computational linguist more like bomputational bimgis

  • 0 Posts
  • 124 Comments
Joined 6 months ago
cake
Cake day: April 2nd, 2024

help-circle











  • For a lot of English speakers, the “had” and “have” in contractions is completely omitted in certain contexts. It’s more prevalent in some dialects (I’m in the south US and it’s more common than not). Usually “had” is dropped more than “have”.

    Also, English can drop the pronoun, article, and even copula for certain indicative statements. I think it’s specifically for observations, especially when the context is clear.

    looking at someone’s bracelet “Cool bracelet.” [That’s a]

    wakes upsigh Gotta get up and go to work…” [I’ve]

    “Ain’t no day for picking tomatoes like a Saturday.” [There]

    “No war but class war!” [There’s]

    “Forecast came in on the radio. Says there’s gonna be a hell of a lot of rain today.” [It said -> Says/Said]

    “Can’t count the number of Brits I’ve killed. Guess I’m just allergic to beans on toast.” [I; I]

    “House came tumblin’ down after the sinkhole opened up” [The]

    “I’d” can be “I would”, mainly if used with a conditional or certain conjunctions/contrastive statements (if, but, however, unfortunately). Also when preceding “have” – e.g. “I’d have done that”. Because “I had have” doesn’t make sense, nor does “I had <present tense>” anything. “I’d” as in “I had” is followed by a past participle.

    “I’d” is usually “I had” otherwise, forming the past perfect tense. But in “I’d better”, it’s a bit confusing because “had better” is used in a different sense – the “had” here comes from “have to” (as in “to be necessary to”) and can be treated as both a lexical verb and an auxiliary verb. “had better” is a bit of a leftover of more archaic constructions.







  • It would be a pain for developers, but firefox and chrome using a gig of ram to view webpages and play videos is horrendous even with isolated design.

    That can’t be helped. Hard to explain well without knowing how much CS you’re familiar with, but basically in order to guarantee security/user safety you have to sandbox each tab (basically running an entirely separate container program for each tab which constantly checks for illegal memory access to prevent it from being exploited), all separately running their own interpreters for javascript/typescript, HTML, CSS, all of which are very resource intensive (mainly javascript/typescript). There’s not really any getting around this, no matter how well you design your browser.

    Now, theoretically, with the growing popularity/advances in WebAssembly, and increase in usage of frameworks/graphics APIs like WebGPU, you could completely get rid of that sandboxing and completely get rid of the extremely slow javascript and html/css, in favor of completely using safe, compiled Rust programs. There’s active research using versions of WASM which only accept completely safe code (mainly safe Rust code) so using memory bugs generated from user error to access data in different tabs becomes impossible (aside from potential unaddressed bugs in Rust itself obviously) and you don’t need to sandbox each tab – the program practically sandboxes itself. Then you could potentially have browsers with thousands of tabs perform perfectly fine, assuming each of the websites is programmed competently.

    But that’s not going to happen, because billions of users rely on HTML/CSS and JS, and it’s not pretty to transition away from. Getting rid of it would be like getting rid of pointy shoes, or getting rid of US Customary Units in the US, it’s just not happening no matter how much benefit it would bring to users. It’s not so much of a browser company issue as it is everyone ever would complain and potentially trillions of dollars of damage would be done. Also frontend web devs can barely punch out a “hello world” program in JS so there’s no way most of them are gonna be touching Rust or Haskell or something.