We did some studies and found that the attribute was almost never used, and most of the time, when it was used, it was a typo where someone meant to write rel=“” but wrote rev=“”. To be precise, the most commonly used value was rev=“made”, which is equivalent to rel=“author” and thus was not a convincing use case. The second most common value was rev=“stylesheet”, which is meaningless and obviously meant to be rel=“stylesheet”.
Sign in with 
OpenID
Well, in general reverse links are useful when expressing relations. It just switches the object and subject around, which is useful when there is no corresponding reverse property for what you mean to express.
An alternative to rev would be to give the ability to explicitly specify the object, which right now is always fixed. Or, to always define reverse properties for every property that you define, but that seems bothersome.
But I guess it is true that in HTML rev is little used. I suppose that probably stems from the fact that rel is little used either, and when it is, only with a limited set of values (e.g. ‘stylesheet’) that have a defined function in representing the document.
In RDFa annotations however, which are much more semantic in nature, and also have a significantly larger amount of defined and used properties, I would argue it would probably see more use. At least I see no reason to prevent this kind of relationship from being described.
rel is used quite a lot in HTML, with a fairly wide range of values -- http://philip.html5.org/data/link-rel-rev.txt lists a load of examples. rev is relatively much rarer.
People do seem to get quite confused about the difference, judging by the numbers of rev="stylesheet" and rel="made".
Philip Taylor - 14th April 2009 20:57 - #
btw, I asked for Ian's opinion on rev="canonical" today via Twitter and got this reply: http://twitter.com/Hixie/status/1517728112
theharmonyguy - 14th April 2009 22:11 - #
In principal you can replace rev="canonical" with rel="x" where x is the opposite of canonical. The very fact that this seems kind of nonsensical is perhaps a hint that canonical is not the relationship we were looking for.
Perhaps rel="tinyurl" would be more apt.