# I am the Watcher. I am your guide through this vast new twtiverse.
#
# Usage:
# https://watcher.sour.is/api/plain/users View list of users and latest twt date.
# https://watcher.sour.is/api/plain/twt View all twts.
# https://watcher.sour.is/api/plain/mentions?uri=:uri View all mentions for uri.
# https://watcher.sour.is/api/plain/conv/:hash View all twts for a conversation subject.
#
# Options:
# uri Filter to show a specific users twts.
# offset Start index for quey.
# limit Count of items to return (going back in time).
#
# twt range = 1 15
# self = https://watcher.sour.is/conv/g2frmqa
It's a beautiful fall afternoon and I have the day off. What are you all working on today? I've been working on a script that pulls in updates for a number of Git repositories at once in order to keep an updated local archive of them.
Today, I'm making it resilient against the maintainer force-pushing an empty branch in an attempt to foil archives. There's still some more work to do, but I just ran a successful test.
The complete history of the repository is backed up in the bundle before the evil maintainer's force push is brought in.
Output of my Git script when detecting a malicious forced push
This can be mitigated under normal circumstances by assigning branches to the dangling commits before they're removed by Git's garbage collection.
However, this sort of malicious forced push can still cause a lot of damage, some of which can be very difficult to repair. It's better for archival purposes to make a full backup and then pull in the updates. A human can sort it out from there.
@mckinley working on the Yarn mobile App.
thanks for asking!
@mckinley I'm working on how to ensure that multiple interactjng agents possibly running on different machines scattered over a network have consistent world state without requiring inordinate amounts of network communication
@eaplmx I keep seeing your name pop up in the RSS feed. Good work, man.
@abucci Interesting. Can you tell us more?
@mckinley Thanks dude! BTW, is there an RSS feed? Cool!
@mckinley I'm desperately trying to "un-brick" my new work Macbook and _actually_ get things installed / setup on it that I need to you know function 😅 -- Corporate IT policies often suck and tend to just "block too much" 🤦♂️
@mckinley I'm desperately trying to "un-brick" my new work Macbook and _actually_ get things installed / setup on it that I need to you know function 😅 -- Corporate IT policies often suck and tend to just "block too much" 🤦♂️
@prologic Sounds extremely frustrating. Is there any weird corporate spyware on there?
@mckinley For my organization I lead efforts to create a large-scale multi-sector simulation of human infrastructure, including things like the electric grid, the natural gas system, water and wastewater management, and several others, together with their interdependencies (most things depend on electricity and water, some things depend on natural gas, etc). We have a nice prototype at this point, and I'm working on extending it.
We've anticipated from the beginning that we would not be able to run all the software involved in such a thing on a single computer. This immediately creates a problem similar to the problems that online multi-user video games have: there's a world state that is relevant to all users, but not all users have immediate access to all aspects of the world state because state changes can be happening on machines remote from them. Keeping state consistent in a distributed software system is not a new problem of course, but this particular application brings with it some special requirements. For instance, since we are simulating aspects of the real world, there's a simulation of wall clock time (the simulated time from the perspective of the simulated world) that needs to be coordinated/synchronized among many agents in the system running on different machines that are potentially in different time zones with different real-world wall clock time. Naturally there are other types of state that must also be managed--for instance, if some component in the electric grid does down, and some agent depends on that component for its (simulated) electricity, it has to be notified of this change and react accordingly. There are two levels of notification/reaction. One is the "simulated physics", whereby powered entities stop receiving power when their power source is cut off or shuts down. Then there's the "simulated cognition" (for lack of a better term), where the agent becomes aware that it no longer has power and has to decide what to do about that (start using a simulated generator e.g.).
@mckinley The bottleneck in distributed simulations is almost always the network, so we pay special attention to minimizing network use and reducing the impact of network latency as much as possible. But I find these are "easy" problems compared to the state management problems I alluded to in the previous twt (they're not easy at all but they are less thorny let's say)
@abucci Fascinating stuff. What is this simulation primarily used for?
@mckinley To be perfectly honest, right now it's used as a proof-of-concept in demo-oriented meetings in an attempt to drum up support for further development. The vision, though, is to do a variety of real-world things with it. The real power of a simulation is that you can set up a wide variety of what-if scenarios, run them, and see what happens, even gathering statistics about the outcomes. You can't rewind and re-run the real world to see what would have happened if you'd made different choices, or if your luck had gone differently, but a simulation gives you (some very limited but still valuable) ability to do that. Armed with something like that, you can use it to plan investments to maximize resilience; test emergency response plans for effectiveness; conduct training exercises; or, even guide emergency response as an event is unfolding.
We focus in particular on what are called black sky events, namely events that lead to a wide-area crash of the electric grid. Natural events like geomagnetic storms can do this, but so can man-made events. Texas came very very close to this scale of crash in 2021. It doesn't seem to be that well known, but if a large-enough segment of the US power grid were to go offline, it could take weeks or months to bring it back; so called "black start" capabilities are limited, and have been degrading due to lack of attention and maintenance. So, a simulation can also bring out just how bad the impacts would be were such an event to occur and thereby spur action.