RFC 666 is just kinda boring. It's a follow-on to what appears to be an unsuccessful early attempt to define a common command language that would be shared by telnet-based protocols like POP3, FTP etc. I can't think of anything funny to say about it at all datatracker.ietf.org/doc/html/rfc666
Proposal: Do a scatter graph of X=RFC number, Y=date of RFC (I predict after you remove an early outlier period, there will be a gradual and smooth slowing-down over time of RFC release rate). Do a curve fit to predict on what date RFC 31337 will be released
Okay wow, the curve of which RFCs were released when is FASCINATING and is not at all what I expected. Look at that pretty-ass sigmoid looking thing
Any data scientists in the house want to extrapolate when RFC 31337 is coming
UPDATE: I RTed this tweet back in 2019, so apparently I am doomed to continually forget and re-discover RFC 69. (I have a similar problem with "the exocrine system".)
So here's a weird question: I'm writing a piece of software that sometimes identifies "addresses" as strings of dense unicode, like
Qme79uUuZfrBCdpVzstgwUBtcbekko71sdHkMfKsNFZjp9
অʭ൮ॿဖÐވѤІ၄ੳࡡਠګ૬Ȥഺශෆuऽωഷ1
爒㨗迓ꈀ镒涀葟磹袅❓珏✭䴠瑲ᯒ抴䃌ʟ
or
馗鑏𢕋䆶蛞揈𥣿𤪙孊ꋎ𡹸𦋔𠨓㶃褼丵
The base58 versions are unlikely to ever accidentally be anything other than random characters; English is high-entropy.
But the version that produces strings of Chinese characters. That worries me.
Is there a good way to scan a string to tell if it is "offensive" in Chinese?
That is, I'm worried that an identifier which could contain random adjacent chinese characters which a Chinese reader would read meaning into, and find it scatological. In testing I've already had it randomly generate a swastika once (inevitable, it shows up at 4 codepoints).
The structure of tech companies puts SO much power in the hands of single, individual men accountable to no one else that the death of one of those men, hypothetically speaking, has the potential to change history & culture in a way akin to, in centuries past, the death of a king
Except maybe more so because a lot of kings were fairly apathetic, hands-off rulers and only really engaged with government insofar as it meant raiding tax revenue to build palaces. Tech CEOs have a tendency toward micromanagement. (Although they also build palaces.)
I really, really hate using software that doesn't come with documentation
UPDATE: After a *lot* of fiddling I have managed to get to reduce the "unknown function code" from 0x935 to 0x7a. Is that better? It's going down. Is going down better?
What I'm trying to do is call this function. This is all the docs I have. Note the lack of a type on "key". In fact, the key MUST be a very special object created by an entirely other npm package, which can itself only be constructed with the output of a third npm package
So… this is odd.
Sublime Text did an update.
Now it is running like mud and pegging the CPU at 100%.
I eventually realized it was doing this "indexing" task.
Indexing… what?
I think it might be doing some parsing operation on multi-megabyte webpack-generated js files D: D: D:
Also the dock icon updated and now it follows the new Mac OS HI standards, which suck :/ They made every icon a rounded rectangle and now I can't distinguish them by shape anymore
It is SUCH an awful feeling trying to type on a MacBook which is overheating. The keys are made of lava
Hey Javascript people. Let's say I'm sitting here and going "I need to sign a binary blob then later verify that signature... I know, I'll use SubtleCrypto.sign". Do you feel a need to fling my laptop aside and yell "NO! You fool!! Use ____ from NPM!!" instead?
Also:
- How do I know which of the four crypto.subtle.sign algorithms to use?
- subtle.sign and subtle.encrypt use a totally different. Is there a way to use one key for both signing and encryption?
- I don't see nothin in the docs about like... IVs. Isn't choice of IV important
This is really hard maybe I'll just use TweetNacl :(