Notices by Erin 💽✨ (erincandescent@akko.erincandescent.net), page 5

Once every two years or so I decide I’ll fly somewhere and without fail it’s an experience that does a perfect job of reminding me why I take the train everywhere even if it takes 12 hours longer sometimes

Two years ago: plane pulled up at wrong gate. Checked luggage went into wrong baggage handling company. Stuck at baggage collection for an hour. Today: security scanner jammed and ate my coat for an hour.

Fortunately the plane is half an hour late so I’m not dealing with “I missed my flight” frustration

@mcc @polotek I think the above captures this well.

There are dozens of existing projects in the Fediverse. Why did the Social Web Foundation decide to only invite three of the long term projects that have existed in this space?

And did the people who started FediForum (who were new to the space) reach out to any of the people behind the projects of the Fediverse? From my vantage point of talking to the people who work on lots of those projects, it sure didn’t sound like it.

As for whether it is truly volunteering: for a lot of projects, it truly is. Many of the projects in this ecosystem are developed by people in their spare time and they’re happy with that. But there’s also a reason I included the “(or grant supported)” paraenthetical, because by and large the projects which have people working on them full time are receiving a lot of funding from grant giving orgs.

Realistically what this comes down to is that a lot of people keep setting up organizations and things around the fediverse but if they’re talking to the people who’ve been building it they’ve certainly not been talking to many of them.

I think this is fundamentally it - I haven’t seen any of these things actually bring money or resources to any of the existing projects. Basically everything here exists due to a mixture of volunteers, donations, and grants from the NLNet Foundation (via the EU Next-Generation Internet funding schemes)

@cwebber (Perhaps this makes me the more jaded of the two of us)

@cwebber I’ve watched these dynamics play out on the Fediverse, on Twitter, on BlueSky, and in real life. They’re facts of human nature, and impossible for technical design to erradicate.

Of course this doesn’t mean that an OCap, non blocklist, potentially fully distributed model wouldn’t produce less of this. I just think blocklists and the human urge to exclude “bad actors” (however flawed that process may be) is universal and eternal.

@eniko @lofty yeah, its not well documented from a “this is how you implement a loader” POV (I guess people expect that every loader you will ever need has been written…)

For ELF PIE its just

1. Read the header
2. Read the program headers, which tell you what to load (and at what relative offsets to each other)
3. Find the relocation information (in REL or RELA formats) in the DYNamic table
4. Apply said relocation information to the binary
5. Jump to the entry point with whatever parameters you want setup

@eniko @lofty For ELF PIE executables, take the highest value of ph.p_vaddr + ph.p_memsz from the PT_LOAD segments in the file

(This assumes that the PIE executable was linked at 0 but that’s the most common case; if it wasn’t, then subtract the lowest value of ph.p_vaddr from the previous)