IRC channel logs

<muurkha>oriansj: yeah, that kind of thing

<oriansj>muurkha: thanks for the inspiration. Hopefully the second draft is better: https://pdp10.guru/stage0

<Hagfish>"But there was a quiet part unspoken."

<Hagfish>yeah, that's like a freedom -1

<Hagfish>the freedom of having the world make sense / your binaries do what the source says they do

<Hagfish>yeah, that page is nice

<Hagfish>it's a good length, very readable and well presented

<Hagfish>the combat / embarrassment analogies seem a bit out of place, but i don't know how else it could be presented as a genuine human story

<oriansj>Hagfish: now the real question, is it good enough to send to RMS (as he asked for an article on this topic)

<Hagfish>hmm, that's tricky

<Hagfish>i think he might relate to it, but i'm not sure what his expectations were

<Hagfish>can you point me to his request?

<oriansj>here is his original: https://paste.debian.net/1243151/

<Hagfish>you just need to get the balance right between, on the one hand, letting him see the similarity of his struggle and yours, and on the other hand, not making it look like his efforts were incomplete and fundamentally flawed

<Hagfish>i think the way it's written keeps that balance well

<Hagfish>hmm, yeah, his email talks about gnu.org and FSF people, so that gives an idea of the audience he has in mind

<oriansj>Hagfish: I wouldn't say fundamentally flawed, just perhaps never could have imagined we would have allowed ourselves to fall this far

<Hagfish>yeah

<Hagfish>you could try to pitch it as a natural extension of the FSF's vision, which they have enabled

<Hagfish>they've made the toolchains, and given the community certain values, and now there is some historical baggage that needs to be cleaned up

<Hagfish>i'm not sure how much of your own personal journey you want to highlight in the article

<oriansj>well, I could write it as a love note to the FSF/GNU community

<Hagfish>yeah, i think they'd appreciate that

<Hagfish>it's almost like a "declaration of independence" for free software, isn't it?

<Hagfish>the GNU project grew out of commercial software

<Hagfish>and didn't even have its own kernel until Linus came along

<oriansj>here I was thinking a Calvin Weir-Fields/Ruby Sparks self-reflection

<Hagfish>i'm not familiar with that

<oriansj>a very campy B movie about a writer which everything he writes becomes real

<Hagfish>ah, i think i've heard of that, actually

<oriansj>So being lonely, he writes the girl of his dreams and discovers the truth about himself

<Hagfish>is stallman the writer or the girl?

<oriansj>I am thinking the girl and this is me discovering the unspoken bit that I should have known

<Hagfish>i think Stallman is an idealist, and it will resonate with him to think that you've pursued this long term goal and built a community around it

<Hagfish>if a goal is important enough, it almost doesn't matter if it's achievable or not, you have to commit to it, because what if...

<Hagfish>didn't someone famous say "i didn't know enough to know that it wasn't possible"?

<oriansj>and incase anyone isn't familiar yet: https://rms.sexy/

<oriansj>and my latest draft is now up: https://pdp10.guru/stage0

<Hagfish>hmm

<Hagfish>i think it's a better analogy than combat, and it's certainly poetic

<Hagfish>i dunno, i just think it requires the reader to assemble too much context

<Hagfish>if they don't know who you are, or what the technical problem is that you're interested in, then the analogy obscures more than it enlightens

<Hagfish>i guess you're trying to convey that it's like an obsession, and that's good

<Hagfish>"a splinter in your mind", as The Matrix would say

<Hagfish>Morpheus, specifically, i guess

<Hagfish>it's somewhat reminiscent of Homer's "do it for her" poster, which is good

<Hagfish>it's like this project is your baby, but maybe the GNU people would think that the free software ecosystem was their baby and she's all grown up now

<Hagfish>or maybe it's your white whale that you know is out there and you're chasing down

<Hagfish>that's a slightly combative analogy again, but hopefully no one thinks the phrase advocates animal cruelty

<oriansj>here I was thinking Deon and Chappie

<tinybronca[m]>Sorry but does this help simplify the MANY-stage bootstrap for modern rustc (much of which might not be reproducably built): https://blog.antoyo.xyz/rustc_codegen_gcc-progress-report-10

<oriansj>tinybronca[m]: I don't know yet, someone will have to see if the Guix build path becomes shorter or not

<tinybronca[m]>I don't know if this is on anyone's radar there but ok

<oriansj>tinybronca[m]: well it isn't a top priority because we do have a working bootstrap chain for rust but yes improvements are a core support task that needs to be done to preserve the bootstrap chains we do have.

<oriansj>well, I think I will stop here: https://pdp10.guru/stage0 for the night. Need more inspiration if I am to take it deeper in a direction

<Hagfish>oriansj: i think it's getting there

<Hagfish>you start with a good hook, "The unspoken responsibility", and then could just jump straight in with "We have an embrassing problem." (which is enough human narrative), then the 4 Freedoms stuff is a perfect philosophical introduction which connects nicely into "But there was a quiet part unspoken."

<Hagfish>everything builds nicely and naturally from there, just like bootstrapping itself :)

<tinybronca[m]>This wiki has no edits or logs made within the last 45 days, therefore it is marked as inactive. If you would like to prevent this wiki from being closed, please start showing signs of activity here. If there are no signs of this wiki being used within the next 15 days, this wiki may be closed per the Dormancy Policy.

<tinybronca[m]>This is weird they close the wiki??

<pabs3>tinybronca[m]: that might be a core mediawiki feature rather than something the wiki admin setup

<stikonas[m]>tinybronca: I don't think it simplified rustc bootstrap path. Rustc would just use GCC backend rather then llvm but it would still need chain of rustc's

<Irvise_>I rhink that is the goal of mrust.

<AwesomeAdam54321>Would gcc-rust shorten the Rust programming language bootstrap chain?

<stikonas[m]>I guess this one might eventually: https://github.com/Rust-GCC/gccrs

<AwesomeAdam54321>yeah that's the one I'm referring to

<stikonas[m]>oriansj: that article reads better now.

<stikonas[m]>I wonder if it should mention kaem

<stikonas[m]>But on the other hand, other non-compiler tools are also skipped (like blood-elf)

<oriansj>I think it has been tweaked enough. I am stopping changes on https://pdp10.guru/stage0 now and will send it to RMS.

<tinybronca[m]>I wonder what distro rms uses these days

<tinybronca[m]>ask per that 😁

<onebitboy> https://stallman.org/stallman-computing.html

<tinybronca[m]>oriansj there is big broken "image" not sure it is very long and links to https://en.wikipedia.org/wiki/Sandpainting but it does not load

<oriansj>tinybronca[m]: what distro a person uses is a meaningless question. One could be using hundreds of distros or made their own. And even if they said Guix, it still tells you nothing meaningful as tomorrow it could be Dragora or Ututo or Hyperbola

<oriansj>tinybronca[m]: if you look at the html, the image name is sand_it_gets_everywhere.jpg and it is on the server you get the html from

<tinybronca[m]>okay well I still would find it interesting, but that is just me perhaps

<tinybronca[m]>Like I thought the MIPS Longsoon laptop was interesting

<oriansj>tinybronca[m]: indeed, just like I find the TV series Jeremiah interesting but just like showtime, most people didn't find it as interesting as I did.

<oriansj> https://en.wikipedia.org/wiki/Jeremiah_(TV_series)

<oriansj>and just like Dr Stone https://www.crunchyroll.com/dr-stone/videos not everyone shares my interests; usually people just put up with them when they find them useful.

<oriansj>also you could always email rms@gnu.org tinybronca[m] if you wanted to communicate with stallman, he does read all of his email and will respond.

<oriansj>just understand if your email looks like spam or trolling it'll likely just get filtered out.

<tinybronca[m]>Hmmm everyone dies but no zombies?

<oriansj>tinybronca[m]: the death of knowledge is just as deadly as zombies

<tinybronca[m]>oh my

<tinybronca[m]>what if zombies were good at hacking

<oriansj>as a small child, could you survive after the stores ran out of food?

<oriansj>or did you know how to make drinkable water

<oriansj>or that a single scratch could kill if not properly cared for

<oriansj>tinybronca[m]: you should write up that concept as a tv series and pitch it to Netflix, I hear they will make anything these days

<tinybronca[m]>Hahahahaha

<tinybronca[m]>I honestly don't know how Netflix makes enough money for original series..... they pay licensing fees for other stuff, huge RAID arrays for video storage and tons of bandwidth costs..... and I think the subscription price is very low

<oriansj>tinybronca[m]: 1PB of storage is pretty cheap; bulk bandwith is fractions of a penny per GB ($1/TB if I remember correctly)

<tinybronca[m]>1PB is cheap?? hmmmm

<tinybronca[m]>guess is relative

<tinybronca[m]>cuz you cant buy consumer or SMR drives

<oriansj>assuming 1M customers streamming 720p 24/7 but paying $1/month; ignoring licensing fees, they should be netting $700K/month

<tinybronca[m]>proper enterprise NAS HDDs are more and you need the rest of the machines as well and redundancy in distributed CDN-style networks

<oriansj>tinybronca[m]: sure you could and backblaze does

<oriansj>but sure let us just imagine $1K per 16TB

<tinybronca[m]>Also $1/TB bandwidth sounds a lot cheaper than I have seen in EU datacenters but I haven't checked in last couple years

<oriansj>so $64K for 1PB

<oriansj>So someone making $120K/year can afford that as easily as a new car

<oriansj>but you can get bigger drives for cheaper

<tinybronca[m]>bigger than 16TB?

<tinybronca[m]>I read they were going to make >20GB but SMR only

<oriansj>tinybronca[m]: SMR doesn't matter at this scale

<oriansj>also you can buy 100TB SSDs for $40K

<tinybronca[m]>Does not seem smart in hot arrays

<oriansj>So even if they do 10 100TB SSDs, it isn't even half what they make as profit in a month (assuming $1/month subscription)

<tinybronca[m]>They would need more than 10 of them however, but ok

<oriansj>fine 11 to have 1100TB which is more than 1024TB

<oriansj>we are still talking about meaninglessly small sums for a business with $29.7 billion in revenue

<oriansj>and $5.116 billion in profits

<oriansj>221.6 million paid subscribers at even $1/month is still 1: 2659.2 million

<oriansj>$2.6592 Billion per year

<oriansj>So even doubling the cost to $1M per PB of storage, you could afford to do that more than a thousand times over

<oriansj>and still have $1.6592 Billion for staffing and bandwidth costs

<oriansj>The reality is today, computers/storage/bandwidth are cheap

<oriansj>heck, even if you have a copy of every single manual and binary distributed in the http://www.bitsavers.org/ archive, it is still only 825gb

<oriansj>more than one could read in a dozen lifetimes

<oriansj>a single 16TB stores enough music to contain hundreds of lifetimes worth of music

<oriansj>right now a lifetime of 360p video can fit on a single hard drive for $400

<oriansj>in a few years its'll be 720p

<oriansj>and a decade after that it'll be 1040p

<oriansj>or even 4K if we do more than just grow hard drives

<muurkha>tinybronca[m]: presumably Netflix doesn't pay for their bandwidth

<muurkha>"enterprise NAS HDDs" are surely not in the running; Netflix is a video streaming service, not an OLTP RDBMS

<muurkha>SMR would be fine; once a video file is written, it doesn't need incremental updates, and it basically only gets read sequentially

<muurkha>but conceivably even SSDs would be a good deal for them in order to occupy less rack space and be able to serve more streams per drive

<oriansj>stikonas: your untar patch has been merged

<stikonas>thanks

IRC channel logs

2022-06-06.log