This is certainly not the case. In the usual contorted way of defining syntax, RFC 822 allows the "local-part" of an email address (the part before the "at" symbol) to be any number of one or more "words" separated by dots (periods or full stops); a "word", on the other hand, can be either an "atom" or a "quoted-string". An "atom" can be any number of "chars" other than "specials", "spaces", or "ctls" (i.e., any ASCII character from 0 to 127 other than ()@,;:\\".[] space, or ASCII 0-31, 127. A "quoted -string" is a quote-mark followed by one or more "qtext" or "quoted-pair" characters, followed by a quote-mark ("). A "qtext" is any "char" (ASCII 0-127) other than the quote-mark, backslash (\\), or carriage-return(CR), but including "linear-white-space" (which, in turn is a combination of optional "CRLFs" (i.e., CR followed by LF) and "LWSP-chars" (spaces or horizontal tabs). Finally, a "quoted-pair" is a backslash followd by any "char".
So the RFC would permit email addresses with local-parts such as john.smith, john-smith, $John!=+^Smith*, John_Smith, "\\\\\\"Jo(hn@Sm]]ith", or even Jo\\ hn\\ Smith and much, much weirder addresses.
Michael D. Sullivan Bethesda, MD (USA) (To reply, change example.invalid to com in the address.)