[uf-discuss] rel="nsfw"

[John Allsopp]

Hi all,

Coming late to the discussion of rel-nsfw[1], a couple of points I  
don't think I've seen raised, one that pertains to HTML, and one to  
ufs specifically.

1. despite rel-nofollow's "success", rel is not the appropriate  
attribute.

As I am sure most people here have read numerous times, rel

"describes the relationship from the current document to the anchor  
specified by the href attribute"[2]

"nsfw" describes the authors opinion of the nature of the content to  
be found at the end of the link, and by no means the nature of the  
relationships between the destination and source documents.

So, it's far from ideal on that count.

2. this is not visible metadata (nor is nofollow, for that matter)

In this case, there is no way, without the use of either explicit  
content (or CSS not supported in IE6 and older, and I really don't  
know about screen readers) of signifying through the use of the rel  
attribute that the content is in the opinion of the linker nsfw. Turn  
off CSS and any indication to human readers will vanish in this case  
at any rate.

It certainly, as has been more than once mentioned, doesn't pave the  
cowpaths (where explicit visible content in the page (though not  
always in the link content) is how nsfw is almost invariably indicated.)

The second concern applies to the extended idea of using a class  
value of nsfw on arbitrary HTML elements, but that at least gets  
around the problem of the first. But, it's really shaky with the  
cowpaths test, because I have only ever seen links advertise their  
destinations are nsfw, not page subsections themselves advertising  
this (which doesn't mean that it never happens, but that anecdotally,  
it is rare. It also happens to be a different problem - marking up  
specifically your own content (rel=tag like) as being nsfw, as  
opposed to marking up other content as being nsfw (more analogous to  
xfolk))

[1] http://pj.doland.org/archives/041571.php
[2] http://www.w3.org/TR/html4/struct/links.html#adef-rel