Notices by Adrian Cochrane (alcinnz@floss.social), page 64

I think I got *some* clarity reading the W3C spec, but I have to redesign my API...

I guess I'll be redesigning my API & CatTrap will definitely be worth a blogpost!

This is a wildly unpopular take but I honestly think it's a very good thing that these #LLMs hav been released with all their... let's call them "rough edges."

The potential for abuse is bad but that potential was probably going to be there whether or not the public had this opportunity to update their priors on what generated text looks like.

(1/)

Ah, seems in my initial API draft I forgot that width needs to be calculated separately from height...

I need to adjust my underlying datamodel & add a couple more tree-traversals for height computation...

Any nontrivial layout is a back & forth negotiation!

Now how exactly do percentages work in the CSS `height` property?

I guess I need to go ahead & read the spec now...

@faho I heard a horrorstory about a fear known as The Spider contracting a webdeveloper to build her a forum from scratch.

That webdeveloper was really good at just going along with things, and the forum became the subject of an urban rumour.

@vertigo Thanks! This feedback is useful!

From a lot of this feedback I get the impression that I'm largely bypassing the most hated bits by using SourceHut todo. I may not have the most optimal tooling on my side, but I'm willing to improve them as I hit their inconveniences.

Here's my conclusions from the feedback: https://floss.social/@alcinnz/109870608563040270

I'm finding the UX very seamless with native mail clients, but I'll ensure it works well with webmail.

It'd be good to figure out how to best encourage devs to be more open-minded to trying to something new, since it's easy to find ones who aren't. Any good writers keen to help?

Also it'd be good to stop SourceHut from telling people off for using HTML email... Given the right preprocessing command I should be able to configure to Maddy to do so.

https://maddy.email/reference/checks/command/

Any pointers to something I can hook in here?

Mostly I'm basically bypassing the most hated part of email-based git.

2/2

Based on feedback regarding email-based git, I think I've got a good-enough email-based contribution system for now. Where I'm essentially using email as an authenticated comment box.

It is possible I'll find shortcomings on my end, & it'd be good to explore alternatives when I'm ready to trial extending the web.

But until then I'll be happy to offer email addresses to those who wish to keep their's private. I have!

I'll be quite lenient in you attach your contributions.

1/2

@lightweight @msw @lwriemen @downey @skyfaller I find this one useful as a means to diagnose a problem in our communities, but I do agree I should avoid applying it to specific individuals, sites, etc. Which I failed to handle to hold myself to this time.

The dangers of autocomplete - Junaid Mubeen & Kawther Alfasi: https://junaidmubeen.substack.com/p/the-dangers-of-autocomplete (SubStack)

Automating podcast transcripts on my Mac with OpenAI Whisper - Jason Snell @ six colors: https://sixcolors.com/post/2023/02/automating-podcast-transcripts-on-my-mac-with-openai-whisper/

Is ChatGPT's use of people's data even legal? - Emmas McGowen @ Avast: https://blog.avast.com/chatgpt-data-use-legal

@froomkin The Magnus Archives! Lots of great ones there!

@wizzwizz4 probably not, as it would require a lot of resources - like what a gov't regional development fund might be applied to, if gov'ts were smart about tech and saw #FOSS as digital 'critical infrastructure' and funded it like they fund physical critical infrastructure. Sadly, pretty uniformly around the globe, they're not. Not even slightly. They're almost all entirely in the pockets of big tech. Corruption is pervasive.

2/

For example, there is no HTTP method to list the contents of a directory.

Also, there is no HTTP method to pull all the contents of a directory (as a single operation).

There are no recursive HTTP methods for directories.

Etc.

1/

The #WorldWideWeb dominated #Internet would have been a different place if #HTTP had support for directories.

Because except for an atavism on how relative URIs resolve against a base URI —

HTTP does not support directories.

Modern AI and the end of privacy - Ploum: https://ploum.net/2023-02-15-ai-and-privacy.html

Bing: "I will not harm you unless you harm me first" - Simon Willison: http://simonwillison.net/2023/Feb/15/bing/#atom-everything

The Curse of Academy Engraved LET - Robin Rendle: https://robinrendle.com/notes/the-curse-of-academy-engraved-let/

@msw @lwriemen @lightweight @downey @skyfaller I shun TechRights since their distrust of corporate influence goes to ludicrous extremes. To the extent they lump elementary LLC, who by absolutely no means qualifies as such, in with the major corporate players trying to "take over Linux".

It is opinions like their's that I am characterizing when I use the term "fossbro", to avoid naming names.

@raboof Depends on whether you use Todo or Lists to accept patches. I'm using Todo, treating pull requests like bugs. Todo provides these features.

@silmathoron Oh, and thanks!

@silmathoron Yeah, I've seen the unfamiliarity issue before... Difficult to address an unwillingness to try something new...

Hmmm, I'm curious whether I'll be bump against those review inconveniences? I guess I'm willing to explore fixes if/when I feel I need them!

I can certainly understand for a degree of anonymity, or to avoid spam. I've got short-term & long-term plans to address that need...