Also, I recall reading about some header file that made, maybe it was Forth or something, into valid (compile no warnings) C code. Woe be unto you, of course, if the syntax is ever so slightly off, and the resulting ill-formed macros explode everything to bits.
Thus, the program becomes a polyglot, when such header is provided.
(It's probably ioccc material, and I've probably remembered the target language wrong?)
More seriously, and back to the original topic(?), I've been more mindful of late, how many things I repeat, or "know that I know" -- but don't actually have references handy to support those positions. Which might range from, say, experiments that I've done, but haven't documented for various reasons (long time ago, poorly recorded -- I wouldn't write an article about some effect illustrated with some crappy decade-old photos; just not very interesting to me, etc.), to articles I've ran across but not archived, to, just... a whole variety of things that I can't properly source (for the above and other reasons) that I have nonetheless synthesized into a more complete state of knowledge.
The latter being the most frustrating of all, because, surely if I could just tell someone all of the truths that have come together to support some particular position I hold, surely they would hold it as well, but what are the chances they'll actually sit and read all of those things -- nevermind if I had those sources to hand -- let alone come to the same inferences?
But, I rarely see others providing references either, so I'm not exactly feeling guilty about it, by way of peer pressure anyway.
So that's, in part, my confessional non-confession, if you like.
Tim