'The Formal Definition of "Harassment"'

The first place to look in determining the scope of harassment law, of course, is the legal definition of "harassment." Speech can be punished as workplace harassment if it's

• "severe or pervasive" enough to
  
• create a "hostile or abusive work environment"
  
• based on race, religion, sex, national origin, age, disability (including obesity), military membership or veteran status, or, in some jurisdictions, sexual orientation, marital status, transsexualism [sic] or cross-dressing, political affiliation, criminal record, prior psychiatric treatment, occupation, citizenship status, personal appearance, "matriculation," tobacco use outside work, Appalachian origin, receipt of public assistance, or dishonorable discharge from the military
  
• for the plaintiff and for a reasonable person.
  

Note what the definition does not require. It does not require that the speech consist of obscenity or fighting words or threats or other constitutionally unprotected statements. It does not require that the speech be profanity or pornography, which some have considered "low value." Under the definition, it is eminently possible for political, religious, or social commentary, or "legitimate" art, to be punished [18].

[...]

[18] The definition also does not require that the speech take place in the workplace; even speech outside the workplace can be considered if it creates a hostile environment at work. See Intlekofer v. Turnage, 973 F.2d 773, 775 (9th Cir. 1992) (relying in part on a coworker "telephoning [Intlekofer] at her home" to support a hostile environment claim); Bersie v. Zycad Corp., 399 N.W.2d 141, 143, 146 (Minn. Ct. App. 1987) (relying in part on a coworker "calling [Bersie] at home" to conclude that plaintiff had made a prima facie showing of harassment, expressly applying Vinson); cf. Bartlett v. United States, 835 F. Supp. 1246, 1262 (E.D. Wash. 1993) (finding that two instances of sexually suggestive conduct, including "[p]laintiff receiv[ing] a sexually explicit card at her home from a coworker," did not rise to the level of sexual harassment, but not even hinting that the card was somehow categorically disqualified because it was received outside the workplace); Myer-Dupuis v. Thomson Newspapers, No. 2:95-CV-133 (W.D. Mich. May 9, 1996), reported in Mich. Law. Wkly., May 27, 1996, at 12A. These cases are eminently consistent with the harassment definition given by the Supreme Court: It's quite plausible that speech by coworkers outside the workplace may create a hostile environment within the workplace.

-- Eugene Volokh, 'What Speech Does "Hostile Work Environment" Harassment Law Restrict?'

TMI: Tired of Swordfighting

I haven't been writing anything in a while. Honestly, I've been experiencing burnout. It's not that I've been working too hard or anything like that; just that when a colleague tells you "we don't want you around", it's harder to convince yourself to go to work in the morning, or even in the afternoon.

But I should talk about the new and exciting thing I'm working on. The Rust compiler is slow, and while there are any number of reasons for that, one of them is that it recompiles the entire crate every time you change any item, even in a way that's unobservable by any code that depends on the item. Compilers like GHC, with its --make mode, and SML/NJ, with its compilation manager, have been doing incremental recompilation for a long time, so there's no good excuse for us not to do the same for Rust. Incremental recompilation means that the compiler automatically determines dependencies between declared functions, and avoids recompiling files if they haven't changed since the last build and if nothing they depend on has changed either. For example, if I added a #debug statement to the body of a Rust function and made no other changes, it would be safe to recompile only the code for that function.

That sounds easy enough, but ( Of course it's not! )