get something like
> "?cat=13" for a tag or something that may not even be the intended tag
> at all.
Yes, Wordpress abuses the rel="tag" spec by doing that, so I have had
to code round it at Technorati. They can't do proper url path tags on
all installs, but the code doesn't omit rel="tag" on the non-tag links.
> Tags may only be placed in the URL path, and only in the last segment
> of the path. Tags may not be placed in query parameters or fragment
> identifiers. e.g.
> is still a URL for the tag "tech", not "fish" or "emu".
Actually, stripping parameters is recommended (that's why it says 'last
path segment'). a previous implementation bug of mine at Technorati
wasn't doing that properly.
> After doing some research on the wiki I see that the
> rel="tag" microformat is based off of existing defacto standards
> (implemented by sites such as del.icio.us and flickr). I still don't
> see why the standard extracts the tag from the last part of the URL
> instead of the information inside the anchor tag. When I see a tag
> and click on it, I expect the visible content, not what's appended to
> the end of a URL. Anyone care to shed some light on this for me?
The 'last path component of URL' part was based on existing practice by
both del.icio.us and Flickr when we did the analysis. This was done to
encourage use of these kinds of tagspaces, rather than just allowing
linking to an arbitrary URL and using the link text. Setting up a
tagspace does take a modest amount of webserver configuration, granted.
More information about the microformats-discuss