Textual UIs with Orhun Parmaksız

This might be our most hyped show of all time because we've been teasing it for, like

This is certainly one that's most planned in advance. I I I I think this is, like, a this whole advance planning is, is new to us. And, of course, we're we're in our our, I would say, Europe friendly time, but but Arun, you're you're in you're in Turkey. Right? So it's like it's not it's I think it's 8PM there. Is that right? It's it's definitely not it's it's not exactly midday there.

No. It's, yeah. It's eight 8PM. I have, like, two hour difference, with Europe. So CAT, like, two hour ahead of me, I think. I'm not sure. I think. Yeah. But that's So it's

still, not I mean, our normal time is is what? 4AM your time.

Four 4AM. Yeah. Yeah. Exactly. It's better than 4AM.

That's friend friendlier. It's all I mean, these days that counts as downright welcoming.

We the oasis is half full. Right. So, Adam, I'm I am very excited for this. I'm I am very excited to talk to his Arun, thank you so much for joining us. We've got some folks from Oxide who've been building on Ratatouille. And so we're we're really excited to talk about this. Yeah. We we we brought

all we brought all your groupies. I hope that's okay. Like, we couldn't keep them away. A lot of Ratatouille excitement at Oxide.

That's sweet. We're we're big Ratatouille users for sure. We we really love Ratatouille. The but, Adam, actually, has a a way of background. I thought I might tell I think I'm gonna tell you something about me that you might not know.

Okay. Wow. Okay. An ox headed prince first.

Making history here, folks. Brian is actually telling me something that I haven't already heard six different times over spanning three decades. Yeah. No. The the

I'm gonna put down the crossword puzzle. That's great.

Put down the crossword puzzle for a second, if you wouldn't mind. So, and of course, you may also be thinking like, you know what? I'll be the judge of that. I'll be the judge of whether

you told me this or not.

So when, so of course, back in the day, I had an IBM PC XT, which is what I I grew up on. And some of this you do already know, by the way. So don't hit your buzzer yet. Okay? Just just totally. Like, I know we've heard this one before. No. No. Stop. Hold on.

No. No.

Oh, man. That thing is like that that is my rosebud, man. That thing I I had that thing all the way through college, and it I it it it was left in a garage in in in Providence, and I'm I I like no, no, it's it's it's not good. Like, if the I I I did that monitor dirty, and whenever and that amber is like got such a kind of an emotional resonance with me because I spent so much time. I had like the word perfect screen was burned in on that on that monitor.

Nice.

And I wanted to have a way of having a a user pick from a menu. God only knows why I wanted to do this. Like, I wasn't making software for anybody else. I don't know. This is just, like, the the things kids do in their nerdy kids do in their bedroom, I guess. The, but I wanted to have a way of having, like, from a I mean, to call a batch file a shell script is far too generous. It really is. It, have you do you have any spent time with DOS, Adam? Is this, like

very little. I think only when I was dealing with, like, Renegade BBS stuff, like, just to also turn to the Wayback Machine. Oh, I I guess I guess not quite as far back, though.

Yeah. I mean, it it it's, you know, raw but it does bad. Batch files bad.

Percent error level percent.

Percent error level percent. Did you scream error level, Josh? I we need to scream percent error level. I I think I don't think that I think that had to be an upper case.

I think it yeah. I think that were case sensitive.

So I had this idea, I I and I guess maybe the the the apple didn't fall far from the tree in terms of my father's delusions of starting a software company. My own delusions, I'm gonna, like, I am gonna write this piece of software that allows you to have this this menu from a batch file, and, like, this is gonna be, like, my ticket to the bigs. Like, this is gonna be I'm really going places. So and I remember vividly thinking when I was in, like, eighth grade, like, if I can just figure out how to do, like, how how to read and write to a file, like, I'm cooking with gas. I'm, like, I I'm this is this is and I, so I did this thing and built this this TUI, really, that would allow you to, like, specify from a batch file, specify a bunch of things, and, made it available as as shareware, which, of course, in 1987 meant that I was uploading it to the Montana Programmers BBS.

I enjoyed this a lot, to be honest.

I I so that was the extent of my distribution. And then I actually, like, alright. Well, I gotta, like, you know, print. I said, I I I printed I I typed on our our IBM typewriters, literally circa 1948, just typed up some labels and was, and I was describing my ambitions to a friend of mine. And a friend of his is like, that sounds really useful. He's like, how I I I'll take that. I'm like, how much is it? I'm like, I think it's $5. And he gave me $5.

Wait a minute. You I I didn't know this at all. So you had a your first profitable software business back in 1985 or whatever?

Yes. 1980 yeah. Seven. Whatever. Nineteen eighty yeah. Nineteen yeah. Eighty seven, '80 '8, something like that. Peter Rubinstein, God bless you wherever you are, for your your charity. Oh, clearly, money meant nothing to you because you gave me $5. An early angel investor.

that'd be that would be almost $15 in today's money. Right?

And that'll be $15 exactly. But, of course, like, $15, then you gotta correct for the fact that I have no job and no way of earning money other than, like, doing food related stunts. So you you you gotta adjust for that. And it's kinda like, it's a little like, you know, like a thousand bucks. You could be like, wow, I can really I can buy, like, two whoppers with this.

Yeah. My first computer was a Mac in 1986.

Yeah. Wow. I mean, did you have the GI Joe aircraft carrier too? Are you one of those kids? Stop.

Stop. You, you, you careful what you say. I was not. I know those kids and I was not

one of those kids. Doctor. No, I was that. Listen, that was what I said was rude and insensitive. And I didn't like to give you that. I, I know based on your ability to socialize with the people that you were not in fact, one of the aircraft carriers. Yeah. You're gonna

ask you if I was homeschooled and forgiven. That's right. With forgiveness to all the homeschoolers who are listening.

But, like, yeah, go play with your aircraft carrier. You do you know you wanna you know you wanna do it anyway. Or invite us over so you can beat us at the video game that you've got. Like, you did these these guys all the time. But so okay.

Yeah. Yeah. So, I mean, like, in in in college, in, like, the graphics courses and stuff. Yeah. I mean, and and I have all the x manuals propping up on the course that I've left around. Yeah.

Yeah. Yeah. Yeah. So we did right. So the the I don't know. I'm not sure if you know X Windows, but, like, the the the hallmark of X Windows is the the end the kind of like the

the The, like, the pride flag of pride flag of books with, like, all the different colors.

Also, I was I was gonna go the Pokemon route, but I think you're also right. It's kinda like the you'll collect them all. You have this kind of, like, you know, whatever it is, 12 books, nine books that are the x window set. And you kinda need that up on your lap as you were doing the stuff because it was just, it was not exactly clean. There's a lot.

It it is notable that the state of the art, as you might recall, the competitor who we were looking at, it looked like command line tools wrapped in pretags. So I do feel like there is sort of a a a circular notion to this and we'll we'll as we'll get to later on.

Yeah. For sure, we'll get to. Yeah. Absolutely. I think there's a circular notion to it. And and I'm like, I was trying to think like when did I first start seeing I mean, because we definitely had to have some TUI programs. I I don't know when the term TUI started to arise. I don't think I heard it until, you know, the last decade for sure. I don't think I heard it before that, but maybe it's been always been around.

Do do you remember we we used to call our thing, I I almost hesitate to say it, a buoy. I think I'm I I feel like

you called them. I think you were the only people that ever used that,

that acronym.

Josh Josh has been patiently waiting for years. It's like, I will stage the next time these guys say buoy, I'm staging intervention. I'm jumping on.

It was such an awful term. I could never get behind it, but it was for a brief moment in sort of the late aughts or whatever. It's it was the term of art. It was awful. And I'm so glad it's fallen out of the lexicon.

I think that I think Mike invented that one.

I don't think so. No. No. No. That that was that people. Yeah. That was that was a thing.

Oh, that was a thing. Okay. Well, and then the and then we did have some things that had to be text based. I mean, there were or something and using curses or end curses.

Right.

Which is terrible. Although, actually, do and Encurse is is also kind of you know, I know we over the course of this podcast, we're really just going through our collective resentment that we've accumulated over our lives. The the Battlecruise knock off, do you know this?

Oh, a Netras.

Netras. And do you know what Netras is? So the for context.

Yeah. For a lot of context, I would just say Battletress, the Tetris, inspired a game that you made that is that was

Okay. So are you saying that this is, like, Sam Altman complaining that that DeepSeek is, like, you're a pain at all Tris. You may wanna sit down. I may wanna, like, lose some of the indignation about being knocked off, pal.

It's like we patented Tryst. It's like, did you though? I'm just I

think there's there's there's there's

maybe even a movie about it. I don't know. But, yeah. But Battle Trist was a game you made with some other folks, and it was your obsession in college. And then I would also note, after we finished was it Solaris 10 or Solaris nine? After we finished Solaris 10, it ground the productivity Solaris nine. No.

No. No. No. No. Nope. Nope. Nope. Nope. Nope. It was definitely not after we finished Solaris 10. It was after we finished Solaris nine.

That's right. That's right.

ZFS hung in the balance. ZFS no.

I like,

this is literally, like, you can go through because there were two things that Battle Trust almost capsized. ZFS, because this is early, early, early days of ZFS.

Yeah.

And the and Brigid, my wife, was, in her PhD program. And the the two worst BattleTress players on the planet, this is a net a network Tetris game where you play Tetris against one another. You get the in addition to the regular pieces, you would get these dice. You'd get that into into to lines that would give you money. And then every time you got 20 lines between you, you'd go to a weapons bazaar to buy weapons to screw up the other person's game.

I'm just gonna say in the last thirty years, you've really tightened up the elevator pitch for battle trust.

Thanks. I'm pretty good on that one. Right. Buzzed me.

Bryan, have you played, like, the modern? Like, this sounds a lot like modern, competitive Tetris.

I this is where Adam would point out that I would be wrong to be claiming that. Well, then then they've knocked off Tetris too.

Yeah. I know. Right?

No. This and actually was inspired by, in many ways by Wesleyan Tetris, which is another Tetris train. It was not networked, but would screw up your game. And, anyway, it was a lot of fun. I would, so I I'm I'm just gonna you're you're being a complete gentleman, and I'm I'm just gonna say it. You're a better Battle Trust player than I am.

I was just waiting for that. Thank you. Alright. Thanks, everyone. That's it for today's episode.

We can actually, like, we can actually now end the podcast. Actually, there's to be no more episodes. But you and I had some absolute clash of the titans on I mean, you and I and the thing is when we go at it, it's like a thermonuclear exchange. It is like it it is like Soviets versus it is Warsaw Pact versus NATO full nuclear exchange. It just does not last very long.

Two kittens trying to kill each other.

Two kittens trying to kill one another when actually the kittens have other things to do, actually. And so, it almost like I I mean, I was getting a little bit worried. I'm like, we like, ZFS might not happen because of Battle Tris. Like, that's how early it was for ZFS. It is it was early, early, early for ZFS. Sorry, Orin. You're getting much more than your bargained for here. Well, I promised we're gonna get the right two. We are we're we're sailing home. I promise.

Yeah. I have used awful awful. I I have used it for, like, the most pedestrian tasks, and, yes, it's awful.

And, Orhan, have you used curses? Have you had the Bright, like, a c application in curses using curses?

I think I did. Yeah. But it was a long time ago, actually. So I don't remember much about it, Sally.

Well, that's good. It's it's it's bad. Just let me just take our word for it. It is very, very bad. And that was kind of the state of the art.

Yes. So my personal history with the terminal and command line actually begins in, I think, 2036 or some some sort of, you know, around that times. I was before that, I was kind of, like, you know, learning programming and, like, trying out different technologies and stuff like that. And then, I basically met with Linux and then, started using Arch Linux, by the way. And, you know, I was just from that point on, I was just, like, a terminal geek.

Well, I think it's interesting you kinda mentioned that, like, the nostalgia aspect of this because the thing that I find so fascinating and, like, why because we do not use TUI's at Oxide out of any sense of nostalgia. We I mean, although, I guess, the nostalgia doesn't hurt, but the, I mean, because you really need the there are so many, especially on the server side, where it's like all I actually have is a is a shell prompt. All I've got is the ability to exit. I got a terminal, but I don't have a browser. I don't have I don't have you I don't have a network connection.

Never. And

Yeah. So the yeah. Go ahead, Orin.

So, somebody in our community, they had this very old, terminal hardware. I think it's, like, for telecommunication or or some some sorts. It's like, like, some, it has some special protocol that it it uses and some kind of, like, you know, special things going on with it. It was, it it was, created in, like, 1980s or, around that time in in France. And this guy just decided to basically put rotatree on this hardware, and he basically, like, shared his whole adventure with it in our Discord.

Well, absolutely. So one, a couple of things. And I I I, if you are if you are French and of a certain age, if you are if if we've got any anyone who's a a Gen Xer and who is French, I mean, the Minitel was, extraordinary. And I, then you ever, you've never seen one of these things or used it?

No. No.

So I, I had a homestay in France, in, 1991. And the family I stayed with, like many French families, had a Minitel terminal, and it was insane. It because it was basically like a it was an it was an Internet. It was a proto Internet, and you could do all sorts of things on this thing. You could it it was like, wow.

Mhmm.

And, and we certainly see that with Hubris running on microcontrollers. Right? I mean, Hubris runs the smallest system that Hubris runs on, an M0 plus has, 64 k of ROM, eight k of of of SRAM. And this is where Cliff is like, yeah, and it should run-in 32 k. But, like, I've never run on a system that's that small in my life.

Yeah. I don't remember the clock rate, but the the rep repo that I just sent to the chat contains all of the information. This guy basically created a, like, interface between, the Minitel and, and Ratatree using ESP 32 and some other other stuff as well. It's like I mean, if there was a, like, a conference talk that that is about this, I would just listen to it. It's super interesting for me and I think this is kind of like, you know, we we need some more details about how this really happened because, you know, this guy showed up on this discord one one time and then next time there was a Minitel running Rataty.

Oh, that's awesome. This kinda tacks into I loved, Adam, where we were talking of to the Shreiber's folks about deep seek and this kind of, like, mysteries that happened in the Internet and kind of, like like, the the the hundred language coin anyway. You're just like, wow. This is this is amazing. That is great.

So, I I kind of got into open source by accident. I I was, like, creating fun projects and sharing them with people on the Internet. And then I discovered GitHub. I started uploading stuff to GitHub. And then I was like, okay. This is called open source. Then, I just wanna keep doing this. Right? And my my mindset was each time I create a new project, I wanted to try something new. So, I I learned about, for example, the Go program language and created some CLI tool.

That's great. Yeah.

And and at some point, I had one friend. He said you should learn Rust. Rust is pretty good. And it was around 02/2018.

I was gonna ask yeah. Yeah. Yeah. I've that that's interesting. I because I do think 2018 is a that's when I learned Rust as well. And I I really think that's a major inflection point for I think kind of Adam in your when's your block entry? Twenty fifteen? When is your block entry? Yeah. That is that just, like, just seems exceedingly painful. And by 2018, Rust is there there are a lot, it's it's I just think it's much more approachable in 2018.

Yeah. Yeah. And, yeah, this this friend of mine, he he was, already, like, creating, small Rust tools and so on. And he was quite hyped about it like everybody else. So I was like, okay, let me just give it a try.

That's amazing. I was gonna say, like, oh my god. Spotify Tooie, where have you been my entire life? I've I've never heard of Spotify Toohey. It looks Yeah. Great. It is is there a more active work than the one I just Yeah. Just stupid? It's like Yeah. I'm like, oh, last commit four years ago. It's like, well, maybe the Spotify API hasn't changed much. And then you go to issues, it's like, doesn't work, doesn't work, doesn't install, doesn't work, doesn't work, doesn't work. You're like, okay.

Yeah. There is this, more active one, Spotify player.

I Okay. Good.

Package this for, Arch, and they seem pretty active, so it should be good.

I, I would love a terminal based Spotify player. This this would be life changing for me. So I'm sorry not to turn this podcast into this particular application, but that would be really, really life changing for me. Yeah. So it was that using what was then two ERS? When does it where does kinda two ERS come in here?

So two ERS is, is a library maintained by it was, like, maintained by a single person, and, Florian, his name is. And he was, like, kind of, like putting a lot of effort into it, creating releases and so on. And I kind of like tried to contribute to the repo a bit back then too, and it was like, I don't know, created in, I don't remember when it was created actually. Hold on. I think it was like a couple of years before my first entry to Rust but I'm not sure.

Oh, it was twiz. It was actually unmaintained. Okay. Yeah.

Yeah. It was, like, maintained, but, like, the the author was, like, I'm I'm busy, guys. I don't have time. Maybe we should do something about it, like we can create a community around this maybe. Yeah.

Great. I I I agree with you about the props to the name. So I okay. So when was this? Was this in with when was this? Because I started using it when it was two it was two ERS when I when I started using it, in, and I think Andrew and Rain would well, I guess, Andrew, you would be the one to correct me. I think Humility dashboard was our first use of two ERS. I don't I don't know if that, maybe not.

Two yes. The dashboard was definitely before Wicket.

So, so, Orin, I I was using two ERS as part of Humility is our is our debugger for Hubris. And Hubris being an operating system, not merely. By the way, if you wanna name an an operating system you want to be searchable on Hacker News, definitely don't name it Hubris because Hubris is used in, like, Hubris is used a lot on Hacker News in its actual

Just call it Dunning Kruger OS. That'll be easier. Call

it Dunning Kruger OS. Exactly. Call it flagged OS. Yeah. No. Absolutely. The, so the Humility is our debugger for that. And, we, after we had done bring up of our initial gimmick board, it was like, boy, we really wanna be able to kinda visualize all of these sensors in the system. And, it it was work that I did kinda over the holidays at the end of twenty twenty one into early twenty twenty two, based on two ERS. Loved it.

But you like, you know, you never wanna make

that complaint. You know, Adam, like, you don't wanna be like, the name sucks. Just like

You're right. Your kid's gorgeous, but

tough name. But tough name, Adolf, really. The I mean, you you just don't wanna make that complaint. And so Orhan, I thought it was great. Obviously, I was not the only one that felt that way. And I think Ratatouille is such a good name.

Yeah. Orin, I really appreciate the description like this like, understanding this because the way we stumbled onto this was, like, updating a dependency. Like, I remember updating some dependencies, seeing that two ERS was, like, pushing us to ratatouille and just, like, following along blithely, not understanding, like, this this really complicated community building that you were doing underneath it where it sounds like the or that original author, Florian, was not unsupportive, just also absent. Yes. That there was some urgency to, like, get everyone rallied behind one thing rather than fragmentation.

Yeah. So I think it it was, like, two year like, two years from now on like, two years two years ago, from now on, that we kind of, like, started this. And, we kind of, like, tried really hard to reach out to the, floor like, the the author, Florian. And, sadly, we we we needed to move really quickly. So, we just, you know, needed to do something about it.

And he did that, like, without any coordination? It was just like, alright. I give up.

No. Yeah. He didn't give give give us any signal or anything. And, yeah, that day, I basically created the the Rust security advisory for, like, renaming, like, you know, changing two ERS references to router three and, you know, two ERS is deprecated and so on. So that happened very randomly.

Let's see if he's to this one. I've got I need to make the check out too. Oh, hey. You are there. Oh, okay.

Yeah. So, he kind of, like, rejected very kindly. He said that, guys, I don't need this. You kind of, like, did the best thing ever by just, you know, keeping this project alive. Thank you. I don't, need any donations or anything. So, yeah, he's he's pretty happy with the current state of the project, I would say.

That's great. That's a real one. Yeah. It's great to hear, like, that he I mean, that that it was also so well supported. I don't know. That that he, you know, didn't wanna hand over the repo, but did wanna see it continue. It just sounds it's a great story just to to hear that continuity.

It is really great. I I noticed the whole, like, trying to talk about contributing back. You know, it was a very delicate process. And you all did a, like, a great job managing that, and I'm glad to hear that the original author is was happy with it. But, like, I also noticed it seemed like as soon as, like, Ratatouille became a thing and became the de facto standard that, like, development picked up a lot.

So, when we migrated to a new repo, we actually wanted to keep the the existing PRs in the original repo. And the only way to do that without, you know, getting access to the repo is basically, like, writing a script that will just create a PR from, like, each previous PR. So I I created a very ugly ugly script that will just, you know, go to each PR and also each issue and, like, take everything from like, take like, any comments and everything and just, you know, create, like, the exact duplicate of it into the new repo. So that way, we kind of, like, transferred all the PRs, and slowly, we started working on those PRs because there were, like, a lot of PRs. Like, there were around 50, I think, or, like yeah.

Well, really appreciate all of that and also really appreciate, honestly, the way you did it. Because I had to go from two e r s to round two e. Right? So I was like, a and I was like, oh god. This is how brutal is this gonna be? I'm like, oh, this actually isn't bad at all. It was funny.

As brutal as a said script. Right?

Yeah. And they it was it was actually new. It was good because it was a little bit more than that because they they did make some breaking changes. And but the breaking changes were all, like, good. I mean, they were all, like, pretty reasonable, and it it just didn't take too long. It was it was fine. It was it it, it was, as interesting, it was, like, well thought out. And, like, okay. No. Yeah.

Yeah. No. It was yeah. Tu e r s, I started with, for Wicket. But, like, so I the project started, like, originally so 2022, like, Oxcon '20 '20 '2, Adam, John, and I sat down to discuss, like, how we were gonna initialize the rack.

that is the management gateway service. So then and that is what allows us to kinda to to go to talk to all of our service processors. That's kind of that that gateway service that sits on, one of our SUDs that's attached to the switch.

It's both it sits on both scrimlets. So we have redundancy. So the scrimlets are her, yeah, the switch attached to gimlets, which are our server SUDs. And And so

And the gimlets are ours. Our compute sleds.

Yeah. We have kind of a interesting architecture there where we process well, I I I'm not gonna get into network architecture. I believe there's plenty of, podcasts on that and documents on that. Yeah.

We'll we'll we'll have something in the notes that we had a couple of episodes on this.

Yeah. So but, like, I was gonna do a CLI, and then Josh, built this manufacturing, UI in his own, like, Tuohy stuff because, as Josh said, like, other people's software is, you know, the greatest frustration. I'm paraphrasing there. But built his own thing, and I was like,

hey. Melville?

I don't know. But, yeah. I, like, I remember pinging Josh and being like, hey. Your thing looks awesome. Like, how do you how do you do this? And I was like, can I get your philosophy on building UIs? Because, like, I'd I'd done a little handwritten UI stuff using DirectX for a video game back in around 02/2003. But not much, lately. You know? I'm a systems guy.

in the beginning. Time Wicket comes up, but we man, we gotta get, like, a the a ASCII movie out there of the the rack the rack setup process because it is gorgeous. It is so Yeah.

But there was also one Friday where I think my wife went to sleep, and, I stayed up late, rendering the rack. So there is a full, like, screen rack rendering of the rack using, anti characters.

And it's It's so I mean, it's I

don't wanna give it away too much, Andrew, but was that that sleepless night when I guess your wife went to bed early. Is that when you made the, like, rack defender game Easter egg that's also built No.

No. No. So that was that was on the way to Oxcon twenty twenty three or '24. That was on the plane. I wrote that on the plane. And it has interestingly enough, it does have like, the the rack falling, racks falling from the sky do accelerate with gravity. So I tried to, you know, make that somewhat realistic. It makes for kind of weird gameplay. And as I believe you mentioned on another podcast, I am terrible at it. So I really need to merge that

into terrible at it. I mean, I I I I mean, like, look, I'm only the second best battle trust player, but, you know, I really expect when you even go and you drop them I mean, it is great. You drop the racks onto a truck. And I think are you driving the truck? Is that the No.

So the trucks random trucks, just go across the bottom of the screen, and you have to drop the racks on, and you get a certain other number of racks. And they have to land on the truck. And then, you know, if they make it to their destination, you get reimbursed a new rack. Like, you've sold a rack now. Right? You didn't destroy it in, manufacturing. And so That's right. Yeah. So now you get a new rack to sell. And it just goes on forever.

Oh, it was sent for Verger.

Oh, wow. Okay. Yeah. Yeah. So, like, this is why, like, it's been complained. I think Laura brought it up the other day. Like, didn't wasn't that demoed, like, two years ago? I guess. Yes. Yes. It was. So maybe I should do that some weekend. Maybe I'll

I mean, you know, frankly, like, I I got we got lots of people that have bought racks so they can play the game while they're installing it. And, like, what do I even buy this thing for? Like, this thing is the this is the boat anchor if I can't actually play this game while it's installing it.

Yeah. The install I mean, I built it because I think the installs took longer a while back. Now they go pretty quickly. They take, like, I don't know, fifteen to twenty minutes to, like, go through the whole process. But that's still fifteen, twenty minutes you could be playing a video game.

You could be playing a video game. Absolutely. Yeah. Yeah. But the and so this is happening over SSH, importantly. Like, this is why, like, why we I mean, any user interface is gonna have to be a TUI that's that is at this because we don't have the rack. The rack is not configured. The rack doesn't speak it has got no networking connection other than the than this connection coming out the front on the tech port. Correct.

Yeah. That's I mean, that's that, like, that's why it was gonna be a CLI, right, until I get inspired. But it's it is pretty cool to be able to just do this over this SSH connection. And SSH provides us other capabilities too. Like, RAIN RAIN is here, and RAIN RAIN, hooked in mechanisms to SSH to allow us to, either open the UI or to, like, upload, a tough repo, which is how we manage our software updates over that SSH connection.

No. No. No. I think, I think I think that was great. I think, you know, one of the things, is interesting where there is a a spectrum of, like, how, you know, how, like, high fidelity it is and, versus, like, how, like, in what places you can use it.

Yeah. That is that is awesome. And I the other thing, Rain, that I also love and or, this kinda gets to one of the principles that you outlined for yourself. You know, you in your Fosdum talk, you mentioned that, Rust is right once, run everywhere. And the only Windows software I've ever developed in my life has been software and Rust.

Yep. So, the the back ends, let me just, start off with the back ends of Resa three because, that's where it gets really powerful. So we we currently support, like, three three back ends. And two of them is, Linux and macOS native, and one of them is actually cross platform. And it's it's called a Crosstern.

just accepted talk driven development. This is the Yeah. Absolutely.

Yeah. Sometimes you just gotta push yourself a bit. So, that's that's what I did. It was quite stressful. But, I at the end, I had something to present, and it works quite well.

Okay. So can I ask, like, just a naive question? So the so when you're using the web as a back end, where where is the actual code running? Is that running as Wasm? You you've you've compiled it to Wasm, and then it's running it it's all self contained in the browser?

Yep. So I'm using, I'm using trunk, trunk. Trunkrs.dev, I think. And it's like bundling your Rust code and your, your files, basically, like the index HTML and other JavaScript files into one WebAssembly, WebAssembly file, and then it just runs. It's like it's very simple, actually.

Yeah. That is wild. That's really interesting. And then if you your website, though, is your website built with Ratzilla then? Mhmm.

No. Not yet.

Not yet. Okay. Got it. It's the terminalcollective.org is the Razzoa, like

Yeah. Exactly.

Yeah. That was really cool. There you go. That's really cool.

Yeah. The the live streams were quite fun, actually, because for, like, for, like, I don't know, two hours, you just watch me suffer. And at the end of the day, I just build something looking like that, and it's quite quite pleasing. But, I mean, sadly, first two hours is just pain because of, because of the rendering stuff. And I actually, kind of, like, render some some very cool looking animations, accidentally during those live streams.

That's really cool. And and so could you describe a little bit in terms of, like, when you because there are some other things I did. Some some things that were kind of approximating this, but you're like, this this is not the way I wanna go do it. Can you describe the philosophy that you have in terms of, like, the way programs should be able to interact with these back ends? Because I think it's kinda revealing as terms of the the philosophy and ratatouille.

So, what do you mean by interacting with the back end?

Well, it's just like I mean, you've got a real emphasis on keeping it simple for the programmer. Mhmm.

Yep. I mean,

I don't wanna put words in your mouth, but that's what it felt like from watching that. It's just like you you you want programmers to not have to know a lot about the innards of these back ends. You wanna be able to have a clean boundary. Is that

is that

up there?

In particular, I mean, you you showed an example in that in the defozedem talk, Orhan, where you're you were saying, like, okay, here's Ratzatouille. Now remember this, everyone. And then, like, twenty minutes later showing effectively the same code or very similar code for Ratzilla. And and that was very impressive. Like, it's very cool seeing, as Brian is alluding to, like, you know, abstracting away a lot of the details here and letting people just, like, do what they intend.

Yep. Yeah. Okay. Got got the question now. So as a as a as someone who comes from a back end and systems development background, I don't want people to learn about, like, front end stuff, and I don't want people to learn about, most of the stuff they don't need to learn.

That's really great. When I part of this, I also love I mean, you've just built a lot of things with Ratatouille. And it feel feels like you're still doing that, which I think is is I mean, I I that to me gives you great empathy for those building with Ratatouille because you're building with Ratatouille yourself.

Mhmm.

And in terms of, like, hitting these as you say, it's a balance. Like, you know, how much do you wanna you you wanna make it easy for people to build new apps, but you wanna also allow people to get into that back end where they need to for things like mouse capture, and I think it's allowing you to hit that balance. It must how do you, make sure I mean, how do you test this thing?

Do you mean or is it three or, like, yeah.

I mean, just kind of in general. I mean, it it just feels that, like, there it it feels like there's some testing challenges with this. I mean, you're you've got, so many and maybe this is where you're able to lean on things like cross term. But how I mean, because you but what you've got is is remarkably robust. What what have you done from a testing perspective?

So in order to test the Resideo application, you kind of have to design your app in a way that you can test it. For example, you you can expose a render function that will take your application state and give you back, something that that is testable or maybe take a back end. In that case, you can use our test back end, which is just, just a very simple back end implementation that that is that can be used to take the whole terminal as a string and can just assert on that string. Another approach is, I which I kind of like these days is, the snapshot testing. Let me just send a link to chat.

Intra brain, have we played around with that?

So we haven't actually done much testing for that. But I actually have been playing around with, snapshot testing for two weeks for a personal thing. So maybe we should hook it up into WCAD at some point.

Yeah. I did not know this even existed. This is cool. I this is very cool. Actually, it leads me to a question though.

Yep. So, when we took over from, two ARS, we are sadly bound to the existing rendering mechanism, which is immediate mode rendering. And, we we should just, you know, keep everything in place. We cannot just break everything, at that point. So there were actually quite, quite a bit of discussions about re redesigning the rendering model.

This is a project called RizUp.

I knew you were gonna love it. I knew you were

gonna lose your mind over this. Word to your mother.

Yeah. So Sorry. There are,

like, lot

of no no problem. There are, like, a lot of other attempts for, changing the, like, the rendering logic and, like, the the whole app, like, application structure for the tree. There's this another one, tree realm, for adopting the the Elm architecture. And, you know, like, they're very, very different projects. But, we want to, like, kind of adopt something based on, masonry and xylem approach and maybe have it as a as a part of fraternity at some point.

Crespo's eye is twitching somewhere. Have you ever talked to Crespo about Elm? No. But oh, so Crespo's got a great story about how he was doing a lot with Elm and had designed a bunch of things around a new feature of Elm or newest feature of Elm. And as he is going on stage to deliver to the talk, the the maintainer of Elm is announcing that he's ripping it out. The thing that he's like, well, oh, well, That may have actually It's

a shorter talk. A shorter talk, I guess.

Yeah. But, the two interrupts. The or this is this is so this is a lot of great stuff here. I mean, you're obviously looking around for a bunch of different kinds of models. And then so, Orhun, in terms of of Ratatouille, what's the, what are some of the things you're kinda looking forward towards?

So these days, we are trying to cook up a new release. It's been in the works for a couple of months now. And this is gonna be actually a big one because we finally modular modularizing the create into multiple parts. So, this is a good thing for, for widget developers because back then, if you wanted to develop a widget, you basically depend on which contains everything. This time, when we push this release, you will only depend on probably just Retoti core, and all the widgets will be in Retoti widgets.

I was gonna ask what what if yeah. What what what Yeah. Will people need to is there a breaking change from, the perspective of those using Ratatouille, or how does that work?

So, there are small breaking changes. Not many breaking changes. Like, there are some unstable feature flags and stuff like that. But, if you're you're just a regular user, you just bump your Ratatouille dependency, and will contain all the widgets and core and back ends and everything. So nothing will change on that side.

That's awesome.

I do have a quick follow-up question about, something that has fascinated me for a long time now is the from an aesthetic perspective, I guess. I feel like, applications in this ecosystem have a lot draw a lot of borders around stuff, which I I find fascinating, like, using box drawing characters. And it feels like there's a lot of, I mean, I suppose Brian and myself are probably the only people that still have a lot of, like, 80 column wide terminals. And and so I feel like my historic aesthetic has involved not using a lot of Chrome on around controls and

stuff, like, trying to

differentiate stuff with, you know, different case or use of bold or or use of background colors and stuff to to to make things visually separate rather than actually drawing lines between things. Do do you know is that is that like an intentional UI guideline thing that people are doing in the Ratatouille community, or or is that, like, a just like a a natural thing that has emerged now that the Unicode box drawing stuff is, you know, pretty well supported everywhere and, you know, you can do rounded corners and stuff?

Yeah. I think there's a so I I also, like, talked about this during my talk. There's something called look, I think, and I believe, this is like this is like the the most typical look that people go for. I think this is not, like, documented anywhere. It's just, like, something people just choose to do.

yeah, one one thing that

like a back a solid background color thing without any border? Or

So, the block is that, but you can add, like, borders to it. And Okay. The the way of constructing that is basically block two columns bordered. So, that's been used throughout our examples. And it's very easy to just create that block, bordered block, so people just use it.

Yeah. And you did something similar, right, for Ratatouille? Because I think that you demoed this at Fozdimm. Mhmm. But I think that, like, I I was unable to fully I was just, like, listening to it kind of on my phone and was unable to fully appreciate it, but I know that the audience was the audience was was oohing and aahing. So it was obviously, terrific on the big screen.

Yeah. This library is built on top of. It's called tuck in effects. And now the the author of this library is actually working on a script like initialization for the for the animations. So you so you will basically write little scripts for creating animations to your widgets and so on. It will be quite easy at that point. And, yeah, it's it's a quite use like, not not I don't wanna say quite useful, but, like, aesthetically pleasing at least. So, pretty cool projects are just, exist in

in the So cool. This is this is incredible. Wow. I'm looking at Exabind right now.

Yeah. I know. It it's and I don't know if, like, I'm actually, like, it's a little bit of a spin the wheel whether Josh I don't know if Josh, this is a step in the right direction or the wrong direction from your perspective. I it's really stunning, though. The I do love the way you're tying this to 80 columns.

Single. Single issue constituency. I absolutely single issue. Yeah. %. Yeah. Everybody's

Well, this

has been great. Arjun, thank you so much. I mean, this is you know, we are obviously really big users of Ratatouille here at Oxide. It's extremely important for us. Again, this is not like it is not a this is not out of nostalgia.

Yeah. Thank you a lot for the invite. It's been a pleasure. And, yeah, I I'm just, I'm just actually one of the people who work on these, thermal projects. So I just wanted to, like, thank everybody who maintains Rust three based tool or three tool or, Rust library for Rust three, and so on.

it's This is amazing.

This is incredible. I I we're we're gonna go find Adam. Can I go, like, can I go whack Adam in this thing? I

think I

I think I see you. Yeah. This is where this could be a bit show notes. This is incredible, and this is like we started I I I think I mentioned, Renegade and BBS's. Like, this is really it's great. BBS vibes.

I think it definitely is. I also feel like, Adam, this is gonna give you something good to do when I'm telling a story for the eighth time.

Oh, there you go. Okay. This is, like, really Even more helpful.

Exactly. This is awesome. And a a a great note, Orhan, to to end on. Thank you so much for, for dropping that in. I I would love to say that I'm going to a meeting, but I'm really just gonna go play this for a while. So,

And just like that, we missed the quarter. Exactly.

It would be the first time. But thank you very much again, Orhan. Really appreciate it. And thank you everyone for, thank you for, Raine, Andrew, and and Josh for joining us as well. And, thanks for joining us in in our Europe friendly time. And we'll do a couple more of these through the year. I think I I I enjoy mixing up a bit. So Yeah. Good stuff. And, getting back on brand, we've got absolutely no plans for next week. So we'll be kind of we spin the wheel. No idea what's gonna happen.

So

Stay tuned. Right. We'll see.

Stay tuned. Exactly. Alright. Thanks, everyone. See you next time.