IRC channel logs
2023-02-08.log
back to list of logs
<dthompson>unrelated to compilers but wanted to share :) <flatwhatson>oh nice! steps towards some tracing/debugging functionality? <flatwhatson>having some good tools for this early will pay off hugely. trying to rebuild this stuff post-incident from log fragments spread over multiple systems is a huge time-waster <dthompson>oh yeah I've done a lot of that over the years <drakonis>i'm eagerly awaiting the future where i can make erlang grade software with scheme <drakonis>that and propagators/minikanren goodness also available in guile <drakonis>i keep looking into prolog and finding it very cool but mildly impenetrable <drakonis>cwebber: but for real though, when is will byrd getting a job offer to do cool logic programming stuff in schemes? <Zarutian_iPad>re this causeway like tool dthompson is writing: how well will it handle other tenses other than the occured? <Zarutian_iPad>say I have a program that might emitt diffrent message timelines depending on some data <EWEW>I didn't realize there was a channel for spritely until just now. Hello everyonei :) <EWEW>Thanks! Something maybe funny and related to the tech here for object cap. Whenever I read CAP/cap now I have to say to myself CAP' (CAP prime) in my head because I otherwise immediately fill it in to be CAP theorem. <shauna>I am generally not a fan of the Silicon Valley trend to just make up random words for their apps and things, but it does have the benefit of creating names that are not completely overloaded <shauna>I'm glad that people seem to be embracing non-.com TLDs, iirc the made up names are at least partly coming from domain name squatting on all the actual words in the dictionary <sneek>Welcome back Zarutian_iPad!! <EWEW>shauna: I agree with you. To try and put it into words the major problem(s) with random words for *things* SV tech puts out. It's not that they're new words. It's that there's a tension between either extending a causal chain of referents, an initial creation of a referent (a truly new word, problematic for widely accepted correct interpretation for some period of time), and an initial creation of a <EWEW>referent that doesn't share meaning with the causal chain(s) of referent(s) that are visually/morphologically/acoustically similar. I'd say the really bad, not tasteful, ways of naming *things* is the third one *always*. And if it's something that needs to be widely known(interpreted correctly) and simultaneously is *not* a truly novel/groundbreaking concept, then it's also imo not tasteful to make this <EWEW>new referent in the third way described. I say that because there could be (and probably are) competing referents that could be overloaded just fine and also due to the initially problem described of getting a new referent "out there" so that it's interpreted properly. But in general these are necessary evils *sometimes.* Referents can get confused not because of visual/acoustic similarity but because of <EWEW>too much overloading in *addition to* not enough context. Because if we only do overloading of words then we end up with descriptions that get too long to the point of impracticality because you need more description, context, as overloading increases to disambiguate(interpret correctly) some word. Dumb example: Base 10: 2000, Base 2: 11111010000. In this example we're overloading in a sense by having a <EWEW>smaller set of symbols. Also notice what i said in the last case about similarities(visual/acoustic) regardless of a good causal chain extension. Base 10: 2000 and 2001 -- Base 2: 11111010000 11111010001. By going from Base 2 to Base 10, we add new "words" which allows for more succinct usage. And if we go to a weird larger base say, Base 33 the problems of something not being quickly/widely interpretable <EWEW>show up. Base 10: 2000 -- Base 33: 1RK. <EWEW>TL;DR FizzlePop is a shit name for an error logging company because it sounds like a soda company. <EWEW>And ryaskrat is a terrible name for a company because i have no idea what that would even be. <EWEW>(until they're worldwide successful *despite* their name) <EWEW>But ryaskrat's are necessary, so that every company isn't "errorlogcompany". So sometimes it's fine. 
