… a rule that is broken too often eventually ceases to be a rule. Insisting with stating it officially while routinely allowing it to be ignored, or trying to enforce it only half-heartedly, especially in a selective manner, does not do much for the credibility of a person or an institution. Sometimes exceptions are in order, but they should be made seldom, clear and precise reasons should be provided, and a reasonably broad consensus sought.
Otherwise, how is one to grant an exception to some, but not to others ? And if the exception becomes the rule, is it not more honourable to do away with the rule altogether ?
Anyone who has ever taught a large class at the college level fully understands that. If a student misses a test, for example, great care should be exercised in assessing whether a make-up test is in order — and usually that is only justified if extenuating circumstances apply (e.g., sickness). For, if a student is de facto allowed to take a test at a different time or place, more convenient for him/her, if oversleeping, traffic, problems with the dog etc., are all regarded as valid reasons, then what should prevent the other 100+ students from legitimately demanding to be granted the same privilege ?
Next thing you know, an instructor would be spending the entire term administering tests at random times — essentially chaos. This is why universities draft very careful policies to which instructors are required to abide, when dealing with situations of this type.
In general, leaving in place for purely formal reasons, policies that are in practice not enforced, or enforced unevenly, whimsically or downright unfairly, is a recipe for malcontent, annoyance, resentment and possible trouble. Particularly distasteful is the case of newcomers being told that a particular restriction or requirement applies, and for them to abide by it, only to find out later on that old-timers got around the same restriction.
A case in point, speaking of rules that exist on paper only, is illustrated in this post by Doug Natelson, who expresses his perplexity over a blatant claim of priority in the title of an article recently published in Physical Review Letters. OK, what is the big deal with this ?
It is still the official policy of the leading American physics journal, namely Physical Review, “not to print claims of novelty or priority”.
What does that mean ? Well, that expressions such as “we report in this paper results of the first measurement ever made of xyz…”, “we describe here a new computational method…”, “this is a novel finding…”, ought not appear on a manuscript published on that journal. Why is that ? To be honest, I am not sure. The first time I saw a paper of mine edited, with the word “new” removed from the abstract, was over twenty years ago. The explanation furnished by the editorial assistant on the proofs, was that “the word new is to be avoided, as by definition any article published in Physical Review should describe something new.”
Despite its obvious silliness, this rule (censorship ?) was enforced almost obsessively, and often times without much common sense, to the point of acquiring grotesque, Orwellian connotations, with awkward end results .
Recently, I have been offered a different rationale for it, namely that “claims of priority and/or novelty should be avoided because they are difficult to assess independently”. This sounds more reasonable.
However, irrespective of its motivation, it is a simple fact that such a policy is not consistently enforced. A straightforward search (using WebOfScience, for instance) yields literally hundreds of articles published in Physical Review over the past twenty years, featuring words like “first”, “new”, “unprecedented”, “novel”, often prominently displayed, in the abstract or even in the title, as in the above example.
I remember being surprised the first time I read one of these words in an article published in Physical Review, and indeed in 1995 asked this very question, out of curiosity, of the at-the-time Editor of Physical Review Letters . The answer I was given was along the line of “Well, that is just a statement of principle, if the author can present a reasonable case to the effect that a word such as new is warranted, we are not really in the business of censorship”.
After hearing that response, I went back to using the “forbidden words” in my articles, when I felt they were appropriate. Invariably, I would find them deleted in the proofs of the accepted manuscript, once they would be sent to me. At that point, I would simply request that they be re-inserted, and I do not recall a single instance in which I had to do more than simply ask, in order to have my request granted. Typically I make the very case that in that particular context those words seem appropriate, that Physical Review does not practice censorship, that exceptions have been made many times (I normally include a list of papers published in Physical Review containing those words), and there seems to be no reason not to make one for me as well.
So, based on my experience I would say that the enforcement, on the part of Physical Review, of its own stated “policy” of not printing claims of novelty or priority, largely depends on the stubbornness of the authors. That is of course like saying that the policy is de facto moot.
If that is the case, would it not be better not to have such a policy in print, and instead have Editors recommend on a case-by-case basis, that authors reconsider certain expressions or words that may seem over the top ?
 The funniest that I have ever seen consists of removing the word “novel” from the phrase “doping such a compound has been predicted to result in some possible novel behaviour”.
 Yes, it was a job interview. Back in those days it was far from clear that I would be able to pursue the academic career, and out of all “alternative paths” that appeared before me, working in the editorial office of a prestigious physics journal such as PRL had some appeal for me.