I was looking for a specific question related to the probability integral transform, i.e., the theorem which states that if $X$ has a continous distribution, then $F_X(X)$ has the uniform distribution. I read this question some time ago and I made the mistake not to favorite it, so I cannot retrieve it any more. This question had comments by user @whuber, which pointed out that the usual proof which assumes $F_X(x)$ to be strictly increasing is needlessly restrictive. I tried to look for questions about "probability integral transform", both on Cross Validated and on Math Exchange, but the ones I found were not the right one: I guess the question didn't have "probability integral transform" in the title or maybe not even in the question body . I'm not sure how I could retrieve the question...I don't think I can search for comments by a specific user, and anyway since the user in question is very active, I would get far too many hits. Any suggestions?
@whuber has provided some good suggestions for search strategies. You may still have difficulty finding it, however, and as he notes, it could have been deleted.
Given the apparent value of the information that was contained in the thread, and the difficulty finding it, another possibility is to ask a new question of your own. You could ask a question with a better title and more appropriate tags, for instance. That would make the thread easier to find in the future. This benefits the sites mission in that we hope to create a repository of high-quality statistical information for future reference. But if that information cannot be found by anyone, it isn't any good. Thus, your new question, while arguably a duplicate, would do the site a service.
It's possible the post or the entire thread was deleted, in which case it will be inaccessible to most (and difficult even for moderators to find). Let's assume that's not the case.
Although our site offers some useful search options (such as searching according to question or answer, number of votes, and so on) it can be useful to exploit Web search engines. For instance, in this case you remembered the name of a user and some key words. You can use them in a Google search like this:
There are some promising hits. The trick (obviously) is the
site: part of the search string. It focuses on our entire site and it can pick up comments.
As a general rule, using my name in a Google search is not going to be helpful: in my role as moderator I have commented on a sizable proportion of all threads. One site search option I have found very helpful, though, is to include the id of any user you think might have contributed a question or answer. You will have to use our own site search capability for this. For example, I recall user "glen_b" wrote at least one nice answer about histograms. A good way, then, to find this answer--and possibly others like it--is to look up his id (search for it on the Users page) and include it in the search:
Sorting the results by votes brings the post I was looking for right to the top of a set of related posts.