News for hackers. Stuff that matters.
Re:Efficient Parallelization of an Ubiquitous Sequential Com... by JoshTriplett on Thursday December 07, 2023 @01:45PM (#38560102)
> I bet the real rule is some reality about feasibility of pronunciation, even though native english speakers see the rules explained in terms of spellings.

The rule as often given in English classes is to use "an" if a word starts with a "vowel sound", rather than starting with a vowel. So, it's "an herb" (since the 'h' is silent in (American) English), but "a ubiquitous".

Relatedly, you can infer whether someone likely pronounces "URL" by spelling it out (like "you are ell") or as a word (like "earl") based on whether they write "a URL" or "an URL".

Re:Efficient Parallelization of an Ubiquitous Sequential Com... by mvc on Friday December 08, 2023 @04:48AM (#38567134)
As someone who uses a dialect/accent of English that means I often pronounce these "silent" letters (e.g. I would pronounce the "h" in herb), I've often wondered if selection of "a" vs "an" is supposed to be accent specific.
Re:Efficient Parallelization of an Ubiquitous Sequential Com... by JoshTriplett on Sunday December 10, 2023 @04:31PM (#38595254)
I think so, yes; if you say "herb" with a non-silent 'h', you'd write "a herb".

By way of example, the name Herb is commonly pronounced with a non-silent H, so even in a dialect where "herb" is pronounced "erb", you'd write "A Herb picked up an herb.".