Sune Karlsson writes
It doesn't only disable stripping html it also disables decoding of html entities. This would cause problems with Author-x-name-last and Author-x-name-first. Unofficial but fairly widely used.
ReDIF-perl has stripped and decoded all values ever since these capabilities was added. Changing things that services might rely on is not a good idea.
It's unfortunate that Jose Manuel has been caught out by this but there was a quite lengthy discussion of this on RePEc-Run when stripping was added in 2012.
It is a fudge but I think the best is to special-case x-bibl.
I think if we can combine the sharp analyics of Sune with my dull vision, we can do better than a fudge. First, maybe we can check if the x- attribute is in a cluster, and if that is the case, don't skip the strip. Second, if we authorize name-last and name-first, as well as Doi, which I think we should as well, we can then preprocess a template from former x- to now allowable structures. -- Cheers, Thomas Krichel http://openlib.org/home/krichel skype:thomaskrichel