Untitled Linux Show 183 Transcript

Please be advised this transcript is AI-generated and may not be word for word. Time codes refer to the approximate times in the ad-supported version of the show.

00:00 - Jonathan Bennett (Host)
This week we're talking Intel in some Windows versus Linux showdowns, and then a new release of LibreOffice. Microsoft open sourced a thing, a new tool. This time Rob has a story about why we don't support bottles or maybe we do support bottles. You have to listen to find out what's up with that. And, oh yeah, it's time to shut down X. It's a lot of fun, Stay tuned. Podcasts you love.

00:26 - David Ruggles (Co-host)
From people you trust. This is Twit.

00:33 - Jonathan Bennett (Host)
Hey folks, this is the Untitled Linux Show, episode 183, recorded Saturday, december 21st. One degree of Rob. It is Saturday and you folks know what that means. It's time to get geeky about Linux and open source and hardware and software, all kinds of fun stuff. I'm your host, jonathan Bennett, and we have we have a panel, we have some guys we're going to talk about stuff. We got Rob and David and Jeff. Welcome guys to the Untitled Linux Show. Merry holidays, glad to be back. Show. Merry holidays, glad to be back. It is good to have each of you back. And we are actually going to start off with Jeff talking about some Intel news, this one not directly related to their CEO, but related to some hardware that just came out.

01:19 - Jeff Massie (Co-host)
Or their chief financial officer for those keeping up on the news. But you know, today I'm going to bring a couple of stories about Intel. You know, got to give them a little love somewhere. This first segment, though, we're going to focus on a comparison of the Core Ultra 9 285 AeroLite processor running on two different operating systems Windows 11 and Ubuntu 24.10. Windows 11 and Ubuntu 24.10. To be more precise, we're looking at Microsoft Windows 11 Professional 24H2, fully updated with all the latest Windows updates and drivers as of December 15th, and we're comparing this to Ubuntu 24.10, straight out of the box, running Linux kernel 6.11 and with all updates as applied by December 17th. But the story doesn't end there.

02:09
Michael Larable at Phronix also explored Ubuntu 24.10, with a few key modifications. First he upgraded the kernel to version 6.12. Then he tested with the 6.13 development kernel and finally he ran Ubuntu 24.10 with a 6.13 development kernel, but this time he switched from the default Intel P-State power save mode governor oh, that's a mouthful to the Intel P-State performance governor. In the article linked in the show notes, michael Arable explains his motivation for this comparison is he wanted to examine how both Windows and Linux are handling the P and E cores of this Intel processor. For those who are new to CPUs or perhaps haven't heard this terminology before, p cores stand for performance cores and E cores stand for efficiency cores, and they're literally how they sound P for higher power, more performance, and E handle the lower level easier tasks but more efficiently. Michael was particularly interested in seeing how Linux has matured in its handling of this hardware and, as some of you may recall and we've discussed this in the podcast in the past how initial scheduling of tasks between the P and E cores in Linux was a bit problematic. The initial scheduling of tasks between the P and E cores on Linux was a bit problematic. However, that was quite a while ago and the Linux scheduler has seen significant improvements since then.

03:31
Because the core Ultra9 285K processors designed as a productivity workhorse rather than a gaming CPU, the benchmarks Michael conducted don't include any games. Instead, he uses a variety of tests focusing on different workloads. These include database sets, blender renders, audio and video compression, decompression, transcoding, as well as common tasks like zip file creation and extraction, which, of course, involve compression and decompression. These tests are designed to push all the cores of these high end CPUs, focusing on multi-core utilization rather than single-core performance. So how did the comparison turn out? Well, I got some bad news For Windows. Congratulations, microsoft on coming in last place Out of 92 benchmarks.

04:21
Michael ran the stock Ubuntu 2410 installation that it was approximately 6% faster than Windows 11. Now here's where things get interesting, though, surprisingly, the newer kernels didn't have a significant impact on performance. And even more surprising was the fact that the Intel P-State performance governor was actually slower than simply running straight 6.13 kernel. Now, typically a performance governor provides a small performance boost, but that wasn't the case in the testing. I do want to point out that all versions of Ubuntu were actually very close and some of these fell into or probably did.

05:03
I didn't do the math, but probably fell into the margin of error. I mean, it was really close. The big gap was between Windows and Ubuntu. For the details on each individual benchmarking test, I encourage you to take a look at the article linked in the show notes so you can go through all the details and see all how each test performed and the based on your workloads, because, like I said, the 285 is not a gaming CPU. So if you're thinking, oh, I'm going to have cores and gaming and no, just no, it's strictly productivity. That's it, so take a look and love to hear feedback.

05:43 - Jonathan Bennett (Host)
Yeah interesting stuff, and as you, as you flip, this is the case every time interesting stuff, you know, as you flip, this is the case every time. We look at one of these. When you flip through the individual results, there are some very fascinating standouts, you know. You've got some where it's like the Windows actually wins on one or two tests, and then you've got one here where Windows is running at half speed, almost exactly half speed, compared to everything else. Um, there's a lot of them where ubuntu 2410, with nothing else, is faster than the new kernels and uh yeah, it's real fascinating. So, with this.

06:14 - Jeff Massie (Co-host)
Well, I was gonna say I should say this is with a geometric mean, that it was, uh, six percent faster, so that kind of takes out mathematically. Takes out some of the flyers.

06:23 - Rob Campbell (Co-host)
Yes so, with this new, uh breaking news that you have here, are you going to be switching, then, from Windows to Linux?

06:33 - Jeff Massie (Co-host)
Yeah, I plan to. I think about 2011,. 2010 is when I plan to switch.

06:41 - Rob Campbell (Co-host)
Sorry, I was talking to David which I was talking to.

06:46 - Jonathan Bennett (Host)
David. Oh guys, that's brutal, that's brutal.

06:54 - David Ruggles (Co-host)
Might be brutal, but it's fair.

06:57 - Jeff Massie (Co-host)
Yes, brutal but fair. There's a show title for you, yeah.

07:02 - Jonathan Bennett (Host)
All right. Well, david, since we're on it, let's talk about some software that you can run regardless of which OS you're on Windows or on Linux.

07:13 - David Ruggles (Co-host)
Absolutely. I have got an article about the fourth maintenance update to 24.8 of the LibreOffice software package and I picked this article in honor of Ken, since I am covering for him, as he is being the real Santa and helping all the Amazon packages get where they need to this Christmas season and he normally keeps us up to date on LibreOffice. I wanted to pull this story out specifically for him. This being a maintenance update, they addressed a total of 55 bugs. The details are linked in the two change logs, rc1 and RC2, but there are no new features. So it brings stability fixes and some regression fixes. It just improves the function of LibreOffice.

08:13
The deb and RPM packages are available for download. Source tarball, of course, is available if you want to compile your software from source and who doesn't want to do that? I mean you know you really got to know your nuts and bolts there and it is getting pushed out to all the repos for the various distros out there. 24.8 was released on August 21st of this year and there was a big privacy push in that release and they're just continuing to maintain it. So in their press release for this they do remind everyone that this is just the community edition and there are ecosystem partners that offer LibreOffice Enterprise family applications if you need support.

09:11 - Jonathan Bennett (Host)
Yeah, very cool. It's neat to see that coming along and LibreOffice continuing to be a great product. You run LibreOffice quite a bit, don't you? David?

09:21 - David Ruggles (Co-host)
Mm-hmm.

09:23 - Jonathan Bennett (Host)
On Windows.

09:34 - Rob Campbell (Co-host)
Yeah, don't you david on windows.

09:35 - David Ruggles (Co-host)
Yeah, actually, I, I'm a multi-desktop individual and he uses mac os also. Yeah, no, I am. I am anti-apple, everybody else I'm okay with. So it was actually a Asahi Linux story that came up this week. They released the spin based off of Fedora 41. And I looked at that and I was like, well, I hate Mac, I'm not going to touch that.

09:55 - Jeff Massie (Co-host)
So Microsoft is okay and Apple's not. That's the line. Okay, we see.

10:01 - Rob Campbell (Co-host)
I mean as far as companies go. I agree I hate Apple as the company. I don't have an issue these days with Microsoft, but I do have Apple products. I still hate the company.

10:14 - Jeff Massie (Co-host)
You were just going to forget the last 20 years. Okay.

10:18 - David Ruggles (Co-host)
Hey, microsoft. Okay, first off, I'm not a Microsoft apologist, but most of my development is on back-end server infrastructure. So you know run Linux VMs and Hyper-V and stuff and I wind up having to support a lot of Windows users. So being familiar with it is valid or not valid, but is helpful for some of my job stuff. But I also have two laptops sitting here that have Kubuntu on them, so I jump around.

10:52 - Rob Campbell (Co-host)
Microsoft is just another company full of people and the people. There's a lot of people that aren't the same today as they were 20 years ago.

11:02 - David Ruggles (Co-host)
Exactly.

11:03 - Jonathan Bennett (Host)
Hey, at least on Windows you can install applications even if they're not available in the Microsoft Store.

11:10 - Jeff Massie (Co-host)
Yeah, going back to the original story, the one thing I think I need to look at is because I'm really intrigued about how Python's coming into spreadsheet programs more and more, and I know LibreOffice Calc has had it for quite a while. But I need to take a look and see how hard that is, because for a while scripting in Python at Calc now this was probably a few years ago when I last looked at it was rather hard. I mean, it was kind of one of those you could do it, but there was multiple steps and hoops you had to jump through to do it.

11:47 - David Ruggles (Co-host)
And I'll have to revisit that, see if it, unless somebody knows if it's easier. Now to Interestingly, I don't do any Python in LibreOffice, but I create Excel files, calc files from Python using other packages, so I go the other way. I haven't done any Python inside, well.

12:04 - Jeff Massie (Co-host)
Excel now is supposed to have Python built in. Yes, I don't know which version or when they turned it on. You said Excel does Lou Lou.

12:15 - David Ruggles (Co-host)
Maresco. He used to host one of the Twitch shows, the Enterprise Twitch show, and now he's on main Twitch. Occasionally he's heading up the Python and Excel department and he's really excited about it. He mentions it every time he's on.

12:32 - Rob Campbell (Co-host)
Well, if those work across both those systems, Libre and Office that would be a very strong reason to go that direction, because a lot of the other things that one or the other has done weren't cross-compatible.

12:47 - Jeff Massie (Co-host)
Oh, and the real power is when you start doing a lot of data analysis, because then you have access to all those Python libraries. Python's pretty strong in data analysis as far as languages go, and I'm very involved with data analysis, so I I'm like, oh, this would be nice it could be quite a powerful choice all right.

13:18 - Jonathan Bennett (Host)
So, rob, you found, you found a fight, you found quite the dust up to talk about here. What, first off, what is bottles? And then what is the deal with bottles?

13:32 - Rob Campbell (Co-host)
so before I get going into the story, um, I want to thank sean for donating 10 coffees to the cause for getting me to try linux mint for two months. I haven't switched to it yet. I'm about the first of the year and I'll do it for January and February and I thought there was a second donation but last time I looked it wasn't there. I saw if I got withdrawn or if I was imagining it. But so I want to thank him because as I get ready to switch over to Linux Mint, this next story suggested maybe it's a good time to move on to something new. As I have been using OpenSUSE for quite a while, it's been good, but recently there has been some petty bickering between OpenSUSE package maintainers and the upstream developers of Bottles. I have my opinions, but I'll first tell you the story and let you guys see what you think. So Bottles is a GUI-based tool to help run Windows software on Linux through the use of Wine, and you know it's similar to like Lutris or Play on Linux other things that just make it easy, a good front end to make it easy for you. The developers have designed bottle software to work in a sandboxed environment like flat packs and say the official supported packages for bottles is the ones in the Arch, aur and FlatHub, and problems that end users then often overwhelm the bottles developers with these bug reports and issues that are often caused by how the downstream package manager packaged the bottles and distributed it. A letter back in 2022, titled A letter back in 2022, titled titled please don't unofficially ship bottles in distribution repositories Outline this and offered help to package managers saying our invitation is addressed to all those who are packaging bottles incorrectly and or do not provide adequate tests, thus invalidating the user experience. We are happy to help anyone who would like to keep their package, adapting it to our quality standards ie making the application work as intended. Ie making the application work as intended. So that was a couple years ago, over a couple years ago, that this open letter saying what's officially supported If you want to package it, you know, let us help you so it can be packaged correctly and work and not cause all these tickets to come upstream to us.

16:44
About two weeks ago, a patch was applied to Bottles. So I guess Bottles was kind of tired of all these downstream problems coming back to them and the patch is to exit the program, if not being ran in a sandboxed environment. So this essentially blocks downstream package managers from packaging the software without removing that bit of code first. Well, this led to a lot of arguing on the pull request, calling the software no longer free and open source, and the package manager packaging bottles for OpenSUSE. They took their own approach, they took this into their own hands. Well, they took their own approach. They took this into their own hands as they made their own modifications, removing the upstream patch that stops bottles from running. That's fine, stops it from running outside of a sandbox environment, and then also adding their own patch that removed the bottle's donation button with a patch file called don't support dot patch file.

17:51
Now, for some of my opinions, bottles is designed a certain way to work best with sandbox environments. That's how they created it. You know, and it's hard to improve a product when downstream providers package in a way that causes bugs and then, instead of supporting those bugs themselves that they kind of created, you know, they get pushed upstream to the developers. So then only supporting officially packaged setups makes perfect sense to me. Package maintainers they're taking advantage of this upstream code. It's like yeah, we're going to use your code and put it in our distribution for their own distribution and then removing things like the upstream donation button. You know they need to eat we often talk about needing money and you know being able to support themselves and pay their bills and you know they're removing that but they're not like taking on that support role for packaging.

18:58
You know it's all legal, sure, because it's open source, but it is awfully sketchy and you know there are distros like Fedora that I think took a much better path. They changed the code that kills a software when not ran in a sandbox and they changed it to a warning, a warning saying that. You know, I don't remember exactly what it is. But basically this is not an official package, be careful. But I believe I also saw that support links for Fedora are also directed to go to Fedora. So if you have an issue with the Fedora package, come to Fedora, come to our community and we'll help you out.

19:50
Obviously, if Fedora developers package maintainers like, yeah, this is an upstream issue, I'm sure it'll get set up. But you know, when I have a problem with software, I first contact my vendor and then, if they can't solve it, they contact their vendor. You know which eventually gets back to the developers. You know I don't go all the way, skip all the chains, so you know you should follow the chains of support. Other package maintainers on Reddit have said they follow the request of the devs and some have said you know, if you don't want to package it as requested or support what you package, maybe stop packaging it. It's already available on Flatpak or Flathub, which you can get on pretty much any Linux system out there. If that's the official way it's supported, there's not really a need to repackage it another way.

21:03 - Jonathan Bennett (Host)
I find it interesting. So in looking through this, I my my first thought was well, are there any, are there any distros that package this with the sandboxing? And I don't think so. I think what they're actually doing is they what this? What the patch does, is it checks whether it's running like as a flat pack, right?

21:23 - Rob Campbell (Co-host)
yeah, I think pretty much yeah, so it's they.

21:29 - Jonathan Bennett (Host)
They intentionally, they intentionally added a patch to make the thing difficult to package, right, and that's interesting um, but it's also comes after years of having problems with downstream package maintainers Right Having to sort through all the bug reports that aren't that shouldn't.

21:49 - Jeff Massie (Co-host)
They shouldn't be getting because of the downstream.

21:54 - Rob Campbell (Co-host)
Which then distracts them from actually making real improvements to, to the, to their officially supported channels and products. And that simple patch I for for developer maintainer. I'm sure it's rather easy to just comment that out from what I saw and move on if you really want to do it in an unsupported method.

22:16 - Jeff Massie (Co-host)
And I think Fedora did it right. You know what we're doing this. It's not fully supported, but we're taking it upon ourselves to handle the support of this, so we're not bothering the original developers and we're we're going to take the brunt because we're doing something that's not officially supported apparently the fedora patch has a problem, though it's got a little check box to not show this again, and that doesn't actually do anything, whoops.

22:51
Side note too this bottles is not CodeWeaver bottles, because when you first were talking about this, I thought this was CodeWeavers and I was looking at the article and I'm like oh okay, this is not CodeWeavers, because they also have bottles.

23:02 - Jonathan Bennett (Host)
Yeah, it's its own thing, it is its own application. Yeah, it's its own thing, it is its own application. Yeah, it's interesting. If you actually go to the pull request in the bottles GitHub, you've got somebody trying to argue that this makes it not open source software, which is not true. I understand the point that Tim77 is making, but he is incorrect and does not understand how open source licensing works.

23:25 - Rob Campbell (Co-host)
Yeah, I don't even think it does anything to the spirit of it, really.

23:29 - Jeff Massie (Co-host)
No, no, I mean at worst you could argue it's bad coding. It's still open source. Oh, it won't run. Well, that's just bad code. That's not, that's not. You can't see it.

23:41 - Jonathan Bennett (Host)
Yeah, it does bring about like an interesting question though, and I know what the answer is. But if you were to say well, I want to write a piece of software. I want to write a piece of software and I don't want Fedora to ever package it. I do not want this to be in the Fedora packages, like the text of the GPL, the text of any OSI approved license, really, so anything we would call open source you basically can't do that. You cannot exclude one of the distros from being able to package your software.

24:15
And so on the one hand you may just kind of think about that and go, oh well, it makes sense that you would want your software. Like if somebody wrote software and really didn't want it to be in Fedora, like shouldn't they be able to do that? And so on one hand I mean, yes, you could write a license like that, but it would not be open source. And the reason is because if you could say I don't want this in Fedora, you could buy the same sort of legal strategy, say I don't want anyone in the country of germany, england, whatever, to be able to use this software. And then I think it becomes a little bit more apparent why that's problematic, and you know that that's a. That's a road that just gets worse and messier the further we go down it. So there there is. There is a reason that, uh, open source software is available to be run by anyone the way that it is.

25:12 - Rob Campbell (Co-host)
This does make me wonder. I've tried bottles I don't know a couple of years ago and it didn't work very well for me at all. Well for me at all. And I don't know if, if I don't remember now, if the one I used was an unofficial package and maybe that is why I had a very bad experience with it. Um, so it really makes it, really makes me think I have to try it again with a, with a. It's an officially supported package.

25:44 - Jeff Massie (Co-host)
And you know, it's almost like a kind of a test too. If you have just a little patch in there and it stops it from running, it's like if you can't easily go in and remove that, if you're, you know, repackaging this for a distribution, you probably shouldn't be messing with it.

26:02 - Jonathan Bennett (Host)
I would think that level.

26:04 - Jeff Massie (Co-host)
Well, really, because somebody at that level, well, really, because somebody at that level, I would think, would just, I mean, this would be like oh yeah, oh done, you know. And I mean they're, they're not going to spend a minute on this.

26:14 - Rob Campbell (Co-host)
And yeah, we'll recompile, we'll do this, it's done recompile, do the thing, ship it out, don't worry about it again. It's like, yeah, if you can't at least do that, and yeah, you probably can't support it either.

26:28 - Jonathan Bennett (Host)
So there is one more thing that I will mention here, and that's because we have somebody in the chat room said doesn't it just mean that if the source code is available to the public it's open source and no, that's what's called source available and that is different from open source. And the reason that we have that distinction is you've got some companies out there I think Terraform was one of them. There's been a couple of big database pieces of software that have added clauses to their licensing that basically says you cannot sell this commercially and again, that is not an OSI-approved license, right? So you've got the open source group, osi, open Source Institute Initiative, probably initiative, anyway that's the group that sort of handles the open source licensing right now, and back when the term open source really became solidified what it means in the software world the people that worked on that they came up with some things that it's like it's got to have. You know, for a license to be considered open source, it's got to have this, and no restrictions on use was one of the things that it absolutely had to have.

27:41
And you know, in addition to that and sort of tied to it, it's like there cannot be a morality clause and be considered open source. So you can't come along and say you know, I don't want and here's the reason why is because nobody can agree on morality, right. So you can't come along and say I don't want anyone of this religious group to be able to use my software, or I don't want anyone in this business venture to be able to use my software, or I don't want anyone in this business venture to be able to use my software Because some people would consider the manufacture of alcohol to be morally problematic and other people think that that's a ridiculous argument. People do not agree on that and that's one of the reasons why, in the open source definition, you're not allowed to include that sort of morality statement in your license you're not allowed to include that sort of morality statement in your license.

28:27 - Rob Campbell (Co-host)
This summer, we talked about a similar issue with uh winamp open sourcing and they did open source. They said you can help us, it's, it's, it's. You can view it, you can help us, but you can't do anything with it yes, they made it source available.

28:41 - Jeff Massie (Co-host)
Yeah, I was going to say that. Very same example. That's the first thing I thought of.

28:46 - Jonathan Bennett (Host)
Yes, All right. Well, we have whipped that one to death.

28:51 - Rob Campbell (Co-host)
Well, I know David worked with a lot of closed store stuff. What's your opinion?

29:00 - David Ruggles (Co-host)
Oh, I was still trying to figure out. When you said OSI, I was thinking Open Systems Interconnection, which makes the OSI model of networking, and then that made me think about how we say DRM all the time, which is digital rights management or something to do with memory, and I was just like. There's too many acronyms, so I'm sorry. I prefer open source licensing and software and I apologize to everyone very humbly that I'm using Windows.

29:38 - Jonathan Bennett (Host)
All right, jeff, let's talk about something Arch is doing, and yeah, it's another Intel story, isn't?

29:46 - Jeff Massie (Co-host)
it. It is, yeah, we're, we're giving intel a little love today. Like I said, you know they're taking, they're taking quite the beating recently. But so this ties in with the first story I had and I was actually going to make it one segment in the show, but I kind of hit a couple pages and I thought, you know, this probably is too much for one segment, so I'm just going to make it into two stories. So now this is also about the Intel 285 processor, just like the last story, and but now we're going to shift our focus to comparing various Linux distributions to see which performance which performs best on the Intel 285 processor.

30:25
Now Michael Erible still included all the Ubuntu 24.10 variants that I mentioned earlier. You know, with the different kernel versions, the performance governor version, all that. It's still in there. But they've added some new contenders. So Fedora Workstation 41. We have an old favorite and you know that a lot of people know Clear Linux version 42790. You know, historically it's been a speed champion. We've got Arch Linux in there, because I know they've been at a lot of people been asking Michael Arable to include Arch in a lot of stuff and CacheOS. Now, most of you probably are familiar with many of these distributions. But CacheOS might be new to some, I know I wasn't really up on it, so I looked it up and it's basically a distribution specifically designed for speed and stability and it's based on Arch.

31:21
So the benchmarks that were used today are largely the same workload focus tests that we discussed in the first story. So there's no gaming benchmarks. This is all production style benchmarks and it's it might be almost the identical run, if not. There's a huge overlap. That Venn diagram really overlaps a lot and you know the results I think you find are quite surprising. The lowest performer was Fedora Workstation 41.

31:56
Next up, you know, a little bit faster, was stock Ubuntu 2410. Now those two were close enough that the difference could be within a margin of error. I mean it just, even if there's statistically a difference, realistically there's not, it's too close to even notice. But then came clear Linux, just slightly faster than out of the box Ubuntu, earlier Linux, just slightly faster than out of the box Ubuntu. And following, you know, still going up the ladder was Ubuntu 2410 with the 613 kernel and the performance governor. Faster still was Ubuntu 2410 with just the plain 613 kernel and the the stock governor. Arch Linux edged that out in that configuration by a small margin, and at the very top spot was CacheOS.

32:51
Now, if you've been looking at Foronix for a long time, if you've been listening to this show, clear Linux usually comes out on top, and often by a noticeable margin. You know, it's kind of the performance juggernaut and kind of has been the benchmark for a long time, so to see it in the middle of the pack was quite a surprise. Now, it's important to keep in mind, though, that the results were close enough that even the slowest to fastest day-to-day use unless you're specifically looking at timing benchmarks, you know you're, you're really not going to see a difference. It's, it's still just a few percentage points. So it's, it's not like holy cow, this is a lot faster, it's, it's, it's going to be just very minimal. But you know it's interesting that you know. But you know, it's interesting that you know Clear Linux, which is known for tuning and things like that of performance workloads, and it's not really meant to be a user-friendly distribution. I mean, you can run it, but it's geared towards more enterprise and tune for Intel processors. But you know, take a look at the second article in the, or take a look at the article in the show notes, the second article on Intel, for the full breakdown of benchmarks, cause again there's going to be certain ones that one one distribution runs away and then there's another one where it's way behind, and you know. But so this is again geographic mean or geometric mean, but you know, know to remove the flyers. But overall CacheOS was number one and on a personal note, when I was looking at it I see that it runs KDE in their desktop variant. So if I was going to jump from Kubuntu, that could be a real contender. You know, just totally go away from something. I know I'm, you know I ran Debian for a long time. I've been in the Fedora, I've ran some Fedora stuff over the years. This would be totally new. So maybe, you know, time will tell but maybe maybe this will be, uh, if I'm going to jump to something, maybe maybe cashy's my new os in the future there's also uh news out today.

35:05 - Rob Campbell (Co-host)
Um, that uh you haven't touched on about catchy os, is that? Uh, they've changed and optimized their kernel even further. It's now being built using a compiler, auto fdo, and so I don't think he's done any benchmarks on it yet.

35:24 - Jeff Massie (Co-host)
Now they've changed this, but it it may be even faster it could be, but now in in kind of the devil's devil's uh counterpoint here ubuntu on 2504 is also now supposed to compile with the 03 option and have some performance enhancements in it as well. So people are really starting to be a little more aggressive on the performance on Linux all over, I guess yeah interesting.

35:56 - Jonathan Bennett (Host)
One of the other things that came to my mind reading through this is that Torvalds, in the past month, has come out against the micro architecture versioning for x86-64, the whole v1, v2, v3, v4.

36:10
There's an article on Pharonix like December 4th or 5th, and he comes out and says this is really dumb and we are not introducing this into the kernel because and this is absolutely correct it's not as linear. This is taking CPU capabilities and trying to make them linear when they are not linear. And obviously we've seen that because Intel is now shipping new CPUs that don't have the AVX-512. So they're still stuck on x86-64v3. So I get why a distro like Cashi or Fedora or what have you would want to be able to do that and say well, we're just going to, you know, starting with this version, we're going to compile against this new, you know, micro architecture version. Unfortunately, that doesn't necessarily work in the real world when the two major manufacturers don't agree on which features and functionality is, uh, absolutely required for new designs yeah, it's, it's definitely a generalization, but on the flip side, the compilers that's how they're classifying some of this stuff.

37:12 - Jeff Massie (Co-host)
Yep, so I mean torvalds can say whatever he wants, but if you know, clan and gcd are going well. Here's how we're doing it, you know yeah yeah, I.

37:21 - Jonathan Bennett (Host)
I suppose it's fair, though, to say that, like the kernel does not necessarily need to have any of that, uh, it doesn't necessarily have to be aware of that. It can still use the old the. The old way of doing it, like it's gcc will sort of magically make that work, and the kernel doesn't have to care yeah, well, like five uh 512 instruction set.

37:42 - Jeff Massie (Co-host)
So even this, this brand new 285 processor, does not support 512 instructions natively. Now they get around it with software and whatnot, yeah, but it's not like amd. So, but that's, I believe, handled all in the compiler, not the kernel itself. Yeah, yeah, that's fair, but you're getting into my gray area here. So if somebody knows better, let me know.

38:07 - Jonathan Bennett (Host)
All right, well, let's talk about a Microsoft release. Is this open source software? It's a new Python tool. I think David's got the scoop.

38:17 - David Ruggles (Co-host)
Yeah, before we jump into that, though, I have just one quick question how many coffees is it going to take to get you on Casio OS?

38:27 - Rob Campbell (Co-host)
Well, I have to get through the Linux Mint challenge first and hopefully I have time before the Cosmic Pop OS, cosmic new Cosmic desktop is fully out. But you know, I guess another 10 coffees.

38:41 - Jeff Massie (Co-host)
We're charting rob's history, hopping for him like three or four hops away now and and I think rob needs to go into march a little bit because he I think he's trying to short chain somebody by doing it in february, so he's a few days shorter okay, I just had to ask we those.

38:59 - Rob Campbell (Co-host)
Well, I'll add those days on there.

39:02 - Jeff Massie (Co-host)
I'm always looking at the data.

39:06 - David Ruggles (Co-host)
So I saw several news articles pop up over the last week about a new Python tool that Microsoft has released on GitHub as open source. I linked to NeoWin because I like to spread the love around. Didn't pick them out specifically, but this is a news article by Pradeep Vishwanathan Microsoft releases a new Python tool for converting files and office documents to Markdown. So if you do any code development, most of the readmes and stuff even if you don't do any development yourself, but you're clicking through to the source code repositories on your open source projects most of the readmes and stuff that you see are written in Markdown and they have an MD extension.

40:00
So the focus of this release is all about leveraging large language models and AI, but I think that there may be some opportunities here for people that just need to convert stuff into Markdown not specifically for the AI side, but their focus is on converting PDFs, powerpoints, word documents, excel, images, audio, html and various other text-based formats into Markdown. This is a Python library and there's some actually connect it to Markdown and use it to Chet GPT-4.0. And then you convert an image and it's going to give you text describing that image, which image description is actually one of the things that LLM seem to be getting pretty good at, so that's pretty cool. It is open source and it is super fast. It's easy to use and it's something fun to play around with.

41:33 - Jonathan Bennett (Host)
I'm immediately struck by a need. I need to do some playing around with taking an image, asking LLM to give me a description and then feeding that description back into LLM and seeing what it gives me.

41:46 - David Ruggles (Co-host)
That could be fun.

41:47 - Jonathan Bennett (Host)
Yes, yes, that's kind of like you take a recipe and you translate it back and forth from Chinese and English several times through machine translation. You try to figure out what it is at the end of it. I feel like your images would be very similar.

42:02 - David Ruggles (Co-host)
Or that's one of those tricks from college where you don't want to get dinged for plagiarism, so you uh take content, you translate it several times and then back into english wow, we do not condone.

42:15 - Jonathan Bennett (Host)
Do not condone any such shenanigans thanks for the tip.

42:20 - Jeff Massie (Co-host)
I am considering some new classes next fall, so the the the previous discussion was for entertainment purposes only and not to be actually used for actual academic it.

42:32 - David Ruggles (Co-host)
it was something I had to be aware of when I was a uh part-time instructor at a local community college. There you go, there you go Now. When I went to college myself, it was well before the advent of easy translation services.

42:49 - Rob Campbell (Co-host)
Okay, but even if you are aware of it, what can you really do?

42:56 - Jonathan Bennett (Host)
There are some different places out there that are trying to offer like is this written by LLM detection technology so you could feasibly try to get the same thing? For is this translated by a machine? Yeah, but is?

43:09 - Rob Campbell (Co-host)
that a thing, yet.

43:11 - David Ruggles (Co-host)
But yeah, from an instructor's standpoint, you basically just kind of have to be aware if the product of work you're getting from a student suddenly seems more impressive than the student has been up to that point.

43:28 - Rob Campbell (Co-host)
That's always a thing, though, even before LLM yeah.

43:31 - David Ruggles (Co-host)
Oh yeah.

43:32 - Jonathan Bennett (Host)
Yeah, that's crazy. I do notice talking about the market down the library itself. It library itself. Um, it supports PowerPoint, it supports word, it supports Excel. No library office formats. It's a little disappointed, maybe. Maybe they'll take a pull request.

43:53 - Jeff Massie (Co-host)
Well, I thought all those Microsoft formats were supposed to be open source. That's what the bro brochure said.

44:01 - Jonathan Bennett (Host)
They're sort of open in a way. No, I know the reality of it all. Right, rob, you want to talk about the, the thing that we disappointed you by not talking about last week I think you probably disappointed ken uh more than me.

44:20 - Rob Campbell (Co-host)
So this is going to be another episode of something ken should buy. So if you're listening, ken, which I know you are, I've seen you in the discord. This one's for you because here is your next pc. So I can't remember if we talked about the the raspberry Pi 500 being released. I know we have fans many fans in our panel and users out there of the Raspberry Pi 400. The Pi 500 is much the same but improved, with the Raspberry Pi 5 architecture and eight gigabits of RAM, the position of the GPIO header and 3x USB-A ports swapped over and the location of the USB-C power ports moved. They're the same ones as before in the 400, just a different order. And all this is available for $90 from Canikid. It's actually there. I looked at it today. I really thought about clicking the buy button, but I didn't think I could get it to Canon time. So maybe next year. But this isn't the big reveal. We all knew the Pi 500 was coming and we knew it would be great. It's the Pi 400 with the Pi 5. But what nobody speculated on was the official Raspberry Pi monitor.

46:03
There are many portable monitors out there. I have one myself for, for traveling. But for pi fans like you and me, this one is it's it's designed for you, it's it's designed for pi fans. Does the specs come with? It comes with a 15.6 inch IPS LCD display with anti-glare, 1920 by 1080 at 60 hertz. Max brightness of 250 nits, 45% color gamut, 80 degree viewing angle, full size HDMI port gamut 80 degree viewing angle, full size HDMI port. Front facing stereo speakers has two 1x2 watts, so don't expect a whole lot from that. And courage it has the courage to add a 3.5 millimeter headphone jack, the courage that Apple has lost years ago and Samsung and all those makers. But Raspberry Pi Foundation keeps that courage going on. It's hilarious. It also has physical buttons for volume, brightness and power. It has mounting options, including a kickstand.

47:20
The portable monitor display I have does not have a kickstand. The the portable monitor display I have does not have a kickstand. It has this, um, just a cover that doesn't even stick to. It just kind of fits over the top and if you fold it upright you can kind of sit in there, but it doesn't hold it up that great. So kickstand, uh, would be a great option. Or it also has the standard VESA mount so you can mount it to anything an arm, a desktop mount or whatever. Vesa is the standard mount that is on almost all monitors monitors.

48:06
This monitor is powered by a 1.5 amp, 5 volt USB-C powered with dedicated. You can power with a dedicated power supply or through a USB-A to type C cable powered directly by a USB port on the Raspberry Pi or another device such as a monitor, with one limitation, or I guess two limitations that the brightness is limited to 60% and volume to 50% when the monitor is powered by a USB port. And all this just a little bit more than the pi 500 itself 100 us dollars. So you get the pi 500, you get the monitor 190 us dollars plus tax and whatever shipping other fees are in there. There you got your your full computer right there and perfect for ken.

48:58 - Jonathan Bennett (Host)
So I've seen I've seen a couple of complaints about the monitor. One, one complaint in particular about the monitor and that is that the ports are behind the VESA mount. So if you have anything connected to the VESA mount you really can't get to the ports on the monitor. Which is not that big of a deal, I don't think, because once you have it mounted up with your VESA mount do you really need to be getting to the ports to unplug and plug things in.

49:25 - David Ruggles (Co-host)
But this is a complaint is it saying that you can't get to them, or are they like blocked, where it can't be used?

49:32 - Rob Campbell (Co-host)
uh, I think you just can't get to them, I think you can still have things plugged in because because of the way it, enough of uh indent or whatever, so that way you could still get them in behind whatever base of plate you have there.

49:44 - Jonathan Bennett (Host)
Yeah, so your your ports on this thing you know you're looking at your monitor they actually plug in sideways. There's a, there's a well, so the cables run in and plug in sideways. You just you cannot get to it because that well is directly underneath the visa mount. So that is. That is one complaint I've seen people have, um I. I sort of have another complaint and that is that none of this stuff supports, uh, usbc alt mode, and when you talk about a portable monitor, that seems to me to be like the killer feature that you really would like to have. Um, so I would have loved to have seen the ability to just put the whole thing on USB-C and get, you know, HDMI over USB-C or or you know one of the other. I don't remember if USB-C alt mode is HDMI or if it's DVI.

50:35 - Rob Campbell (Co-host)
Anyway, oh, so you can't just have a single cable even?

50:39 - Jonathan Bennett (Host)
I don't think so. I've never. I've not seen anywhere in the specs that it supports it. Even if you could though I don't think through any of the Raspberry Pis support that I don't think there's a Raspberry Pi that has has USB-C.

50:49 - Rob Campbell (Co-host)
Because it said it's USB-A to USB-C. Yeah, oh, that that is. That's that is disappointing.

50:57 - Jonathan Bennett (Host)
As cool as it seems, now you can do it with two cables, so like it's still, it's still usable. Um, it's just not like a top of the line, um, portable monitor, right and and. So this is something else I would actually like to see on a future Raspberry Pi is a fully forked, fully featured USB-C port that can do power delivery out. That this this is something that I know is sort of niche, but I would love to see power delivery out from Raspberry Pi.

51:27 - Rob Campbell (Co-host)
Out and the display all in one cable.

51:30 - Jonathan Bennett (Host)
Well, yeah, so on a laptop there's USB-C ports. You can do that. It'll do USB-C power out and it'll do alt modes, that you can just have a single cable to run your display.

51:41 - Rob Campbell (Co-host)
Oh, you have the Pi having the out.

51:43 - Jonathan Bennett (Host)
I would love to see the Raspberry Pi 6 have a fully featured USB-C port on it.

51:48 - Rob Campbell (Co-host)
I thought you meant the display, so you could daisy chain them.

51:51 - David Ruggles (Co-host)
Would beefing up the components and board to support the power out increase cost significantly? Oh.

51:58 - Jonathan Bennett (Host)
I'm sure, yes, I'm sure Right, and so that's why they haven't done this. And that actually leads nicely into if we're talking about the Raspberry Pi 500 as well the big disappointment that everybody has with the Pi 500, and that is that it's got an NVMe port on the motherboard but none of the components populated to actually make the thing work. So the traces are there, but there's just no NVMe header and there's a couple of. Let's see. There's the header itself to actually physically be able to plug something in. There are two little tiny service mount capacitors and then a 3.3 volt power supply. That is what is missing to be able to get an NVMe port on the Pi 500. And the fact that the case so the actual plastic case, the bottom of it there's no flap there to be able to get to it right.

52:55
There's a couple of other really interesting things. Jeff Geerling was the one that reallyred the 500 apart and looked at this. There's also a place that seems to be for power over Ethernet, which is also very, very interesting. The Raspberry Pi Foundation was thinking about Pi 500 with power over Ethernet, and so these are two. You know, like for certain people, for me, for some of us, these are really, really cool features that we didn't get, that we would love to have, and I'm pretty sure what we're seeing here is. It's just there's going to be a Pi 500 Pro, it's almost inevitably coming, and the reason they didn't just make these things standard is it would increase the price, so they wouldn't be able to sell it for $90. They'd have to sell it for, I don't know, 105, 110, something like that, when you talk about adding all these different things to it.

53:47 - Jeff Massie (Co-host)
Yeah, the circuit board to increase the power wouldn't be that much. I mean you'd have to beef your traces a little but cost-wise it wouldn't be terrible. It'd be more of the components you'd have to add and any licensing that goes with that Right?

54:02 - Jonathan Bennett (Host)
So with the 500, with the Pi 500, the actual keyboard, the traces are there. It is on the board itself. They literally have the traces there. In fact, there's a Hackaday article we cover this where someone sourced the components and made the NVMe slot work. Huh, it's just, there's like four different components. You have to add to it at least four, maybe even more than that with some resistors on there.

54:24
But again, it's just, it's going to. It's going to increase the bill of materials, which is going to increase the cost of it, and they were hoping to keep probably a sub hundred dollar price point is what they wanted to hit. And you start adding a bunch of things to it and it's going to be over that. So all of that to say the Pi 500 itself. Yes, on the CPU side it's an upgrade, but it doesn't have enough of the killer features for me. I'm not ordering a Pi 500. I'm holding out for the Pi 500 Pro, Pi 550. I don't know what they're going to call it Pi 500 XL. That's the one that I'm waiting for. The XL will include base amounts.

55:02 - David Ruggles (Co-host)
I hope so. I'm waiting for the XL will include VESA mounts.

55:04 - Jonathan Bennett (Host)
I hope so. I legitimately hope so. I asked Evan, I interviewed him and I asked him. I said we need the VESA mounts on the Pi 500. I was disappointed to not see it. I'm hoping it's coming on the Pro model.

55:18
Where you'd mount a monitor directly to the no, it doesn't necessarily have to be VESA, but I've just thought for the longest time that it would be nice to have on the bottom of the Pi 500 keyboard so you know, your Pi 500, it's a keyboard, you mount it to things. It would be nice to have a couple of certs in there, whether it be the VESA mounting pattern or not just a couple of places in there, to be able to actually bolt something to it, to be able to mount it.

55:48 - Rob Campbell (Co-host)
You could put it on a swing arm then, just like the monitor, exactly.

55:52 - Jonathan Bennett (Host)
You know how cool that would be To be able to have a Pi 500 on a swing arm connected to a Raspberry Pi monitor, also on a swing arm. That would be cool.

56:04 - Jeff Massie (Co-host)
That would be cool.

56:06 - Rob Campbell (Co-host)
I could see some use cases there.

56:07 - Jeff Massie (Co-host)
Mm-hmm, first, you know. You know, I was just kind of looking too. You know, when you're adding components to a Pi, you know there's some NUCs that get kind of cheap in there too, and at some point you're going to, you know, crossover if you, depending on how much you throw into it and you kind of then go on, well, I can have a low end PC for what? Yep yeah.

56:34 - David Ruggles (Co-host)
And that is a NUC is it comes with a windows license.

56:38 - Jonathan Bennett (Host)
I don't the one I was looking at. I don't know, as it did. It may or may not be a legitimate Windows license, but it may have Windows installed on it, sure, yeah, I don't know that I would trust that install of Windows, but it's probably on there, yeah. Well, the other thing you have to think about too, is that Raspberry Pi Foundation. They're very much still thinking about the educational side of things, still thinking about the educational side of things, and so they are. They're. One of their main focuses is they're making these computers for education and so trying to keep the costs down, for that is a big deal for them too. So, like I don't, I it does not bother me that they made the pie 500 and that they made it with the reduced feature set. The thing that annoys me the most is that there wasn't also an announcement that something more was coming.

57:25 - Jeff Massie (Co-host)
Well, think of it this way too. Here's, here's a heck of a uh a technician, electrical technician, uh, trainer. Okay, in today's class we're going to get the parts and we're going to surface mount uh parts onto onto this and we're gonna make this thing, the nvme, work, and we're gonna make the poe work, and those coupling capacitors are so tiny, they are very small there's your educational focus you're looking for that's true, yeah, and there's.

57:57
There's people that I've got buddies that they solder under microscopes. I mean that's kind of what you have to do now. The old days, the big old spool.

58:10 - Rob Campbell (Co-host)
And that's pretty much it. If you got the skills for that, you might as well just go be a surgeon and make lots of money.

58:16 - Jeff Massie (Co-host)
Yeah, but if you pop a capacitor you're not.

58:20 - David Ruggles (Co-host)
Yeah, your downside's a lot higher yeah, I.

58:24 - Jonathan Bennett (Host)
I suspect that the guys that can do that with the soldering irons actually make more than the surgeons do in some cases I don't know about that, but they're they're making good money.

58:34
I will say that believe, believe it or not, in a lot of cases, points are totally off topic. In a lot of cases, though, most of that money you know you spend a hundred thousand dollars for a though most of that money you spend $100,000 for a surgery Most of that goes towards the hospital and not the actual surgeon. Anyway, let's move along to something more technical related Before we move along.

58:53 - Jeff Massie (Co-host)
I have a little bit of hardware. It wasn't enough to make a story, kind of a side note. I would love to hear if anybody has an experience on it. But I'm kind of looking at the uh Flint to wifi six router. It's from GL period I net, I guess, is how you say it and it basically runs open WRT it. Uh, they have a kind of a customized version, but I was doing some research and you can go between the generic OpenWRT and their little customized version easily back and forth. It's pretty close to stock OpenWRT and it's just a new Wi-Fi router that's basically built to run.

59:40 - Jonathan Bennett (Host)
You might also consider the OpenWRT One one. Yeah, if you can get it now, I don't think I don't think amazon has it locally yet, so it's still aliexpress, but uh, those are. Those are a couple of the really uh, couple of the really impressive ones right now.

59:55 - Jeff Massie (Co-host)
I've used glinet before and they make good products as well yeah, so I, and it's pretty, seems pretty open source friendly if they're using open wrt. So they're. You know, yes, some nice, nice hardware and and it comes with the wi-fi where the, the, the one doesn't really.

01:00:14 - Jonathan Bennett (Host)
I mean, it's not the you can get it with wi-fi. The big downside of the openwrt one is that it only has a single 2.5 gig port, whereas the Flint, I think, has at least two of them.

01:00:26 - Jeff Massie (Co-host)
I think it's 210 gigs and it's got several networking ports on it. There's two 2.5 gigs and there's four 1 gigs and it supports Wi-Fi 6. Yeah, so it's kind of the step up from 1 gigs. Yeah, and it supports Wi-Fi 6. Yeah, that is impressive. So it's kind of the step up from the 1. Yeah.

01:00:47 - Rob Campbell (Co-host)
It's cool. The price isn't bad for all that either.

01:00:51 - Jeff Massie (Co-host)
Yeah, no, it's really good. Yeah, so there's just more hardware out there that we might be purchasing.

01:00:59 - Jonathan Bennett (Host)
So, Jeff, you jumped in to tell us about hardware right before it was time to hand it back over to you. It's tell us about hardware.

01:01:08 - Jeff Massie (Co-host)
I'm seeing a theme, yeah yeah. So laptops, you know, and I bring this up every once in a while, I kind of keep watching them evolve and kind of exciting. I kind of keep watching them evolve and it's kind of exciting. Well, my last segment here is about a review that someone did on the Framework 13 AMD laptop on Linux. So you know we often discuss the Framework laptops on this podcast, but I found this particular review quite unique. Not only does it focus on the AMD 7040U processor, which is a new processor from AMD for mobile you know, laptops and whatnot but it also takes a gaming and Linux-centric approach and it goes beyond the simple listing hardware specs and delves into the actual user experience and impressions of Linux on this machine. Now, the review does cover some of the hardware improvements, such as a better screen Now it's got a 2880 by 1920 resolution and 144 hertz refresh rate. The reviewer also mentions an improved hinge compared to older generations of the laptop and there's other enhancements that you can read about in the article. But they've had the previous generation, this one, so they can kind of compare and contrast the changes that framework is still making to their product. But what we're most interested in here is the Linux experience 404 on the laptop and noting that they aren't a typical Ubuntu user, I mean and they even provide instructions on how to completely remove snap packages from the distribution and prevent them from being installed again. You know, but that's not the main focus of the story, but there's a bit of a rant in there about snap packages.

01:02:58
It's just one little paragraph so it's not huge, but the review includes benchmarks comparing this 13th generation framework to an Intel 12th generational model, and the AMD system performs very well and is noticeably faster than the Intel system. And the Intel had a Core i7-1260p CPU and both CPUs have 16 threads CPU and both CPUs have 16 threads. The Intel does have some P and E cores, so some of the E cores don't have hyper-threading where the P's do, but bottom line they both can take 16 threads. The author also discusses battery life and while it's generally good, they point out that sleep mode isn't working perfectly in Linux. They said closing the laptop at 9 pm and opening it again at 9 am results in a 5 to 10 percent battery drain. However, they also mentioned that the battery is larger on this model so it's not as significant of an issue as it was on the previous version.

01:03:57
Interestingly, the older Intel version had issues with gaming because the drivers weren't mature enough and some games failed the launch. But the author is very enthusiastic about the AMD version's gaming capabilities. Not only can it launch games, but it also plays them very well even without a dedicated GPU. To quote the author, the 7840U APU that's the graphics part of this chip is a beast. The author is happy with the laptop after two to three months of use, which gives this review more depth than the typical.

01:04:31
I used it for a week review, so you know they do mention a few minor issues, but nothing show-stopping. So it's pretty detailed. Having used it for that long, you know they really got the ins and outs of it. So it's pretty detailed. Having used it for that long, you know they really got the ins and outs of it. I appreciate the article not only for its detailed analysis but for its comparison between this year's model and last year's model and I highly recommend checking out the article linked in the show notes for a thorough review of both the hardware and software experience. But bottom line it was really good and he loves this laptop and thinks it's a lot better than the previous years yeah.

01:05:05 - Jonathan Bennett (Host)
So I bought a framework 16 for my dad. Uh, he had a job change and had a work laptop, needed to get something, and so he he went to me. He goes, can you build me a laptop? Like, well, actually I can if we go with this one company. So we did a framework for him and so I got to. I got to play with it a little bit, getting it set up for him, and I was. I was quite impressed with the 16 um and so I've got a. Uh, I've got an amd framework 13 on my wish list, um, so if someone wants to donate to me $1,644 worth of coffee plus tax, I'll have to work that out later Exactly what that ends up being. A lot of coffee.

01:05:53 - Jeff Massie (Co-host)
Would you say what? 1,600?.

01:05:55 - Jonathan Bennett (Host)
It's $1,644 is the framework 13 as I would want it set up? Yeah, 200 something, 329 coffees there you go. That's a bunch of coffees. He did the math, I did the math, but yeah, I'm I'm a huge fan of the framework. Something I've been thinking about here recently is, um, doing some custom 3d printed framework modules to go into the slots.

01:06:18 - Jeff Massie (Co-host)
There's, there's some fun ideas I have for that now to to be to truly review, though, I think you also need to get a system 76 laptop so you can compare and contrast them, and and it the newest model of both companies.

01:06:34 - David Ruggles (Co-host)
Yeah, the 76. They just released a uh one based on the Ryzen 9. I saw that article come across my feed. I can't put my fingers on it right now, but it looked impressive.

01:06:50 - Jonathan Bennett (Host)
As soon as you can find a sponsor to take care of the cost on that, I will be glad to pick up those laptops and review them for you.

01:06:58 - Jeff Massie (Co-host)
If someone can donate about 700 coffees, I think we'll be good yeah.

01:07:06 - Rob Campbell (Co-host)
Yeah, let's quadruple that and get one for all of us. There you go.

01:07:13 - Jeff Massie (Co-host)
Well, that would be two for all of us Well right. So if we had a good 12, well, probably 1300 coffees, we could each get one.

01:07:26 - Jonathan Bennett (Host)
You know, spare change for some people, you know yeah, yeah, there are probably more cost-effective ways to do that rather than going through. Buy me a coffee, um, anyway, oh my goodness. We are so far. We are so far off the path now let's talk about uh oh, jumbo data packets.

01:07:44 - David Ruggles (Co-host)
Oh, interesting stuff happening in the linux network stack yes, so in um let's see the patches have been merged to next dash net dot get making it material for the networking code in the next kernel merge window, which is Linux 6.14, and it should be opening up late January. But with that context what we're looking at is jumbo data packet transmission and rack TLP coming to Linux. The jumbo data packet transmission is code, so David Howells of Red Hat has been working to implement this, and it is code to enhance the ability to send UDP packets that are significantly larger than the default MTU of 1412 bytes. Mtu is maximum transmission unit and it is a number that tells networking equipment how large a packet of data they can transmit at one time. Between endpoints the MTU can actually fluctuate because, depending on the path you're going over, some networks may be able to handle jumbo frames and some may not, and so that's where you get packet fragmentation and things, which is where they have to break up a packet to get it across. A segment of the network that doesn't have well, has smaller MTU sizes and then reassemble it. Well, has smaller MTU sizes and then reassemble it. But some of the things that he mentions is an MTU of 8,192 would allow five of the standard 1412 UDP packets to be transmitted as one. And then you also have the issue, and he said an alternative and possibly more efficient way would be to be able to expand and shrink the capacity of each data packet to match the MTU and the save on header and tail gap. Overhead Tail gap is where you your packet is smaller than the maximum MTU but because it's smaller you still have to fill up that full size of the NTU. And he mentioned specifically that IPv6 over Wi-Fi doesn't. The NTU is smaller even than the standard 1412. And I haven't looked into it, but I would assume that's because the header information for IPv6 with the larger address fields would be taking up more of that 1412 maximum capacity.

01:10:31
The other interesting thing that he talks about is RAC TLP, and RAC TLP is an enhancement to the TCP IP acknowledgement process. Taking advantage of so the original ACK process is where the server sends a whole bunch of packets. Then the client, when it gets the packets and it's reassembled them, it acknowledges them, it sends an ACK back and then the server waits for that ACK for a certain amount of time and if it doesn't get, it goes. Oh, they must not got it, so it resends it again. Well, if the timing is off, um, if some of the packets didn't arrive, but they arrived out of order and the reassembly didn't happen quickly enough, you can get into situations where you're retransmitting data that was actually received and then that's just adding to congestion. So an enhancement that they made a while back was called a SAC, or Selective Acknowledgement, and that's where you can acknowledge specific parts of the entire window of data that you're supposed to receive so that you can say, hey, you know I got these three but not those two.

01:11:40
And then RAC builds on that with some timing, some dynamic timing. So basically it's constantly monitoring the transmissions and the acts, it's getting the sacks, it's getting back um to see how long it's actually taking, and each time it gets a selective acknowledgement back, it adjusts that timer um so that it has the actual time plus a buffer, so that it's not wasting time on resending. But it's also not just resending stuff without knowing that it didn't get there. And if you click through to the RFC that's linked, which RFC is request for comment and is how pretty much all of the standards that operate networks and the internet are developed in the open, they actually go through an example in there of how that can look, where you're having multiple losses, but it just brings a lot of efficiency.

01:12:39
And then the rack-tlp. The tlp is tail loss probe, and that is where, at the tail end of a network communication or conversation, you can wind up not sending as much data because maybe there's not as much to go, and so you get a situation where you don't have this windowing information so it can actually intentionally send packets just looking for acknowledgments, just to be able to rebuild that windowing information. So anyway, these two enhancements being built in, they've already been developed and they're being implemented in network equipment, but having them here in the operating system is just going to make for more efficient networking moving forward, which is always a good thing, interesting stuff.

01:13:28 - Jonathan Bennett (Host)
A lot of that doesn't have a whole lot of use case for small networks. It's more like big iron stuff where you've got like 10 gig back ends and you want to be able to just smash as much traffic as possible through there.

01:13:43 - David Ruggles (Co-host)
But a lot of that rack. Tlp stuff is where you've got, like, your data centers with that 10 gig but then you're sending it out over the internet. So you need to be able to adjust your mtu's and stuff um on the fly so that you're not using your 10 gig to crush your internet connection or two or three isps out yep, yep, yeah, neat stuff.

01:14:03 - Jonathan Bennett (Host)
And and of course you know you see these things they'll land in the kernel, they'll connection or two or three ISPs out. Yep, yep, yep, yeah, neat stuff. And and of course you know you see these things they'll land in the kernel, they'll get used in data centers and, the next thing you know, somebody will have some implementation in a router that makes use of some of it in one way or another. Right, so it does. It does sort of trickle down as it eventually. All right, uh, rob, is this a, uh, is this a political statement about a certain social network?

01:14:30 - Rob Campbell (Co-host)
yes, it is time to get rid of x. No, I'm not talking about twitter. I'm talking about x11 or xorg or you know whatever name you want to call it this decade. You know the x, that was x long before twitter, and that's why you don't name your company a name that's already taken, you know. You know, whenever I've started a company I've, you know, I usually google first. I was like anybody got this, all right, I'll take it then. But usually the first hundred names I pick are already taken.

01:15:10 - Jonathan Bennett (Host)
Company names are just like email addresses. All the good ones are taken.

01:15:13 - Rob Campbell (Co-host)
Yeah, so I don't take any X. I mean, come on. Anyway, I'm talking about another X that is also circling the drain, as one of the last major desktop environments has finally implemented Wayland as an experimental feature, and that desktop environment is XFCE. The desktop, long known for being slow to change but stable as a rock, has released XFCE version 4.20. And only two years since their last release of 4.18. But for once it's kind of packed with new features, and the biggest new features is experimental Wayland support. That can be started with a simple start xfce4 space dash, wayland. However, it's still super early, with several xfce components that aren't going to work yet, such as xfwm for the window manager, x dashboard, xfce4 screensaver who needs a screensaver these days? Xfce4-window ck-plugin and xfce4-xkb-plugin. I don't know what most of these do and haven't had a chance to look into them. Um, sounds like a lot of things that don't work yet. Maybe you can get by without them, I don't know. But the XFCE devs say quote so far XFCE does not feature a compositor which supports Wayland. If you want to run XFCE in Wayland, labwc or Wayfire will give you the best results. Plans are underway to add Wayland support to XFWM4 while preserving its existing X11 functionality. However, such a restructurization will be a major effort and we cannot tell yet when slash if it'll be done. So please don't hold your breath. You know you got another couple of years for the next release and then you know there'll be a little more there in a couple more years and we'll see Other changes included in this.

01:17:44
There's support for svg wallpapers on large screens, support for setting custom colors uh for icon labels and icon backgrounds. A new option to store folders uh, before files uh. Enhanced context menus, support for configuring all XF desktop shortcuts. Thuner, xfc's file manager received support for IPv6 remote URLs, along with other improvements new toolbar buttons, search improvements. Like I always say, I'm still holding out for IPv8. There's also more configuration options for the panel launcher improvements, better support for handling dark themes, which many people love, and many more. The desktop environment that rarely changes is improving and, if all goes well, hopefully XFCE will be fully whaling ready in a few years or so.

01:18:46
Uh, just the time to fully finally deprecate x. So if you're looking to get your hands on a new xfce 4.20 um desktop environment, slackware based, porto 1, spelled because I don't really know how to say it. I even looked on Google how to say it and it sounds French and I can't. It was tough for me. So it's spelled P-O-R-T-E-U-X and the X is capital. So I don't know if that means it's Porto X. So I don't know if that means it's Porto X or if it's Portox or I don't know. So that's how it is. That's what it is. It's a Slackware based Slackware, not known to be one of the easier ones, so it's probably not one of the easier desktops for people new or who just want something easy. I don't know. But it is one of the first to ship with XFCE 4.20. But you know, within the next couple of release cycles, just around the corner, we'll likely see 4.20 coming in many other more mainstream distros in the next few months 4.20 is one of those interesting numbers.

01:20:04 - Jonathan Bennett (Host)
more mainstream distros in the next few months. 420 is one of those interesting numbers. I just can't help but notice.

01:20:08 - Rob Campbell (Co-host)
Were you giggling every time? I said it Sort of yeah, 420. I mean, it really made me think of X and Elon Musk also.

01:20:17 - David Ruggles (Co-host)
Yes, that's why it just all tied together I was totally expecting you at some point to say that it will have Wayland finalized just in time to start deprecating Wayland.

01:20:32 - Jeff Massie (Co-host)
Yeah, I thought that's where you were going to.

01:20:37 - Rob Campbell (Co-host)
I have much higher hopes than that, just because X lasted for a long time.

01:20:44 - David Ruggles (Co-host)
you hope wayland will for a few more years oh, I know, I know the point is not that wayland's on the way out, the point is that xfc moves so slowly I?

01:20:54 - Rob Campbell (Co-host)
I just have higher hopes, wishful thinking, maybe because x needs to go and Waylon needs to take the lead Everywhere.

01:21:07 - Jonathan Bennett (Host)
Indeed Interesting stuff. Alright, so we're moving to some command line tips. I think it's time.

01:21:15 - Rob Campbell (Co-host)
Let's do it Alright.

01:21:16 - Jonathan Bennett (Host)
Jeff is up first with COM.

01:21:20 - Jeff Massie (Co-host)
Yeah. So today's command line tip and it's actually a command line tip is the com command C-O-M-M. So this command allows you to compare two sorted files that's an important detail sorted line by line, identifying lines that are unique to each file and those that are common to both. And this is particularly useful if you have lists or log data you need to compare. So this is a little different than, like, say, diff or something like that. So the com command performs a line by line comparison and by default, generates a three column output. So column one contains lines unique to the first file, column two contains lines unique to the second file and column two contains lines unique to the second file and column three contains lines common to both files. Now, if you look at the example in the show notes, there's a nice article in there about using it. But you know, and the three column format should be easy to visualize because it's just three columns.

01:22:19
Now, most commonly used options for this command are a dash one, a dash two or a dash three, or it can be and dash three. So these options simply suppress what you specify. So, for example, using dash one with the command will suppress the first column. You can use more than one of these options at the same time. For example, using one and two together will suppress both the first and second columns, leaving you with only the lines that are common to both files. Conversely, using dash one and dash three will show you only what is unique to the second file. There are other switches you can use with com as well, but they're not really that common. I mean, I wasn't trying to make a pun there, but it's just the dash one, two and three are kind of the most used ones. Take a look at the link in the show notes or consult your man page you know man space C-O-M-M to learn more about this useful command.

01:23:18 - Jonathan Bennett (Host)
All right, very good, let's see Up. Next is David talking about expansion, bash expansion. David is muted. We cannot hear you, david. Yes, I am.

01:23:34 - David Ruggles (Co-host)
So let me pop over here, all right. So I am confident that bash expansion has been talked about before, but it has been a long time ago and, as we often mention on here, while we sometimes find brand new command line tips also, sometimes we either have to deal with a problem or get bit by something. And I got bit by Bash command line expansion, specifically dot versus asterisk. So I'm going to explain why that's a big deal. So I created an example directory here and if I do an ls, I see I've got a dir1 and a file. But if I do an ls minus a, which is going to show everything we can see, I have some hidden objects here I have a hiddendir2 and a hiddenfile. So what was happening was I was trying to locate some contents in the file using grep. So we've talked about grep in the past.

01:24:36
I'm going to use grep in this example, but I'm not going to explain it. But it lets you search inside the files and you can give the path a dash R for recursive searching. So if I grep ZZ I'm sorry, if I grep ZZ in the current files, it's going to tell me that dir1 is a directory and file contains ZZ and file contains zz. Now if I add dash r like this, that's going to tell it recursive. So now it's going to actually go down into dir1, and it's going to tell me that inside dir1, there are two files file which has zz and file which is a hidden file which also contains zz, and then my file, zz, but it's missing the entire dir2 folder. So if I change that asterisk to a dot, now it's going to search in the current directory instead of enumerating what it can see, and now it'll give me all of the examples.

01:25:41
But if you notice, whereas before it said dir1 slash file, dir1 slash, dot file file, now it's saying dot slash, and then that same information, but it also has the hidden dir2. The reason for that is because the dot says search in the current directory, but the asterisk is just doing the same thing as a basic LS and it is saying here's dir1 and here's file. And we can actually see that by echoing the asterisk, which will give you the same results as doing just an LS. So it's not a big thing, but it was something that bit me because when I was in a subdirectory I was getting a different result than when I was in a root directory and I realized what it was, was the directory I was changing in and out of was hidden and it wasn't showing up. And I was like, ah, forehead smack. And I was like, hey, you know, I'll share this.

01:26:38 - Rob Campbell (Co-host)
Have a good laugh at myself and, uh, just remind everybody that you know, sometimes these things matter indeed, you know, I've always just been using all that with dot slash, but I suppose that slash uh is unnecessary just one extra keystroke. But you don't need and I've never saved the keystrokes I've never even thought to use the asterisk, but now that you've taught me the difference here and also that I don't need that extra keystroke, that'll save me lots of seconds that will add up over the years.

01:27:12 - Jonathan Bennett (Host)
Yeah, a lot of RSI saved as well. Yes, all right, jeff has been in contact with the FBI. Apparently. What that's not Jeff's Is that not, yeah, I'm like wait, wait. Rob has been in contact with the fbi.

01:27:30 - David Ruggles (Co-host)
Sorry, let's try not to blow your cover, rob, and that is much less surprising yeah all right, let's show my screen here, because this is a command line tip.

01:27:38 - Rob Campbell (Co-host)
Let's show my screen here, because this is a command line tip and you all viewers need to see this one because it's a visual one. Even though it's command line, it's visual. So FBI is frame buffer image. It's a frame buffer image viewer is what FBI is. So right here, those watching this is on an Ubuntu server, no GUI, no desktop environment, just Ubuntu server. It's a VM, but what you're seeing here is the equivalent of just a TTY. If you had a monitor attached with no GUI and it just boots up into your command line.

01:28:19
So FBI, if I type F, oops, let me type that on the right screen here. Put my cursor over here, fbi, and then dash T1, that's the TTY, so you can have it show on TTY 1, 2, 3, whatever you have there, have there. And then I'm going to put a dash A is going to auto zoom to fit, so it's going to, it's going to make it, the image, fit the screen. Then I'm going to do a dash T, space three, which is dash T and the three is second, so it's for a continuous slideshow. And then I've selected two files that are in the directory that I'm in, and you do have to use sudo, from what I've noticed, because it needs access to the TTY Maybepg space, robertcampbelljpg. So I don't need to have all those. But I'm going to show you a bunch of different commands here. There are flags that you can add onto this and so you could use this, flags that you can add onto this and so you could use this.

01:29:43
So if maybe you want to just do a screensaver and you don't want to, you have like a screen, a Raspberry Pi attached to it. All you want to do is have a slideshow. That's what I meant to say. You don't need a full GUI for this. You don't need a full GUI for this. It will do it right in the frame buffer.

01:30:05
Now, that's a really small picture, so that's why the picture of me is blurry. But yeah, you don't need to install a GUI. That's all you need is just a basic Linux install and you can do a slideshow. Now I removed the dash A, so now it's not zooming in. That's how small the picture originally is.

01:30:29
There's more details on the bottom. You can't necessarily see. The image is behind me, the image name Way over on the right, the size 100%, and then it has the actual size. The antenna length shows 600 by 600. The antenna length show is 600 by 600. The second image is 150 by 202, and there is an H for health. But you have to escape out of that so you can have it just boot up into that and easily have a slideshow or whatever other purpose. Maybe it's a headless machine but for some reason you have a monitor on and you want to just boot up some logo or I don't know what you want. But you can have images in the frame buffer without x or wayland or any gui on their frame buffer interface I like it.

01:31:17
Interface, there we go. That's what the eye is.

01:31:19 - Jonathan Bennett (Host)
Yes, all right. So my tip is one. I actually use this, not this week, a couple of weeks ago, uh, it's, it's absentee or abcde, and that is a better cd encoder. And this is a really simple uh, cd audio ripper for the command line and for uses. You just go to the folder you want and you can just run ABCDE and it will go out and find your CD drive, rip it. I think it goes to Flacker MP3 by default, I think, but there's all kinds of flags you can give it to give it to tell exactly how you wanted to do it.

01:31:58
But yeah, it just for me at least, it just works, and I had a CD and I needed to get the audio off of it. I've got a link to their main page, which has links off to documentation about how to do things. It is pretty simple though, and from what I can tell, absidy is installable from the repositories in most of the major distros. I know it's in Fedora and it appears to be in Ubuntu as well. So if you need a simple command line CD ripper, there you go. Fun stuff, all right, we have made it through the show, and I want to give the guys each of them a chance to plug whatever they want to, or if they want to get the last word on something. We'll start with Rob, and he is asking for a whole lot of coffee lots of coffee.

01:32:48 - Rob Campbell (Co-host)
I'm thirsty guys. All right, as always, you can come connect with me. Find my links at robertpcampbellcom. On there. There's links to my linkedin, my Twitter, my Mastodon I'll get blue sky up there someday. And here's the important one, where you can click on this little coffee cup to donate me coffees. Sean recently donated 10 coffees, so that way I would have to use Linux Mint for two months. As I said on a previous show and I'm not sure if he's the same Sean, but if you guys remember, there was a Sean that donated coffees to everybody and then donated coffees to me when I felt sad because he didn't donate them to me. So I don't know if it's the same Sean or another Sean. Maybe it says if I click in but Sean's a good guy.

01:33:45 - Jonathan Bennett (Host)
Shaking Sean down for coffee money.

01:33:47 - Rob Campbell (Co-host)
Thanks for this coffee. Sean, Sean or Sean's, I don't know All right.

01:33:53 - Jonathan Bennett (Host)
David.

01:33:56 - David Ruggles (Co-host)
Oh. So to be slightly serious, but also, you know, not to be depressing or anything, but this time of year is can be uncomfortable for some people, that you can be going through stuff, and I only mention this because I've had this situation myself in the past and so I'm very aware of it. So if you're listening to this and you're, this just isn't the best season for you, just know that there are people that care for you, that you are important and reach out, reach out to a friend to find some support structure to be around you, but you're important.

01:34:37 - Jonathan Bennett (Host)
Yeah, absolutely, absolutely. All right, Jeff.

01:34:41 - Jeff Massie (Co-host)
You can find me on LinkedIn and M mastodon. Just go find rob and then I'm connected to him in there. That'll, that'll work, but more importantly for this last show of the year. So we got to make a good one poetry corner. This has got to tide you over for a week, for two weeks. Roses are red, my server is gray. I'm computer nerd, don't expect me to rhyme. Have a great week and a great holiday everybody.

01:35:10 - Rob Campbell (Co-host)
So, to be clear, to find Jeff, you go to robertpcampbellcom, click on LinkedIn, which was the far right one as you look at the screen. Click on that, search for jeff massey and you'll be there that's hilarious.

01:35:27 - David Ruggles (Co-host)
Technically, you should be able to find me that way too all right, david ruggles is there too yeah, we're, we're all, we're all at one degree of rob there's the show title. Oh, all right.

01:35:46 - Jonathan Bennett (Host)
Uh, just a couple of housekeeping things. We will not be having a ULS next week. That is the Saturday between Christmas and new years. We're taking the day off. I think all of twit is taking that time period off and get some much needed rest and relaxation. Um, if you are a Floss Weekly listener, we are taking a couple of weeks off there as well. We will actually be back on Floss on the 31st to talk about FR Routing, which is a really cool project doing some deep internet routing stuff. And I don't know, maybe I can talk to Robin to coming with me to be the co-host on that one. We'll see, but that's what's there.

01:36:25
And then you can follow my work over at Hackaday on not only Fossil Weehoo but also the security column. We didn't run one this Friday, not because of the holidays but because I was sick as a dog. On Thursday I spent all day out here staring at my computer trying to write security article, but just I had the brain fog too bad I couldn't do it. We should be back next week for that as well. But yeah, you can follow me there and appreciate the guys being here for the show today and we appreciate everybody that watches and listens both the guest live and on the download. We will see you in two weeks on the Untitled Linux Show. Thanks. We will see you in two weeks on the Untitled Linux Show. Thanks.