Data structure (over)design as a form of procrastination
@dansup When you take VC, you sell the company. No ifs, no buts. We keep getting surprised as the sale usually just takes several years to complete. C’est la “exit”.
This is why VC investment is incompatible with sustainable businesses with social missions. And why I keep harping on about how we must support projects that benefit the commons from the commons (https://ar.al/notes/encouraging-individual-sovereignty-and-a-healthy-commons/)
Library design: where every decision you make is sure to ruin someone's day at some point in the future
Compared to the Lisp compilers I grew up with JS engines have spectacular garbage collectors and are disappointing optimizers. I guess this often skews my perceptions of what's fast.
Well, that was a dead end worth three days of work, I guess. So it goes.
Maybe I've internalized the 'lines spent' idea a bit too well
I like ridiculously dense code and I cannot lie
I figured I might at some point have to leave gmail because of data handling. Hadn't expected they'd be so incompetent that things regress to the point of becoming unusable. Today the search box doesn't work in Firefox anymore. Editor focus and scrolling been broken for ages.
Pre-netflix: Too tired to do anything. Guess I'll go to bed early
[wake up well-rested next morning]
Now: Too tired for anything other than watching dumb series
[click 'next episode' until 2am, curse myself next morning]
I guess 'I deserve passive entertainment goddammit' isn't the best self-care angle
Hm, apparently #letsencrypt now provides mor than 50% of the web certificates.
On one hand: FUCKING AWESOME.
On the other: Shit. They're getting very powerful and we are centralizing our trust. I think we could reallu use 2-3 new orgs like Letsencrypt, with similar technology and mission statement, but entirely independent.
Are there any out there yet?
This is a great write-up on the kind of things ProseMirror is making possible: https://open.nytimes.com/building-a-text-editor-for-a-digital-first-newsroom-f1cb8367fc21
Automating the release process of ProseMirror was time well spent. Releasing often and with accurate release notes is much easier when the notes are in commit messages and the actual process that generates tags, commits, and changelogs is a script.
I still do have to run a command—not comfortable with having every bit of garbage I push to github immediately becoming a release, and I do like to combine a day's work into a single release without the complexity of a dev branch.
love to write code against a spec that's not kept in sync with the implementation
This toot was brought to you by me watching the a11ycasts videos, which have _great_ accessibility information but a terrible google polymer bias (and some rather bad code)
I don't feel any of the tech being added to the web under the name 'web components' is very helpful, and I'm worried that we're adding more complexity to an already overcomplex system to accommodate a dead end.