mfsch

  > Added 20442 new packages
  > Updated 20641 existing packages
  > Removed 17532 outdated packages

Perhaps of interest: I came across this retrospective of the author of that site and how he moved it to static hosting a couple of years ago [1].

I would also say that this library covers more or less the “lower half” of solo ball juggling in terms of difficulty. With lower ball counts (say ≤ 4), there are a lot of these patterns that have complex arm movements and can be difficult to explain with words, so having such a listing with animations and step-by-step instructions is very valuable. Starting with 4 balls, there’s less and less time for moving your arms around and it is more about the sequence of heights of the throws, which are well described with just their numeric “siteswap” pattern and you can learn them just from knowing the number sequence. The site has only the most basic of those (e.g. 534) and even very common 4-ball (7531, 633) patterns are missing with hardly anything beyond that.

[1] https://ianconvy.github.io/projects/other/libraryofjuggling/...

If you are just creating a simple document with default styling, the main advantage you get from Typst is near-instant compilation speed. Pandoc to HTML is similar though, but if you’re generating PDFs with LaTeX the compilation delays can be pretty annoying.

If you are creating more complex documents, the advantages become more pronounced. Styling in Pandoc means modifying templates, at which point you’re just writing LaTeX, and styling in Typst is much nicer than in LaTeX. You can also hit the limits of Pandoc templates quite easily, at which point you have to write Lua filters. I have found those to be quite cumbersome, and now your document logic is spread out over the Markdown source file, the LaTeX template, and the Lua filters. In Typst you can have a single file with your whole document in a clean modern format, and you can decide for yourself how much you want to separate content and presentation.

It’s also what Thunderbird is using to build their paid email hosting. Seems like a very ambitious project mostly done by a single person – impressive!

The issue I have with this is that it deletes the whole message, not just the media. In WhatsApp, you can delete media from the images/video folders and the messages remain in the conversation, they even still have the blurry preview iirc. In Signal, you end up with gaps in your history instead.

From what I can gather, they are not yet incorporated and they are working through organizational questions in [1] and the issues thereof. The `ACCOUNTS.md` file there gives an idea about the main people behind the project and the donation page on Open Collective [2] also documents team members and how they spend those donations.

[1] https://codeberg.org/comaps/Governance [2]: https://opencollective.com/comaps

Technically the big O notation denotes an upper bound, i.e. it doesn’t mean “grows as fast as” but “grows at most as fast as”. This means that any algorithm that’s O(n²) is also O(n³) and O(n⁴) etc., but we usually try to give the smallest power since that’s the most useful information. The letter used for “grows as fast as” is big Theta: https://en.wikipedia.org/wiki/Big_O_notation#Use_in_computer...

That’s exciting! Is there a summary of what’s new in 1.0 somewhere? Usually the version announcements contain an overview of the most important changes, but this one is more of a retrospective and general overview apparently.

Does “no lost data” mean that all drives were wiped or simply that there were backups?

You could still move e.g. /nix to the BTRFS disks and only keep /boot on the thumb drive, no? That’s more or less how I have NixOS set up with ZFS. But your solution is of course much fancier :D

Phoronix has looked into some the financial information available from their filings: https://www.phoronix.com/news/Purism-75M

I’d love for Purism to succeed and they appear to be doing a lot of valuable work in the software space, but I can’t help feeling like that 75M self-assessed valuation is a bit on the optimistic side…

I’m not sure about these stable packages, but for the nightly packages they introduced recently they explicitly mention on [1] that you can keep browsing:

Following community discussions, we have updated the post to highlight that Firefox can continue browsing after an APT upgrade, allowing people to restart at their convenience.

[1] https://blog.nightly.mozilla.org/2023/10/30/introducing-mozi...

From what I can tell, the “Ubuntu Mozilla Team” does not consist of people from Mozilla but rather just people packaging Mozilla software for Ubuntu. The latest packages in that PPA have been uploaded by Rico Tzschichholz, who does not appear to be affiliated with Mozilla.

I’m continuously impressed by the care and effort that goes into the developer experience of Elixir. Those improvements of the (already pretty great) documentation and the error messages look like they will make a real difference.

Related: The podcast “On the Media” had a story recently about how infrasound may have affected pidgeon races – highly recommended!

https://www.wnycstudios.org/podcasts/otm/segments/curious-ca...

Since this font is published under AGPL – is Times New Roman available under an open license? Or is this based on an open alternative?

The way I see it, the problem is the C in CSS. If you rely on `.author-bio > h2`, then every time you make a change to `.author-bio`, you have to consider all the contexts in which this class may be used globally in your site, and if you change the styles for `h2` even more so. Sure you can say you encapsulate every context with a globally unique id, but since you’re probably already using some kind of templating system for you HTML, it’s easier to specify the styles there. It’s basically the difference between having variables scoped global vs. local by default.

I would say there is one more problem it can solve, though less related to sustainability: Building up a modular system may make it easier to more offer customers a device close to their ideal configuration with a limited number of SKUs.

One distinction that I had missed for a long time is that the en-dash is used instead of the hyphen to connect words in the “and” sense, such as “read–eval–print loop” or “Myers–Briggs personality type”. I find that the en-dash makes it a bit more clear that the words are sort of “on equal footing” and it’s not one word modifying the other one.

I wish it was possible to add FIDO keys to an account without having physical access to the key. Without this, it is hard to balance the convenience of adding your keys to new accounts and the risk of losing all your keys. Ideally, I’d want to keep one key in a safe location far away and just have some public key data that I can upload to new accounts. Does anyone know why FIDO doesn’t work this way? Is it simply to make it harder to lock yourself out of an account?

I’ve always liked the idea of Netlify CMS, to have some (statically deployed) JS that talks directly to GitHub/GitLab and provides a clean interface to edit the content that’s stored in Markdown files. There are many sites that don’t need much CMS functionality but for which you’d still like to provide a way for someone non-technical to update some content. Being able to do that while keeping all code and data in a single repository is pretty neat.

However, I’ve been hesitant to use Netlify CMS in any project because it appeared to be mostly abandoned for a long time now [1]. It remains to be seen whether “PM” will be able to revive the project as they are planning to [2]. An earlier fork [3] seems more promising to me at the moment.

[1] https://answers.netlify.com/t/is-this-project-dead/70988 [2]: https://techhub.p-m.si/insights/introducing-decap-cms/ [3]: https://www.staticcms.org/

You can also use `git init --separate-git-dir=~/something /etc/nixos` to keep the git data in your user directory while the config remains in the default location. I like this solution because it keeps the stuff I want to do as root (configuring the system) separate from what I want to do as regular user (tracking changes of the configuration).

99% Invisible did a podcast about P-22 very recently – worth a listen if you want to learn about this story.

https://99percentinvisible.org/episode/cougar-town/

I wish something like this was used for the Signal message history, similar to what WhatsApp does. As it is now, you either have an ever growing message database full of pictures you don’t need to keep around, or you start deleting them and end up with weird gaps in your message history. Keeping a few bytes of data as a placeholder for deleted images would be optimal in my opinion.

StreetComplete has actually been adding a new feature called “overlays” recently, which let you switch the map to a special mode for editing a certain class of features. The upcoming version 48 has an overlay for shops and street addresses, which might cover much of what Every Door is aiming at: https://github.com/streetcomplete/StreetComplete/releases

I’m a bit conflicted about the goal of separating the “hard, cold, objective” science from “soft, warm, subjective” things such as reputation and trust. It may be the case that some publications can be perfectly evaluated based purely on the text of the manuscript, but in most of the work I’ve read and written there are countless little details that are not discussed in the text. The manner in which the authors have handled such details can greatly affect the quality of the research. How carefully did they write the one-off script that produced the figure? How closely did they follow the study protocols? How much did they look out for issues that might undermine the results? Such an evaluation may only be possible after following the work of an author over time and perhaps through direct collaborations.

In practice, we may get better outcomes if peer reviewers completely ignore these aspects and evaluate all papers purely based on what’s on the page. But I don’t think it is obvious that any reliance on trust and reputation should be derided as bias to be eliminated.

Perhaps worth mentioning that the Phoenix creator Chris McCord has written a number of posts on the Fly.io blog since he joined them a year ago (exactly a year today?): https://fly.io/blog/how-we-got-to-liveview/ https://fly.io/blog/how-safari-ruined-my-tuesday/ https://fly.io/blog/livebeats/ https://fly.io/blog/liveview-its-alive/

I guess that’s part of the reason why the Phoenix blog is less active?

Looks like this is the case for apex domains, but not for subdomains, for which you simply need to configure a CNAME record: https://developers.cloudflare.com/pages/platform/custom-doma...

It’s worth noting that the quoted article was published on April 1 [1]. While the PineBuds were confirmed in the following community update [2], the PinePod plans are only mentioned in the description of the development board as a “platform for earbuds and a digital audio player”. It seems to me that the plans for a future DAP are still quite nebulous and I wouldn’t give that much weight to the wording used in the April Fools’ announcement.

[1] https://www.pine64.org/2022/04/01/introducing-the-pinebuds-a... [2]: https://www.pine64.org/2022/04/15/april-update-no-more-unico...

I assume that the requirements for a browser implementation are more strict than for existing pre-processors such as Sass because 1) the performance budget is much more limited for interactive use compared to pre-processing and 2) browsers need to handle adversarial inputs gracefully. Are there any other reasons?