hpr-knowledge-base/hpr_transcripts/hpr4425.txt

Episode: 4425
Title: HPR4425: Introducing Linux Matters
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr4425/hpr4425.mp3
Transcribed: 2025-10-26 00:36:13

---

This is Hacker Public Radio episode 4,425 for Friday the 18th of July 2025.
Today's show is entitled Introducing Linux Matters.
It is part of the series podcast recommendations.
It is hosted by Ken Fallon and is about 38 minutes long.
It carries a clean flag.
The summary is long overdue.
We share a teaser of another excellent Creative Commons podcast.
Today's show is licensed under a Creative Commons Attribution Non-Commercial License.
Hi everybody, my name is Ken Fallon.
You're listening to another episode of Hacker Public Radio today is another in our series
of recommendations for Creative Commons podcasts.
This one is going to be Creative Commons Attribution Non-Commercial 4.0 International License.
And it is Linux Matters, which you can find at litinixmatters.sh.
If you go to the OAB page, it says, join three experienced open source professionals as
they discuss the impact of Linux on their daily lives, upbeat family, friendly banter,
conversation and discussions for Linux enthusiasts, casual observers of all ages.
This of course is our very good friends of the podcast, Alan Pope, Mark Johnson and
Martin Wimpress, who we met as a camp, except for Martin, who I managed to avoid the whole
weekend, but that shall be rectified at the next one with any luck.
This is a very nice, easy-going podcast actually.
Little less on the free years of freedom and more on the programming side from my personal
experience.
But if you're not subscribed, the links to do so will be in the show notes for this episode.
So sit back, relax and enjoy the podcast.
I have successfully pulled off the great GitHub Space Heist.
Did they catch you?
No, they didn't.
It's brilliant.
Well, they're not talking about it on a podcast then.
It's all right.
Nobody's listening.
It's doubly awkward because I've made my space high-stay GitHub action.
So one way or another, they know I've done it.
OK.
What are you talking about?
OK.
So I build lots of things with GitHub actions, and that's all just fine and dandy.
But occasionally I need to build big things.
And what I'm about to explain is kind of focused on building things with nicks, but actually
the little life hack that I've found, you can apply to other things as well.
And I'll get into a little bit of that in a moment.
But the concept here is, I needed lots of space to build large nicks configurations,
particularly my OS configurations for my workstations, which have lots of big stuff in them,
like browsers, DaVinci Resolve, audio players, a whole bunch of electron applications,
full development environments, lots of language ecosystems, all of that sort of stuff.
And consequently, that exhausts what at first glance appears to be 20GB of available space
on a standard GitHub runner.
That's probably Ubuntu.
Yes.
So it's running Ubuntu latest or whatever version you want to use.
And there's a GitHub action from determinate systems that installs nicks for you.
And then you can do nicks things in an Ubuntu machine.
And like nicks builds for like small packages are just fine, but for these OS things,
not so good, because it runs out of space.
And the reason why I want to build my whole configuration is because if I build it,
then I can cache the results in my own private, flake cache.
And then that means that my system updates are fast and I don't have to do any local
compiling or anything like that.
So that was my motivation.
I wanted to figure out how to make more space.
And looking around in the GitHub Actions Marketplace, there are actions which delete lots of
things from the Ubuntu runners in order to claim back some space.
And I thought, well, that's an interesting idea.
We could do some of that.
The problem with that is it takes a long time, we'll get back to that in a little bit.
But what I did is I made a little debug thing to sort of go an action to go poking around
in the Ubuntu runner to sort of tell me about it.
When I discovered there is a slash MNT partition in a GitHub runner.
Historically, those used to be 14 gigabytes, which is interesting.
And this is 14 gigabytes of space that you wouldn't ordinarily have access to.
But today they are now 65 gigabytes inside in the standard GitHub runner.
That's free space.
Can I upload all of my photos into there or something?
Yeah.
The idea I had was what if I created a GitHub action that creates a partition inside that space
as the slash nix volume, where nix does almost all of its things.
So the long story short is that's exactly what I did.
And I called this GitHub action nothing but nix.
And this is for those people who are running a nix workflow and all they care about
is having enough space for nix.
And the first thing it does is it creates a sparse file using fallocate for the size of
the mount partition, less a little bit of reserve, which you can configure.
And then it formats that sparse file with butter FS in a raid zero configuration.
And then mounts that volume as slash nix and have wallet.
You now have 65 gigabytes of space for your slash nix volume.
Two questions.
Number one, I find it interesting that you pronounce it fallocate.
Moving on.
Is it any particular reason you use butter FS?
Yes, I'll get to that in just a moment.
It's because you can do clever things with butter FS that you can't do with the other
file systems unless you get into using LVM and stuff like that.
And I wanted to keep it as simple as possible.
So having created my 65 gigabytes of space, I then thought, hmm, well, there's this other
sort of approximate 20 gigabytes of space in the root partition.
What if I could claim some of that back as well?
So then what it does is it goes and creates another sparse file in the root partition.
It then enrolls that sparse file into the butter FS volume.
I've already created and then grows and balances that file system.
So now you have 85 gigabytes of space available.
So this is all going swimmingly at this point.
And what was interesting about this is that you do run into issues.
If you exhaust all of the space in the root file system, then you know, bad things start
happening.
And there are some things that are happening in the root file system that you need to
preserve some space for.
So I then had the bright idea of well, let's do the other strategy that everyone else does,
which is going to delete a bunch of stuff.
So instead of just doing big deletes of files, I thought that we could do this properly.
So the action creates the 65 gigabyte volume, which it can do in about one second.
So you know, instantly 65 gigabytes, you can then move on with the rest of your work
flow.
So in the background, nothing but nicks, then creates a new process that runs in the background
that goes off and elegantly apt removes and purges a whole bunch of the stuff that the
GitHub runners automatically provide in order to give, you know, every language imaginable,
every build tool imaginable.
You talk about doing it properly with apt, you know, auto removal, whatever.
So some of the other actions just like RM, RF, like slash user bin and stuff like that,
they do.
And in fact, I have to do a little bit of that as well and we'll get into that.
So whilst the rest of the action is running, nothing but nicks has kicked off this large
apt get remove purge in the background and it removes gigabytes of stuff.
Then the next thing it does is in slash user local on the runners is a whole bucket load
of just random stuff that's been jammed in there that hasn't been apt installed in a
correct way.
And I thought, well, I don't need any of that.
So I can just RM, RF that just brutally extract it.
And this is where you run into an interesting problem turns out that takes a long time.
RM minus RF on a large slash user local.
I think it's like 10 or 15 gigabytes of stuff that's in there.
It took between 11 and 16 minutes to RM minus RF that from the VM that is the Ubuntu
host runner, which is a bit too long, like you couldn't.
And this is the problem with some of these other actions.
They do that.
RM minus RF, you know, as a as a blocking action before you get the space back.
Anyway, I found out there's some tools and they're called the fast unix collection
or something.
Anyway, they have two commands, CP Z and RM Z and they are super fast implementations
of RM and CP written in rust that are hyper paralyzed and that deletion of slash user local
that 11 to 16 minutes goes down to about 30 seconds, which is quite astonishing.
So you ran this on the Ubuntu latest or whatever, the default GitHub action runner that
most people use is, have you tried this on any others like I don't even know what other
options are available on GitHub or are some of the alternatives like blacksmith, I think
is an alternative.
Do you know if they present like huge volumes of unknown like blank space to the runners?
Blacksmith.sh do just present lots of space in the root file systems.
You don't have to do any of this shenanigans and if you're looking for a fast alternative
to GitHub actions that support GitHub actions are a really good option, you just change
one line in your workflow to say you're using blacksmith and all of the other stuff just
works.
It's great, but I didn't want to pay for that and I didn't want to pay GitHub, I wanted
to have my cake and eat it too.
So yeah, I'm doing all of this crazy stuff, but what happens now is that nothing but
nicks creates the 65 gigabytes, runs this background task that uninstalls a bunch
of stuff, then goes and deletes a bunch of stuff as fast as possible and then creates
this additional volume, which is now 40 gigabytes.
And you can create somewhere between sort of 100 and 130 gigabytes for your, your
nicks volume.
Sorry, my, my math is sailing me there.
You said the root position was 20 gigabytes and then you delete some stuff from it and
you've got a 40 gigabyte volume.
How does that happen?
Sorry.
The free space in the, oh, I see, repartition by default is 20 gigabytes and then nothing
but nicks goes and deletes another 40 to 50 gigabytes of stuff out there, so it's
actually understood.
In the end, the core Ubuntu that's left behind is only like four and a half gigabytes,
but there's like an 80 gigabyte volume that's been allocated for the root partition.
So all of that deletion and purging happens in the background and then when that space
is available, that nicks volume is dynamically grown with butter FS to be as large as it
needs to be.
And so I'm very happy I can now run my large nicks configuration jobs and all of my stuff
is cash to all my updates are super quick and I'll have links in the show notes to nothing
but nicks itself for anyone that is using nicks and also a blog post that I explained
just like how this works under the hood.
And one of the things I was thinking about is, well, this is an interesting trick like
if you need to build large container images, you could do the same trick but mount up the
partitions for where you want to put your container images, for example, maybe you want
to put in lots of container images in order to complete your build.
But the idea I've had is taking something like QuickMU, a project that the three of us
have worked on in the past because in GitHub actions, the free tier and the paid tier,
you can enable KVM acceleration to do pass through acceleration to virtual machines that
you can run on the virtual machine that is the GitHub runner.
So my idea is, you could use this extra space to provision your own virtual machine image
in a GitHub runner, enable the KVM acceleration and then boot your own VM inside a GitHub runner
in order to do like your golden build using your validated certified and secured build
environments or build tons of container images that you wouldn't have space to do, you
know, otherwise.
So that's my little life hack for GitHub.
I'm really pleased with that and I'd be interested to hear if anyone else finds a need for
creating gobs of disk space on GitHub.
This episode is sponsored by TaleScale.
Go to talescale.com slash Linux Matters.
TaleScale is a proper VPN for connecting your devices and infrastructure together, not
something that's marketed at people watching foreign streaming services or who want to
hide their browsing habits from coffee shop staff.
It's the easiest way to connect devices and services to each other wherever they are.
PCs and laptops, servers, cloud instances, 3D printers, network storage, connect anything
and everything you want wherever it is and get started in minutes.
TaleScale's personal plan will always be free and you get up to 100 devices and 3 users
for free with no credit card required.
The free personal plan is enough to get you started if, like me, you want to host services
on your LAN and access them from anywhere on the internet securely, no messing with
routers, no port forwarding and no messing with firewalls.
But what gives you the confidence to stick with TaleScale is their straightforward business
model.
If you want to use TaleScale as an industrial scale with more users and devices or take
advantage of features like mobile device management, auditing, network flow and logging
and priority support, you can join over 10,000 companies as a paying customer.
So show support for the show and go to taleScale.com slash Linux Matters and try TaleScale out
at home or work today.
That's taleScale.com slash Linux Matters.
I rebooted an old website that has been dead for years and it's not mine, it's a website
that somebody else made and then for some reason it shut down and I decided to reboot it.
And I've been thinking and talking about doing this since about 2021 maybe before then.
But over coffee the other day Martin kind of gave me the nudge and I just did it.
So now I now have another domain and website to look after.
To be fair, I didn't realize that I was nudging you to do this thing, I was showing you something
else.
It started by moving your blog onto a different platform, but then it kind of snowballed.
Yeah.
So this is called nerdy day trips and it was started back in 2011 by Ben Goldakos, a famous
doctor and kind of celebrity doctor who writes for a number of organs, although it doesn't
quite so much right now and he was looking for like a database of nerdy places he could
go when he's in a particular town and he wanted a map so he could like find where these
nerdy places are and go to them like it could be a museum, it could be a very specific
museum like a transport museum or a pen museum or it could be just some architecture that
people go nerdy about, right?
It's not necessarily a computer thing, it's whatever you think, whatever somebody finds
to be nerdy.
Yes, exactly.
And it was basically just a big map with loads of pins in it, a Google map with loads of
pins in it.
And the data was crowdsourced back then starting in 2011 and there were a lot of pins put
into it and then the site kind of disappeared in about 2016.
Now there were problems.
I remember seeing spam attacks on the site and seeing that they didn't have very good
spam protection and so there were lots of nefarious posts being placed.
And I don't know quite why they shut the thing down, it could be a combination of things
but the fact that Google changed the way they license Google maps could be one, maybe
getting sick of all the spam will be another, it's a project that maybe ran its course.
So while nobody's like publicly acknowledged why it went away, it just went away and that's
fine.
You know, sometimes things die but also sometimes things are reborn and so I decided
to like the Phoenix.
Yes, I decided to rebirth this Phoenix and it started like I said in 2021, I was chatting
with Wilcook and Stuart Langridge, my good friends and we thought and we'd chat it and
every so often we'd like throw together some mockups or a bit of data and one of the
problems is seeding a website like nerdy day trips with the pins, like the places.
If you start with an empty map, it's not very motivating for other people to crowdsource
that data.
You know, if Jimmy Wales opened a blank website and said, Hey, this is Wikipedia, everyone
get go crazy.
I don't think anyone would be throwing money at him.
So I was chatting to Stuart about this some years ago and he went through the internet
archive and pulled some of the data, scraped some of the data, not all of it but only some
of it because of the architecture of their site, it was the internet archive doesn't have
a full copy of the data, it only has whatever was on screen when the archive caught it and
so we have some data like about a thousand pins or something like that and I took that
blob of Jason and I mashed it a little bit and turned it into some markdown files and
because everything gets fixed with Hugo in my mind, I used Hugo to build a little website
in which there is a map with all the pins in it and there's some buttons to do various
things like search and look at the details of each of the, each of where the pins are.
So as you describe this, Hugo is a static site generator, this, what you're describing
doesn't sound like a static website, so how did you pull this off?
Well, I'm glad you asked.
Each of the locations is a static page, it's a piece of markdown with a bit of front
matter that's the GPS location, that long, an address, a title and maybe some other metadata
about the place, right?
And there's one file per location and they're in subdirectories based on country, region,
that kind of thing so that you haven't got just like one big folder full of markdown files.
It's about like 30 folders full of markdown files and the site gets built and published
like, initially I did this locally just on my local workstation and it works like it,
the site uses the data in the markdown files to build a JSON blob which gets loaded and
we use leaflet.js to show the map and we're using OpenStreamMap instead of Google Maps
and you see all the pins and yes, it's a static site, there is no editing, it's not like
where you be, you can't just press a button and edit what's on the page, however, I pushed
all of this into GitHub, so now you can edit that stuff.
You can go to GitHub and go to the nerdy day trips project and you'll see website is
in there and you'll find all the little markdown files buried in a folder or bunch of folders
and you could edit those to update them or you could create new ones.
What then happens is if any changes are made either to the website itself or to that data,
it gets built and pushed to Cloudflare pages.
I could have used GitHub pages but I used Cloudflare pages, thought I'd try something new
to me anyway and so every time the site is updated, it gets rolled out a new build and
within seconds the site is updated, the life site that is.
The problem is it's a bit brutal to ask normies to go and edit a GitHub repo, they might
not even have a GitHub account and so what I'm currently building is a frontend to that
which can resist spam and makes it straightforward for normies, non-github users to propose a new
location to be added to the map which goes through a review process, some automated some
manual and then that turns into a pull request against the repository and so by appearing
as a pull request it's just like any other change in amongst other people's edits to
the code or edits to the database as it were and they get reviewed and someone can see
a preview of that as well and merge it if it's a data that is desirable in the database.
I say database, a bunch of mark down files on disk is a database in my mind.
I'm thinking about one of the projects I was looking at a little while ago is that translation
of 650 game strings in 56 languages, that started out as a Google sheet, you've got
a GitHub repo but that doesn't stop anyone from turning the data that's in there into
other formats, right?
You could turn it into a SQL database or whatever you wanted it to be because the data's
open.
Yeah, I guess you could take that data and do something different with it.
What I would really like people to just contribute to this and yeah, obviously someone might
come up with a better way to render this information, which will be great, especially if
they contributed and improved the website.
There have been some iterative improvements like we didn't have search and then Stuart
contributed a little search option.
It's very much MVP at this point, like you can scroll around the map and you can zoom
in and out and you can zoom straight to a location and there's a few like helper things
in there but it's not finished by any stretch.
The big problem right now is making it easy for people to add locations and edit locations
and by the time you hear this, that may be done but at the moment, as we're talking here,
you have to edit the pages directly on GitHub itself.
You say it's MVP, I mean, how does it compare with the original website when it shut down
in terms of its features?
Well, it's an awful lot uglier for one thing.
It's another thing.
I'm not very good at CSS and if someone wants to design a logo and contribute improvements
to the CSS, so it looks nice on both desktop and mobile, it's okay.
It's usable right now but it could look better and it could respond better in mobile environments
and so on.
If someone wants to contribute that, that would be awesome.
The old site had a nice logo, a nice theme and obviously a larger database of points of
interest.
Right.
The main motivator for this is to not let what happened to the first site happen again.
I would like the data to be crowdsourced just like theirs, but I want the data to be kept
in the open, which is why I put it on GitHub.
I could have used GitLab, I could have used anywhere else, but I just happened to use GitHub.
And yeah, it could move in the future.
That's part of the beauty of it being all like open and built with Hugo because you could
build that anywhere and you could publish it anywhere.
I happen to be using Cloudflare, but it could be published to GitLab pages or it could
be published anywhere else or to your own server.
This is just like the minimum possible nerdy day trips website.
I don't have their original domain, their original domain was nerdydaytrips.com, I grabbed
nerdydaytrips.org because I thought that went better with the idea of crowdsourcing and
it's just like a fun little project.
I've just had an idea, which is because it's Hugo and it's a static site, there'd be
nothing stopping you building the static site and actually having the static site on your
phone.
And it could be usable completely offline, feel free, I want to be doing that.
I do have it on my laptop and it's very easy to develop for like literally you clone
the repository, you know, create a branch and you can just run Hugo serve and you'll end
up with a local copy of the whole thing and it operates exactly as it does remotely.
The only difference is I'm using Font Awesome for some of the icons and because of the
free license I've got with Font Awesome, it only renders those for certain pre-selected
domains.
I think local host is one of them and nerdydaytrips.org is another, but if you published this as is
to a host of your own, that bit probably wouldn't work, that's a full warning and I probably
should put that in the read me, but if anyone wants to contribute to this, I would welcome
it both in design and functionality and the data itself, hopefully by the time you listen
to this, there'll be an opportunity to add whatever nerdy places you like to go to the
database so other people can discover those places.
Linux Matters is part of the Late Night Linux family.
If you enjoy the show, please consider supporting us and the rest of the Late Night Linux team
using the PayPal or Patreon links at linuxmatters.sh slash support.
For $5 a month from Patreon, you can enjoy an add free feed of our show or for $10 get
access to all the Late Night Linux shows add free.
You can get in touch with us via email, show at linuxmatters.sh or chat with other listeners
in our telegram group.
All the details are at linuxmatters.sh slash contact.
I have been pushing my steam back to its limit with a new game.
Do we need to ask?
Oh, yes, sorry, I completely forgot the signs there.
Is it a game with road-like elements, by any chance?
No, it's not.
In this case, games with road-like elements don't tend to be that demanding, do they?
It's like we don't know you anymore, Mark, the last two games that it's just like
no.
So I've been playing a game called Avald, which is a first-person RPG set in Aura, which
is the setting of the Pillars of Eternity games, which I have spoken of with great joy before.
So Pillars of Eternity is an isometric party RPG in the sort of Baldur's Gate tradition,
but this is set in the same world, but a very different kind of game.
And when it was first announced, it was a short trailer showing the first-person view
and someone walking around with a sword and casting spells, and everyone immediately
assumed that this was going to be a sort of Bethesda style RPG, like Skyrim or Fallout.
And there's elements that are similar, but it's not really like that to play at what
I've found really comes to mind when playing, is actually the most recent Tomb Raider games.
So Tomb Raider, Shadow of the Tomb Raider and the other one, which I can't remember
right now.
So there's a lot more focus on exploration and sort of crafting and quests and stuff,
and it's not just one big open world where you can pick anything up and back stuff
with your sword and it goes flying across the room, which confused me at first,
because you know, I'd climb up a bell tower and hit it with my sword and it didn't go
dong and I'd go, well that's a little bit bad.
So is it kind of, when you say it's like the Tomb Raider ones, it's kind of linear
in that like you've got an open area that you can look around in, but you get taken
from area to area throughout the game.
Exactly.
So you've got quests within that area, but you're free to explore and it's worth exploring
because there's always something to find in each area, but it's not, you don't have
puzzles to solve like in Tomb Raider, it's more narrative driven and character driven.
And the environments look amazing, like they're really like graphically detailed, even
if they're not sort of logically detailed in the same way that say Skyrim is.
And there's some points which are obviously very deliberately made for you to find that
you can then see across like the land and the music ramps up and you think, well that's
actually quite impressive, like even on the Steam Deck with all the graphics settings
low, there's still quite an impressive draw distance and you can see a lot of detail
in all of this.
So is that what you're alluding to when you say you push the Steam Deck to its limits?
You had to like dial stuff down because it can't cope with it.
So is it, is it just that it's possible if you've got a bigger GPU, it could do more?
Well, this is the thing is the games rated as playable on the Steam Deck.
So Valve has a rating system where they have I think five criteria, which a game has to
meet for it to be Steam Deck verified and it's rated as playable, but it's rated as
playable because you have to manually invoke the on-screen keyboard when you name your character.
Aside from that, all of the other criteria, including that the default graphics settings
give it, you know, are playable, is playable with the default graphics settings it gives
you, that criteria is met and that's not quite what I found when I played it at first.
Repairing it on the TV in big resolution or handheld.
A bit of both.
I've tried it in both just to see where the limits are and it's certainly better in handheld,
but not perfect and I'm still playing it rendered at 720p, just blown up to a big
telly.
Right.
So it's not like I'm trying to render it in 4K, that would be a tad silly to expect, but
I found a few specific issues which I actually spent quite a bit of time when I started
playing fiddling around with to see if I could work out where it was they were coming
from.
So the first thing I found was that in dark areas, I kept on looking at walls and seeing
this weird sort of black stippling effect all over, which wasn't quite lined up with
the shadows that were being cast, I couldn't work out what was going on.
So I went through flicking every graphical setting to work out what was going on and I
found that if I disabled ray tracing, that went away.
Always turn off ray tracing.
Always turn off ray tracing, right, noted, because yeah, normally it's not on by default
and I don't tend to turn it on.
Right.
But yes, in this case, it was on by default and it was causing this problem.
And once I turned this off and then I saw the shadows where they were meant to be, I
realised that they looked like very jagged polygons, like sawtoothed edges on all of these
on all the shadows.
So I went through and found the settings again and found that I could just about get away
with nudging shadows up to medium.
It didn't have the much of an impact on the frame rate and they looked a bit better.
Then a little way into the game, once you get through the prologue, you find yourself
on a dock like next to some ships and there's ropes along the edge of the dock and then
there's sea, which is very pretty and very nicely animated.
But then I found if you look at the interface between the sea and the thing in front of
the sea, there was this weird, almost like a Jpegar defect all the way along the screen.
So I wasn't very happy with that.
I was finding this was breaking my immersion of it.
So again, I went through and found there's a setting in the graphics, which is just called
effects.
So I not I not stopped the effects and this problem went away.
It wasn't just on or off.
It wasn't a barrier.
No, this is a slider.
This is starts off low and goes up to ultra and whatever.
So I've moved it from low to medium and this problem went away.
So quite happy so far.
The frame rate still decent enough.
When you say decent, like what is the frame rate and is this game that's like frame rate
sensitive?
It plays about 30 frames second fairly solidly, right, which is reasonable for a game of that
type on a steam deck.
And you notice when you come into a new area, when you load up, you see the textures
load everything.
It's not like everything's like, bam, you know, eyes melt with the glory of it.
You've got to set your expectations accordingly.
But once you once everything's loaded and you're playing, it's going at 30 frames a second
quite, quite nicely.
You don't really notice a huge amount of slow down too often.
But there was one remaining problem and that's a weird ghosting effect that I've noticed,
which was most apparent when I mentioned before, it's a, you know, a first person swords
and sorcery type game.
And when you draw your weapon as the blade moves across the screen, you see a sort of ghost
of it.
And the best way I can describe this is imagine you're looking at a picture on an e-reader
and then you change the page to another picture and it leaves a few dots behind of the previous
picture.
But you're seeing this in real time on a rendered game or the other thing that it really
makes you think of is if you ever played like an early graphical adventure like Mist,
when they wanted to, you know, you pick up an item off the background and they wanted
to show you fading out, but they didn't really have transparency effects then.
So things would kind of dissolve.
Oh, like a stipple effect.
Exactly.
You get this same effect, which yeah, it kind of gives me this really weird nostalgia for
like all the white dolls in early Windows games.
But within this amazingly beautiful 3D environment, which is quite strange.
Do you know if this is because you're running it on a steam deck, which is lower power
or is it Linux or is it the driver's, is there any indication from other people's support
conversations?
What could be causing these kind of things?
Yes, I did find a promising thread, which was discussing this.
And some people, you know, one person saying, you, I've got it.
It's definitely not this, this or this.
It's definitely not a, you know, the drivers, the hardware, it's just, it happens.
And other people say, well, I don't see this.
And what they narrowed it down to was a feature of the Unreal 5 engine on which this game
is built called temporal anti aliasing.
Yeah.
And essentially what's going on here is it's inventing additional frames of animation.
Right.
So I suspect if I was playing on something which could deliver a higher frame rate, and
then I set a frame rate limit of what I was actually seeing on my telly.
So say it could, you know, I could pay it at 30 or 60 or whatever.
Then I probably wouldn't see this, but because I'm playing it on, you know, a reasonably limited
device, it's giving me a smooth experience.
But the payback is it's not actually rendering all of the frames I'm seeing.
It's making some of them up as it goes along.
And in some cases, not doing a good job of it has others.
Unfortunately, there isn't any option to turn this off in the graphics settings and seeing
if that makes a difference.
It's kind of baked in.
Yeah.
I was just looking it up, actually.
And I found the Unreal documentation where it tells you how to turn it off in the editor
if you're the game publisher, but not like, I couldn't see a command line option which
would be ideal.
Sad.
Yeah.
I wonder if there's an any file in the game data somewhere you can go and twiddle that.
Yeah.
So this is an interesting discovery.
And you know, it's something I've got to, I've just decided that I'm going to live with
this.
It's slightly annoying, but really, I'm enjoying the game and it doesn't pull me out
of it too much that it would upset me.
What it has led me to wonder though is are we sort of starting to hit the point where
modern games aren't going to be able to target the Steam Deck as a sort of lowest common denominator
anymore because I think, you know, Valve's obviously always trying to get people to be
Steam Deck verified and setting these criteria.
You don't have to be able to play it with any particular level of graphical fidelity.
You just have to have a setting that it will pick up by default on the Steam Deck where
it works.
But if it's going to have issues like this and if that's going to cause consternation
with gamers, maybe game developers might get to the point where they say, well, there's
no point because we can't get it running nicely enough that we don't end up with a load
of complaints about the game.
And will we reach the point where Valve can come out with a new revision before that happens?
I think you might be surprised at how much effort developers are prepared to put into
getting a game running on smaller devices because the Steam Deck has a massive footprint
now of like users as a console.
And the efforts that some of the developers have gone to for sort of big grand 3D games
have gone to to port it to the Switch are just astonishing.
And the things that they've had to profile and debug and improve in order to get their
games on the Switch because there's an audience there.
I think there's probably enough people now with Steam Deck's that there's enough motivation
to make it work well.
I'm looking forward to all these YouTubers who do low-end gaming on like old Pentium
fours.
In a few years, they'll be doing low-end gaming, right?
Steam decks.
Oh, look at this arcane device that can't do modern gaming, but hey, I can play Doom on
it.
Yes.
So when I play my next game, I'll report back and let you know if I'm still having ghosting
issues.
You have been listening to Hacker Public Radio at Hacker Public Radio does work.
Today's show was contributed by a HBR listening like yourself.
If you ever thought of recording podcasts, click on our contribute link to find out how
easy it really is.
Hosting for HBR has been kindly provided by Anonsthost.com, the Internet Archive and
R-Sync.net.
On the Sadois status, today's show is released under Creative Commons, Attribution, 4.0
International License.