no WASM in 1991

Superlogical has officially named their terminal multiplexing app “Rex”.

I have to believe that someone vetoed the name “Ttyrannosaurus”.

*Youth Pastor Voice*

You know what ISN’T worried about the eMudhra certificate compromise? That's right: #NeXT machines (and everything else 35+ years old).

TIL: Anthropic on GitHub is officially `anthropics`, not `anthropic`.

Anyway, it's weird that there are widespread supply chain attacks, lately, right?

@foulab (Just noticed I put this in the wrong thread. Oops. This was supposed to be linked to my NeXT machine ramblings: https://scoat.es/@sean/117374736710583768 )

@foulab I got early WorldWideWeb.app running. It declares `Copyright 1990,91, CERN.`

1) it only speaks HTTP/0.9. That means no name-based hosting. So basically nothing on the 2026 Web works.

…there are ways around that… but…

2) There's obviously no TLS (nor SSL) here. And even if I manage to build and use something like BearSSL… this is a 25MHz CPU, so… probably looking at almost a minute to handshake HTTPS connections. So basically nothing on the 2026 Web works.

This is a weird project. (-:

@foulab Finally some progress on the #NeXT Slab… it boots! I made a disk image from the legit original NeXTSTEP 3.2 CD, and installed the OS in Previous (emulator), copied to SD card and hacked up this cabling mess.

@foulab I keep hitting dead end after dead end. If anyone in #Montreal has an internal (50pin) #BlueSCSI I can borrow (or buy in the next couple days), please let me know. I’d also take some working SCSI (not ultra) hard drives in lieu. 😴

If you're coming to #VCF #Montreal ( #VCFMontreal ), and are a #MeshCore user, please bring your companion with you.

If you do, you'll be able to interact with a couple of my anachronistic vintage computing mesh projects (which will be at the @foulab exhibitor table).

Adding “experience installing NeXTSTEP from scratch from SCSI CD-ROM” to my [un]marketable skills.

This is not shade toward the developers, but the #php2shell vulnerability was reported in August and the fix was published in October.

I have to think that increasing funding for a project as big as #PHP ( @thephpf ) could have reduced that timeline. In the era of analysis heavy machinery, it’s pretty unlikely that *only* the reporting party discovered this vulnerability. https://mastodon.social/@sovtechfund/117491065698939254

An interesting (to me, at least) recurring conversation I have with people at @foulab who are visiting the museum, when they ask about the NeXT machines, goes something like this:

What OS is that?
It's NeXTSTEP.
Weird.
Yeah, it's BSDish. But very different. It's roughly what became Mac OS X around 25 years ago. That evolved into macOS today. I write macOS/iOS apps sometimes, and we *still* use the `NS` prefix for some APIs. There's direct lineage from this machine to the iPhone in your pocket.

“Teaching” the NES client to speak [a very small subset of] Unicode is… an "interesting" problem.

(Scare-quotes absolutely intentional.)

The plot thickens… (I'm not 100% sure this display (that we had on a storage shelf) was originally attached to the Cube, but at this point, I think it was)

#McGill #NeXT #NeXTcube

This might be reckless, but I do respect the boldness.

(`mitmproxy` added `—emudhra` and the leaked certs/keys for easy not-even-self-signed request forging)

https://github.com/mitmproxy/mitmproxy/commit/169234499088fc2c414644a67efaed3d64ae2e3b

If you think today's stop-everything-and-patch-now critical vulnerabilities were a problem, you just wait until half of the vehicles on the road are meshed with Starlink and they act as one big Tesla swarm to “own" the other brands by totally-accidentally running all non-texan cars off the road and *totally-accidentally* dangerously punishing anyone who profiles as someone who might disagree with Fleet Commander in Chief.

Tried to participate in tonight’s local #Montreal #meshcore “net” (roll call) with #MeshNES and captured a brownout on camera. This actually explains why some of my messages get dropped (long ones take longer sustained current on the cart rails). Good data. Need to fix this before #VCF. Buffer caps are definitely out of my wheelhouse.

If you've worked anywhere around me in the past 15 years, you've probably heard me say something like “this software is bad. all software is bad. even my own.”

Today's (well, August’s… well, 2020 and beyond’s) PHP vulnerabilities don’t mean PHP is bad, but it does fit into my thesis that *all* software is bad.

And the reason it is bad is that—in the names of progress and efficiency and profit—we have built software at 5% of what it ACTUALLY costs to build software, for the past 30 plus years.

If Dead Internet Threory wasn’t real before today, try adding a “everything powered by PHP is offline for maintenance” aspect to the narrative.

Everywhere I go feels down right now.

Despite much of this reading as a confirmation of what I’d fear about working for a government organization, and a sad account of why things didn’t work as well as they’d hoped, I’m so glad @sboots wrote about this.

I’ve been casually curious about these kinds of mechanics since CDS really impressed me when I took the opportunity to interact with them during the pandemic: https://seancoates.com/blogs/how-i-helped-fix-canadas-covid-alert-app https://mastodon.sboots.ca/@sboots/117446091056210663