 Computers keep changing the world, but their power and safety is limited by their rigid design. The T2TILE project works for bigger and safer computing using living systems principles. Follow our progress here on T Tuesday Updates. This is the 49th T Tuesday Update, let's get into it. So last week we learned the difference between M-Sleep Interruptible, which kind of isn't, schedule timeout, which is interruptible, and we got a hundred times improvement from six milli-air to 650 milli-air. This week my official task was aluminum, and I didn't really succeed at that, but I'll talk about it. In the next week, this coming week right now, starting tomorrow, is the SFI, Biological Computation Workshop. It's going to take up about three days and so forth, so that'll interfere with whatever other things are going to develop, so I'm not going to put anything else on the official hot plate for me. So, all right. Yeah, I said last week was the last call for stickers. This is really the last call from stickers. I guess the only thing I am putting myself on the plate for is to have the these stickers sent out before the next update. So if you'd like to get some, there's still a chance. You just got to get me a physical address somehow that I can send them to, and we'll get those sent out. That's that. All right. Yeah, so still haven't worked on true intertile, because that's going to take some design. I need a solid chunk of time to do it, so I've still been just working with Loopback, but I am getting further improvements. Is it possible to see this? See if you can see this. That is Dreg and Raz running on the Keymaster with full Loopback, and it's updating about once a second, and it's doing somewhat less than one air in the process, but that's really pretty cool. I mean, the Dreg are actually going out. It's very weird, because when they go out through this cable, they come back in through that cable. So, right now, since it's all full up, you can't really see, but when there's only a couple of music goes out of the bottom here and comes in over there, it's kind of weird. I also did a little bit more software hacking to speed things up a little bit, and I made an improved air estimation that when we get the stats guy up here, if we can bring him up, it just takes a second. If you can see it right down there, that number followed by an air, followed by an A, what is it? 0.67, so that's 670 milli-air that it's currently running, and that'll probably go up a little bit. But the main thing is, look at all these numbers. Look at all these giant arrows. It's going crazy sending all the cash updates. It's 5,000 bytes in 700 packets and so on and so forth, and that's just what business as usual is going to be once the engine is actually having events and sending bytes through the packet all day long. So there, you can see the air over the last period of time is now up to 790 milli-air. And in fact, I often see it in the 80s. And this is just a snapshot over a period of time, the total number of events that happened divided by the number of sites, divided by the amount of time, and we're often seeing in the low 80s, 800 milli-air. So that's okay, not too bad. And that is the software story this week. I really wasn't focusing on it so much. I really need to get myself to settle down and tear up the cash update protocol so that I can introduce new messages at the beginning and reorganize stuff and so on. And I'm just terrified because it was so hard getting all that stuff working the first time that I'm once burned, twice shy. We'll get to that after we get through SFI this week. All right, so that's that. And do we have anything here? It was just the picture of all of those packets coming and going in each direction. 4,600 bytes in 770 packets. The angle bracket, meaning that's priority traffic, which is what the MFM uses unlike the bulk rate, which is what the common data manager uses, can't even see that here. Although the common data manager is running, it's just all that stuff's getting pushed off. It's not nearly as much and it's getting pushed off the screen. All right, so that's that. Hardware. We have to get frames that we can take these individual tiles. And when we connect them together with the ITCs, that is an electrical connection that's not that strong, a mechanical connection. We cannot hang weight on it. We need to have some kind of frame that fit the whole thing on the back and hold the things together so that we can actually hang these things up on the wall. And the goal was to look into aluminum, how people would use aluminum, people in Nabah-Kirky that could do it, and so forth. The long and short of it is, I started looking into this stuff and, number one, I got terrified by the price. So, you know, an eighth inch thick, which is not even all that thick as far as rigidity goes, although I'm sure it would probably be fine, or three sixteenths of an inch thick. A power zone, four by four tiles, will fit in a two foot by two foot. It's something like 20 by 15 or something. So, that's the smallest size I could get if I wanted to actually do it. They're from 50 to 120 dollars each. And how many of them have to go through in order to figure out what's going on and how to get everything set up? So, that was terrifying, but I was saying, okay, well, even if I'm going to do that, how am I going to do the design? I'm going to do the design the same as I do in plastic on the 3D printer, presumably I'm going to use OpenScan, the program I use, and design up the holes and the cuts and whatever it is. And somehow, that could be turned into instructions for a CNC to machine one of these things and throw 80 percent of the bar of the sheet of stock or whatever it is. So, the idea was, well, maybe I should mock this up in plastic. And even if it's too big, the frame is too big to fit on my build plate, well, you know, you can slice it, you can cut it up, you can do pieces and glue it back together again. So, that's the road I started going down, saying what if I could do some kind of pieces that, you know, would be glued or whatever, somehow put together and that I could fit on the build plate that I could at least try in plastic. And what it really started driving home to me was I had been imagining a big rigid thing like the acrylic sheet, you know, which is just the same in all directions, but I don't really need strength in the same strength in all directions. It has to be relatively strong to take the weight of the power zone plus other power zones below it hanging from a wall or a pipe or something like that, but it doesn't actually need all that much strength laterally. And it needs to be resisting the to sag backwards because if we're going to grab it at the back, if we're going to grab one of these, if the frame is at the back of this thing, it's going to want to tip like that. So, it needs to be able to resist torque in this direction, but not in that direction. And so, well, maybe I should just start taking advantage of those particularities of the situation that focus on something that is stronger vertically than it is horizontally and focus on something that only resists torque in one direction. And what that got me to, starting to design stuff up in Inkscape, I was kind of imagining ball and socket joints between the power zones. So, you could kind of stick one of them together. And then as I started realizing that it only needed torque in one direction, instead of sort of a ball and socket, that way I imagined sort of a tab and a groove. And I came up, well, so there's a, that is nine power zones, which we will have. We could probably have 10. We could maybe even have 11, depending on how many are left and so forth. But nine or 10 power zones, I think we're going to have. That's the sort of size of the thing that we might look like. So, I started designing these things and it came up with this idea of, okay, what about a tongue and groove? The tongue and the groove would resist the rotation and then it would have a bunch of like ladders that would interlock for hanging strength. And so, here's what I came up with. Here's my little demo. If you can see it, it's got the grooves up here. It's got slots in the back and the idea, here's another copy of itself. And they mate with each other and they just click in like that. And they're really pretty darn strong. Now, this one I've worked enough that it actually comes apart pretty well and resnaps. Now, a lot of the designs I have are still pretty stiff. And a lot of times you have to fuss in order to get them to latch. But, and once you do, it's hard to get them to unlatch. But that all can be tweaked. So, that seemed kind of plausible. So, I took this thing and I started elaborating it out. There's a view from below. And now that has to go together with the feet that we had built these. We originally had these round feet, this kind of thing. But that was being printed vertically and these things tended to break. So, it wanted to be printed horizontally and we ended up with this kind of thing here. And so, we made one that has little daggers on the bottom that would stick through slots in the mounting plate. Originally, I was imagining sticking through slots in an aluminum plate, but well, we could try it in plastic. So, right. And so, I added these slots for one of these things to go in. I added sort of keyholes so that we could have some lateral stuff going through it and just started to build it out. So, here's the kind of the where I've gotten it to. It's got these keyholes line up with each other once the things are staggered because since the thing is brick wall, we have to get them to go in two different phases. So, it's kind of like this. So, here's two guys that are linked to each other. This is all just pretend on the slicer and the build plate. We stick in those, snap in the feet that are going to grab onto the tiles. We have rows of these things. So, the vertical direction is meant to be going this way. The horizontal direction doesn't need as much strength, but it still needs some. So, I had this idea of making keyholes in it and having little straps that would go in it. I don't really like the keyholes because they weaken the strength of the straps. And if it was only for holding them, it could be done with little clamps that would go around the outside. But the good thing about having actual keyholes like this is that if we wanted, we could actually hang them on screws on a wall in the keyholes as well, which would be harder to do if we had something that's more exotic that was designed for clamping with a plastic piece. So, we get the multiple pieces together and that's what it starts to look like. Alternating rows go up and down so that you have grabbing one pair of them, grabbing the next pair of them back and forth, and that's what it would look like hanging up, or the idea is. And so, I started printing them. They printed okay. No calamities with a 3D printer. Thank you very much. I started thinking that I actually wanted to print them diagonally so that, well, it's probably not worth taking the time to go through, but because my 3D printing knowledge is so weak, I ended up printing them diagonally. They print fine. They do what I want that way. So, here's three of them coming off the printer that way. All right. And so, the long and the short of it all is this. Welcome to the world's first complete power zone. And I have powered it up once and it sort of kind of worked with the same old problems of the system D races and so on, but really what this is about is about the mechanical connections. So, here's the back. Here are the vertical straps that are interacting with each other, latching on the horizontal that grabbed three, three, three, back and forth, three, three, three, back and forth, and so on. And the idea is that they would hang off of this thing and the next below it could hang off of those guys down there and so on. So, this guy's pretty heavy. We're going to say we're going to have ten of these. This is going to be a significant installation, but in a lot of ways this is really not bad. It's not as strong as I would like and when you take out a tile, it's kind of flimsy underneath because these things don't have as much crow cross support. See, there's an example. I just popped one of my little side key things out there. Maybe we better put this guy down before we destroy it any further. And in fact, when I was trying to put it together, I ran into an issue where I couldn't see one of the ITCs and sure enough, there it was. Can you see it? I bent one of the pins and, you know, if you don't notice that, if you don't notice it in back off and you just push harder, you bend the pin worse and can actually end up destroying the bracket. I was pretty careful. There it is again after I've pulled out all the ITCs because now the goal was to remove this guy out of the middle of the power and there I got it out. And it was possible. It worked, but it wasn't really great. It does want more lateral strength because it's really, it has a whole separate set of forces when you're trying to pull one guy out of the middle. So if this thing is going to work, it's going to need some kind of design for more rigidity so that the guy in particular, this guy below it, which was the edge of the power zone, he was just basically flapping because when he wanted to come along with the piece that I was trying to pull out because there wasn't enough to hold him down. So it's going to need more work there. But I think there's a good chance it'll end up working out anyway. We'll see. There's a good picture of it, how it looked. I put it under the microscope. I went at it with the tweezers. I got it pretty well straightened out well enough to take the the ITC correctly. And in fact, what you just saw was the fixed one in it. So I think it'll take some more work. And also, I mean, you know, like these guys that these straps that I've built, you know, they're just completely sharp 90 degree edges and they're all very painful to work with. And they're quite stiff. So I mean, this one, this one goes pretty nice. And it's, it's, it's way strong considering all. But so the corners need to be rounded out. And just in general, everything needs to be tuned up a little bit. But it goes together comes apart pretty nice. So that's the progress for this week. We've got a candidate, let's say the best one so far, for a power zone solution that will not only hold them together, give them support, allow you to remove things out of it, at least if you're careful, and also interlock with neighboring power zones. Exactly how that would happen if the power zones were completely assembled and you were trying to tab one into the next one, I'm not sure that's going to work out so well. So the assembly process still needs to be considered. But we shall see. And that's it. So I will stop here and hopefully have a good time at the biological computation workshop. We'll see if I put together a 10 minute talk to spread the word a little more. Hopefully something will come up and it'll be interesting. And that's it. Thanks for being here. Have a good week.