I would tend to agree with placing standard/baseline algorithms in the “not OK” (although not flat-out “rude”) category. First if it’s a good baseline, chances are there is a free implementation available somewhere. If you ask me about that, this is a good indication that you haven’t done your homework.

Second, I would be reluctant to have my implementation of somebody else’s work perceived somehow as a “reference” implementation.

I guess that the reaction you will get to your request depends in large part of the perception of how seriously you prepared that request. I would suggest the following rule of thumb : if you have put less effort in your request than you expect me to put in the reply: not OK.

I guess my take is that most young researchers are relatively shy (me included) and don’t send enough emails asking for advice/explanations/code/etc. I think emailing distinguished researchers should be encouraged not discouraged, and most researchers are indeed very happy to receive such requests. The lesson here, I suppose, is just to write a polite email if you want a polite answer.