Replies

lzg, 5 hours ago to random

the worst part of genocide is now. but the second worst part is after 20-30 years when sooo many assholes will say to your face "there was no genocide" and netanyahu gets turned into an inoffensive grandpa who gets gushing obituaries in big newspapers. people will try to make you feel like you're exaggerating the horror, this old crazy person, like you didn't SEE IT. like you didn't see the mass graves.

tef, 5 hours ago

@lzg at some point we're gonna see tiki torch marchers flying the flag of david in support

i mean, far right people fly the confederate flag and the star of david

tef, 5 hours ago

@lzg northern ireland, same group

the red hand ulster flag with a star of david

glyph, 7 hours ago to random

@mcc me thinking about using systemd: "I don't understand why people are so mad about this, launchd is better than sysv init, this is just launchd for linux, I guess they couldn't just port it because some low-level stuff is different but I'm sure it's basically the same"

me actually using systemd: "fuck fuck what is this shit what the fuck why didn't they just actually port launchd to linux, there's no excuse for this"

tef, 7 hours ago

@glyph i feel entirely hoist by my own petard here as i'm normally the one bringing up display pdf's secret history

tef, 6 hours ago

@glyph so in the case of the linux kernel and devfs i think it was something like this

• let's say we integrate devfs, we now need to write policy over which device goes where, and then try and enforce this on vendors
• wait, device driver writers will 100% hard code a name across their driver, and distribute a binary kernel module
• maybe we'll need to have some way of managing this namespace to let users move things around
• that sounds an awful lot like what we did already, pre devfs

tef, 6 hours ago

@glyph now in an ideal world, the userland stub of the linux kernel would be developed inside the same repository

but well, that stub becomes a great place to do hotplug, to do suspend, to do service mangagement

as it actually knows what the hardware is, and when it's available...

and then, well, yeah, systemd

tef, 6 hours ago

@glyph again i'm not trying to say these are the best ideas around but yeah

moving things out of core development to go faster shouldn't come as any surprise to a python dev :-)

tef, 6 hours ago

@glyph alas as i repeatedly rediscover, writing a good interface between components is harder than writing components with a good enough interface

so although i do think we could live in a better world where we admit to ourselves we're writing a microkernel architecture, it's a lot harder than bolting freebsd to the side of mach and calling it a day

tef, 6 hours ago

@glyph i think the only sign i've matured as a developer is that i have lived through so many what-the-fucks that i can reverse engineer the good intentions that paved the way

well that and i don't use "butts" as a placeholder word as much anymore

tef, 6 hours ago

@glyph punting out to dns is a great idea but it does tend to remind people of Java, which is the main downside afaik

tef, 6 hours ago

@glyph @xgranade hi this is prolog calling

tef, 6 hours ago

@glyph i like telling people things like "a hard problem is one not amenable to brute force" or "a complex system is one where local changes have global effects" and sometimes "scale is when the one-size-fits all stops fitting"

but i do absolutely love detailing about the butchery in how the sausage is made, yes

tef, 6 hours ago

@xgranade @glyph

i think my fav version of this is the hype cycle

• a new idea arrives, and has a new name attached
• people attach this name to whatever they're doing so it looks cool
• the name now means whatever the most popular project says it is

i might stop here before i start breaking out into "roy fielding's thesis was about web browsers" though

tef, 6 hours ago

@glyph every day we stray further from god's light (using UUIDv4)

tef, 6 hours ago

@glyph @xgranade honestly, prolog's just regular expressions over sql, so if it did escape the lab it would be one less query language to write

tef, 6 hours ago

@glyph @xgranade i think my favourite thing about "building restful apis" is that we've gone on to build electron apps, or browsers, but for exactly one website each time

despite everyone using json no-one uses the same json, i guess

it is genuinely a miracle that html is so widely used in a mostly interoperable way, but yeah i said i wasn't going to go off on one