n8   Doesn't follow you    @n8@mastodon.social

ꜱʜᴏᴡ ʟᴇꜱꜱ

@paperdigits Also let's all give the finger to whoever it was at Autocorrect HQ who decided that "wiki's" had higher priority in the word list than "wikis". 😠 Once I'm in charge, there are going to be some purges over there. Entire divisions.

@paperdigits Do you have any pointers / example projects of folks who have built a wiki site using SSGs?

I mean real, honest-to-jimmy "multiple people make edits when they feel like it" wiki's?

Okay I'm gonna come right on out and say what everybody's thinking: software text handling needs to have "short-message–breakpoint" dictionaries for languages, in addition to hyphenation dictionaries.

Brotographer.

Dunno why it took me this long to realize, but I did not see NEARLY ENOUGH pics of pets wearing #Halloween costumes this year.

Please correct this now thanksxoxo

We all die a little bit on the inside every time somebody says "innersource."

In unrelated news, I am coining the term "uppersource". Don't know what it will mean yet, but whatever it is it needs to make everyone who unironically says "innersource" seethe with rage.

Once I get this time machine working, priority #1 is heading back to the 1960s to slap some sense into a few researchers who insisted on leaving only handwritten notes in their personal papers. We can deal with the Hitler stuff later.

It's a standalone Windows-or-OSX-app in Qt with a clunky UI.

The code is AGPL; I am unclear if there are any patents involved on the algorithms. But the author did publish a lot of background material on his personal site, which may preclude some claims on it being all trade-secrety. If anybody felt like re-implementing it.

This project "SmartDeblur" is nearly 9 years old, but it works miracles for recovering text. https://github.com/Y-Vladimir/SmartDeblur/

Really interesting blog post here from somebody who used to do all self-hosting, and has decided to stop: https://www.happyassassin.net/posts/2020/11/24/site-and-blog-migration/

(lots of other bits in there, as this is a "tech refresh" post; the self-hosting is just part of it)

"It's Doomsday Somewhere! ™"
– let me know if you'd like to buy this slogan on a t-shirt

Also, it may very well be public knowledge who the account holder is and I just never encountered that; trying to deduce it was still entertaining though.

I don't want any spoilers on the following:
- On Twitter, I follow the @FakeUnicode account and it follows me back
- It's great
- AFAIK, the account-holder's real name isn't public knowledge
- I have not tried to find out who it is via searching/asking, because I wanted to see if I could figure it out just by context and/or deduction
- I think I have figured it out
- That said, it's not clear how to test my answer under reliable conditions
- I realize virtually nobody thinks this is interesting

This sort of thing is why the people who work on text-rendering support do not drop everything and spend hundreds of hours to ensure that your ancient ASCII bitmap font looks pixel-for-pixel the same today as it did under Cygwin WindowsNT.

(So does "unfinished laser cut birch plywood with the burn marks still on every edge.")

I don't have any desire to use an Amazon Alexa or Google Home speaker/mic, but I would like some pointers on how to make those fabric-covered enclosures that they use for projects of my own.

The "just 3D printed in some weird random color" look of most DIY hardware stuff makes me sad.

Can somebody please check in on Matthias?

Today in Pango, someone has opened a bug to complain that the letterspacing of the Microsoft bitmap fonts they extracted from Windows 10 don't look right on a modern GNOME desktop system.

FYI, I now bookmark all of these stories for reference later on, in the annual LGM planning process, when the inevitable barrage of Eurodudes objecting to any suggestion of hosting the event in the US because of "its policies" starts up again.

I should clarify that I do not. Though I am still trying.