Show more

Doing research is the best type of procastination

Also, a lot of the UX innovations developed on VC money can be conveniently absorbed into the general culture and be applied in other contexts.

Weird to find myself saying good things about startup culture in several conversations lately. The theme was UX, and my point was that growth-obsessed companies have seriously upped the stakes in how much thought we put into interfaces. And that's a good thing—there was (and remains) way too much "RTFM" mentality in open source culture, and corporate software houses still don't really seem to get UX at all.

Autocomplete can spread a typo all over your code before you even notice it (I just corrected 8 instances of 'precededence')

New electronic message from a stranger...

1996: Wow, exciting, let's talk to them and maybe become friends.
2019: Report spam.

Marijn boosted

RT @colmmacc@twitter.com
Does anyone else feel that they basically lucked into a high paying career because their interests randomly aligned with scarcity and a rapidly growing field, and that it's nowhere near as hard, or as societally valuable as what teachers, medical professionals and care-givers do?

And if you read this toot and squat those names I will _end_ you. 😠

What's a better name, 'grammarian' or 'lezer' (reader, but sounds a bit like laser)?

Goddammit we've reached the point where every English term, as well as most variants of it, already exists on npm.

Time to start naming packages in Dutch, I suppose.

Hahah I'm going to call my parser 'parsley'. So clever.

[searches web]

Oh okay there exists a parser system named parsley in... every single programming language already. I guess that was a bit obvious.

Every time I second-guess this paper I end up realizing it was right after all. Efficient and Flexible Incremental Parsing researchgate.net/publication/2 . It's from 1998 but none of the more recent papers in this space describe as ambitious and practical a system as theirs.

It's just how my design process works, but the way I often spend days typing long-winded inner dialogues into text files could be reasonable cause for concern about my mental health in most contexts.

Marijn boosted

At this point the project I'm working on can only be described as "doing terrible things with parser tables", and it's starting to get fun

A/B testing as a replacement for empathy

So yacc/bison-style ambiguity resolution with precedence and associativity declarations is really a giant hack. Are there any less dodgy conventions/notations that do the same thing (disambiguate an LR automaton statically)?

Don't invasively mess with editable fields in web pages unless you know what you're doing—take it from someone who's spent a career doing just that. There's SO MUCH you can screw up, so many accessibility landmines and obscure workflows that you never even heard of.

This tweet was brought to you by Airbnb's message fields being completely unusable in Firefox today

Explanation: all the references to this type of parser (which litter LR parsing wikipedia) come from the same crank user

So "Canonical LR" parsers claim to be more powerful than LALR with similarly compact parse tables (en.wikipedia.org/wiki/Canonica). Which makes me wonder, if they are so great, how come I never heard of them? (The technique has existed since 1977.)

BNF and related forms seems a rather awkward way to express grammars. Is anyone aware of work that looks into other representation? (Equivalent ones, not talking about stuff like PEG or combinators.)

Show more
Mastodon

Server run by the main developers of the project 🐘 It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!