Since other "named" tags got mentioned, I would like to make some general remarks first.
I see (at least) three different types of tags or tagging involving the name of a mathematician, and I think it is important to keep them separate.
The purpose of ask-johnson and ask-noam are to signal that a question is such that Bill Johnson and Noam D. Elkies, respectively, seem like "canonical" respondents. These tags are a reference to two key-users, something like an in-joke, part of MO-folklore, or something along these lines. Anyway, they cannot be measured according to the usual standards of a tag; 20-questions would be another tag in that category.
Tags for mathematics named after a certain mathematician. We have several of them. Over time we tried to make them "specific" by which I mean that different notions named after the same mathematician have different tags and/or the name of the tag is the notion not the person. There is still some work to be done in this direction, for example gaussian is still a bit unclear. Though it is at least not "gauss."
Tags about (the work of) a mathematician. Some question about specific mathematicians are on-topic. In my mind there is no reason a priori why there shouldn't be a tag "erdos" to tag questions about Erdős (or his contributions in a specific way). And the same goes for any mathematician about who (or whose work) there is a meaningful number of questions.
To sum this up, I think we can have such name-tags if there is a need for a specific tag. Can you find, say, five questions about some given mathematician? If yes, I will not object if you tag them like this. (The "you" is an abstract you.) If no, the tag seems pointless. (Obviously there is nothing "holy" about five, but please let us say at least two, single used tags get eventually auto-deleted.)
However, we should be careful as to still keep type 3 and type 2 above separate. This is an indirect reason I can see somewhat against these tags (and this is further motivation to try to keep their number somewhat under control via imposing that threshold above). It would not be a good idea to tag some general question about euler-characteristics as euler and possibly in addition characteristic. And, in my mind, there is no questions whether this will happen, the only question is how often it will happen.