Notices by Christine Lemmer-Webber (cwebber@octodon.social), page 45

@Wolf480pl @nolan @david_ross Ha... well I'm not trying to compete with anyone. But I do think it would be good if we could push our society towards using machines for longer and making them more user-repairable.

holy shit this soundtrack https://www.youtube.com/watch?v=gSGk1bQ9rcU

@er1n Racket: https://github.com/cwebber/racket-linkeddata/blob/master/json-ld.rkt

It's a port from the Guile one

Down to 37/78 failures on the json-ld expander

well, that's better than 70/78 failures, which is where it was last week

bad news everyone: your purely functional and capability-safe program is running on top of a side effect'y stateful world

the correct way to solve this problem is to generate new single use self-contained purely functional universes and do message passing between them, and corrupt the fabric of time between the message passing substrate in such a way that timing and other such side channel attacks are not possible

@er1n because javascript (and php?) people were mad they couldn't do js.type and had to do js['@type']

In fact some people wanted to oppose json-ld altogether for this silly reason, so the alias was done, which is considered "best practice" these days for that very reason by many

IMO it would be better to use the @, I think it's a good signal that it's different than other attributes in that it's not a normal property

SocialCG call today in about 40 minute! Hope to see you there: https://www.w3.org/wiki/SocialCG/2018-02-14

TFW you do an example from the tutorial and it keeps !@$%'ing erroring and you keep trying to figure out what you did wrong

then you read one more line forward and it says "surprisingly this errors and here's why"

(Don't reply with this telling me to not use github then or why it's nonfree; I'd like to but life is complicated and thanks I already know)

How did I get subscribed to so many projects with so many notifications on github

I don't even *like* github

org-mode to dogfood, generate its org-mode manual from manual.org (including info manual output) https://lists.gnu.org/archive/html/emacs-orgmode/2018-01/msg00348.html

Great news!

@bob @drequivalent the fundamentals of needing software freedom haven't changed, but the consequences of having or not having it have become more general and severe.

@drequivalent Also my suspicion is that we're going to *need* a libre hardware version of these in terms of "computing is becoming part of you more and more physically, and user freedom issues are no longer nerd rights but human rights"

@bob I have this feeling that blockchain technologies will move in the following directions:

a) bitcoin will stick around but people will give up on it being a currency... instead it'll be an e-gold value store to back other currencies
b) blockchains are not unlike signed git logs. Why are we trying to put everything in one repo? Instead the multi-repo approach, along what DAT and SSB are doing, will become more popular (with maybe consensus of "latest values" in one big chain")

(1 of 2)

@djsundog @KitRedgrave Macaroons are super cool, and are a big influence on the work I'm doing on Linked Data Capabilities. They lack some of the power of composition (as does linked data capabilities) as language capabilities, but this might not be a problem. See here for more context: https://github.com/WebOfTrustInfo/rebooting-the-web-of-trust-fall2017/blob/master/final-documents/lds-ocap.md#object-capability-programming-languages