Show newer

Alright. Time for my first vacation off work (for two weeks) since I started working full time. 🤪

No extra obligations, just relaxing and not focusing on productivity or commitments.

Social media is NOT about numbers of likes and retweets, it’s about finding out weird stuff the CIA did in the 60’s

Page 37 of this PDF really spelled it out for me. :O

Each of the "stores" can be IPLD Prolly Trees and the connections can be via distributed compute pipelines. 😍

Even the query layer can be distributed across multiple workers in order to take advantage of locality and parallel processing.

dbgroup.cs.tsinghua.edu.cn/lig

Show thread

Who needs earthly pleasures when you have distributed database engineering. 😎

Just learned more about how HTAP databases work and can't stop thinking about creating pipelines for auto-generating p2p search indexes over raw p2p data collections. 🤓

Java programmers should be communists all they do is focus on class

Wrote up my blog post on databases on top of / Prolly Trees. Hoping to switch up how databases are approached with a focus on performance and the ability to combine datasets determenistically.

blog.mauve.moe/posts/peer-to-p

@technobaboo Yeah, maybe once you have some of the foundations stabilized you'll have more time to look into what sort of stuff can be built on top of them. :o

Luckily the state of VR/AR hardware still sucks (especially with open platforms) so there's plenty of time to iterate before people can actually start depending on this stuff. :P

@mauve A lot of the problems I've had to solve involve coordinating rich interactions between multiple programs that have a separation of concerns that is intuitive to people (to avoid permissions and therefore permission fatigue) while also allowing maximum functionality.

@technobaboo Yeah absolutely! I think your work on stardust is some of the most exciting in terms of VR environments. :o

I'm not sure if you've been attending the omi scripting group meetings, but that might be a good place to discuss standards with others so we can make objects more interoperable (assuming that's a shared value in the first place 🤪).

Might also find a way to get funding from Protocol Labs or something to work on loading interactive objects from IPFS via WASM or the such.

@deltatux @andrew regarding clients, if more clients used the activitypub api instead of the mastoson api we'd be able to reuse them accroaa the fedi way more easily :x

I sort of want a set of Unix-ish single purpose command line tools for ActivityPub with which to do things like:
`readnote --wait @foo@example.com | jq '.object.content' | striptag > foonote.txt`
Doing auth and the rest with cURL is a PITA but with a little automation fun could be had. Someone has to have already made this, right?

All I want for Christmas is a domesticated raccoon that wants to eat lasagna with me & go on quirky adventures. It would also be nice if the raccoon could do magic but I understand that is asking a lot and therefore, it is not required.

@andrew I think this means that the fediverse is in need of a "private groups" feature in general that's not just tied to instances and the instance moderators. E.g. a private group can span many instances and have cross-instance moderation for adding and removing members. 🤔

It would probably not be a mastodon feature but a new fedi-capable system.

@andrew but then only ottawans on your local instance can see it and ones that are on others (like myself) can't. Do I to choose your instances moderators and infra to engage in my phyaically local community? Multi account setups make this easier I suppose 🤔

I've got spatial computing on the mind.

Had some chats today about what interaction with virtual objects would even look like. So far it's a big question mark.

Another is the interfaces that should be exposed to 3d objects to be able to make them interactable but which keep them sandboxed from wreaking havoc on the rest of your space.

My tool for making a simple, searchable, themeable archive of your public tweets and threads is now live:

tinysubversions.com/twitter-ar

The tool runs entirely on your computer, in your browser. None of your data is uploaded anywhere in the process. The output is a zip file of a basic HTML website that you can upload to a web host if you choose.

The site also answers most of the common questions I get. Please read it first before asking me questions here!

Show thread

There are accounts from the first settlers to Turtle Island that the forests were so abundant with food, and that the #indigenous folks they encountered seemed to eat along their pathways while traveling easily. This wasn't an accident. The forests were cultivated food forests, a great technological advance that was then mostly decimated, chopped down, and made into monocrops of things like wheat and corn. #permaculture and #foodforests are Indigenous innovations. Let's bring them back!

@heapwolf Heck yeah. Sounds a bit like the Goblins thing @cwebber@octodon.social has been working on at Spritely. :o

Have you seen the Prolly Trees article yet for using content addressibility for building and searching B-Trees? 0fps.net/2020/12/19/peer-to-pe

I'm working on standardizing it over IPLD, but it might also be useful for SocketSupply apps down the line.

Show older
Mauvestodon

Escape ship from centralized social media run by Mauve.