Thankfully, there's FreeBSD documentation that explains how to fix that. You can generally follow the steps explained there about virtual_oss, but it boils down to:

pkg install virtual_oss
service virtual_oss enable
service virtual_oss start

Then you have to determine your sound devices:

me@fbdesktop:~ % cat /dev/sndstat
Installed devices:
pcm0: <Realtek ALC887 (Rear Analog 5.1/2.0)> (play/rec)
pcm1: <Realtek ALC887 (Front Analog)> (play/rec)
pcm2: <Realtek ALC887 (Rear Digital)> (play)
pcm3: <Intel Cougar Point (HDMI/DP 8ch)> (play)
pcm4: <Intel Cougar Point (HDMI/DP 8ch)> (play)
pcm5: <Intel Cougar Point (HDMI/DP 8ch)> (play)
pcm6: <XiiSound Technology Corporation GG-HS420W> (play/rec) default
Installed devices from userspace:
dsp: <Virtual OSS> (play/rec)

So in my case I need pcm0 (rear output) and pcm6 (my wireless headphones). To actually switch audio output you need one of these commands:

virtual_oss_cmd /dev/dsp.ctl -f /dev/dsp0
virtual_oss_cmd /dev/dsp.ctl -f /dev/dsp6

I have to point out that documentation uses /dev/vdsp.ctl file which doesn't exist on my system.

But these commands require root privileges, so I added this to my doas.conf:

permit nopass :wheel as root cmd /usr/local/sbin/virtual_oss_cmd

...Aaaand Firefox still doesn't care about this. But we can fix this too.

If you use Firefox or any of its fork, you should go to about:config and look for media.cubeb.backend parameter. If it doesn't exist (it probably doesn't), you should create it (String parameter) with the value "oss" (without quotes). It basically tells Firefox to use OSS as a backend. mpv, qutebrowser just work without any tinkering.

With these simple steps my audio now switches on the fly, no need to restart an application or pause/unpause.

You can write a simple script using dmenu or rofi or create a hotkey (depending on your preferred WM/DE) to switch audio outputs.

Accordingly, now every web service in existence brags how private it is. Reality, of course, is merciless.

Facebook faces $101 million fine for storing passwords in plaintext. In 2023 hackers stole Microsoft Azure Active Directory certificate which basically gave them access to all Microsoft cloud services: Outlook, Office, SharePoint, Teams, etc. etc. You can say that the whole Microsoft has been compromised. They don't even disclose the details — this might tell you how bad things are.

Trusting Microsoft, Facebook, Google and any other «well-meaning» corporation with all your data is, of course, not for the faint of heart. But there are privacy-conscious services, you might say. For example, Protonmail (and all its different services: Calendar, Drive, VPN, even AI chatbot), Tutanota, Signal, Telegram (really?) and so on. They are protected by the pinky promise and local laws.

That's basically what we got in terms of guarantees of privacy: pinky promises and local laws (of Switzerland, Germany and some other European countries).

Pinky promises

What's the difference between promises of corposhell of a human being from Microsoft and CERN physicists that build Proton services? I think there is none. You can't peruse the code, you can't be sure that servers run EXACTLY THAT code, and more importantly, you can't do anything against bugs and vulnerabilities.

Qubes OS has a very good principle: «distrust infrastructure». When you apply it everywhere, life gets easier. You don't have to trust anyone, you know every system you do not control is compromised by default. My own desktop is my fortress, it well might be compromised too, but I trust it somewhat more than anything else. If I encrypt data on my own desktop with my own key, I deem them somewhat protected. Everything that went to the Internet I consider compromised and untrustworthy.

It means I can use any service I want, I just have to know which data I want to share with the whole world. I don't need your privacy measures, I have no way to ascertain they are real. Hasn't Facebook undergone strict security audits that found nothing wrong?

Local laws

Local laws that protect user data is a great thing. The problem with laws is that they aren't Twelve Tables of Rome, they aren't set in stone. They may change any time. And they actually change.

Proton was protected by strict laws of Switzerland but now it moves its services from Switzerland, and their chatbot Lumo is the first to move. They move to Germany for now.

But when ChatControl proposal will pass (not if, but when) — well, they'll have to move somewhere else or it's a Game Over situation. Check their canary often — but I think they don't have one, only Transparency Report.

Privacy services are targets

Services that brag about their privacy (real or fake, it doesn't matter) quickly become targets for governments that aren't interested in losing control over people's lives, so they quickly become banned or restricted.

We have many examples like that: VPN technology that was misused to death so that most VPN services are now banned in many countries (and turned into huge data mining machines), «private» email services like Protonmail (6378 legal orders in 2023, 11023 in 2024), Telegram that was first banned in Russia, then unbanned, and now it's on a verge of ban in the West (you gotta keep up with who's the enemy of which state, but Oceania had always been at war with Eurasia).

What should be the conclusion of this quite chaotic line of thought? I think we should look at the things simpler.

Let's not upload to the Internet things we wouldn't show to any random stranger.

Let's create, not encrypt.

And let us take care of our garden.

I wrote a post on how to use XSLT to style RSS feed.

Articles

• Google is killing the open web

This is an extremely interesting article about the war that Google waged against open standards, particularly against XSLT. Also this is a links galore, huge amount of material with meticulously expounded history. Cannot recommend enough.

• AI is a Mass-Delusion Event

Ruminations of Charlie Warzel about AI. Worth to read as a demonstration of «sobering» amidst the mass-media hype about Word Prediction Machines.

• Emacs Elevator Pitch

It's just hilarious. Imagine you explain Emacs to a total stranger in the lift.

On the other hand, there's quite an interesting links to things you can do with Emacs. I bet list of things you can't do with Emacs would be much shorter.

• The Advantages of Text-Based Information Versus Videos, Audio or Images

Quite interesting thoughts about digital dark age and the role of text as still the best medium.

Smallweb

• Michael Rawdon's Home Page

This is one of the surprise finds you can stumble upon when you use Wiby's «Surprise me» feature. What is this? It's just a personal website, remember them? With its last updates in 2006, it's still built with <table> instead of <div>.

Why all this is cool? Because of small and pleasant discoveries.

From this page we can find out that Michael Rawdon likes Sci-Fi, comics and role-playing game. He's been into Magic: The Gathering for quite some time.

There's a lot of reviews for comic books, games, books and music. And a lot of links to other pages, including homepages of his friends.

If you read his journal you can find out that on April 30, 2001 he had a date with Debbi he hadn't seen since high school.

And you know what? Looks like they had it great. Journal ends in 2006, though...

But wait a minute. Since then Michael moved his website to FascinationPlace.

I read there an interesting review on James Gunn's «Superman» movie there. After all, Michael is a big comics fan (unlike me).

And I'll have you know, he's still with Debbi.

They got married in 2015.

Based on my almost one day of experience in XSLT I can say that XSLT is basically a technology of XML transformation. But for the purpose of this post we'll call it a CSS for your RSS feed.

XSLT lets you define what do you want to do with XML data (any RSS feed is literally an XML file) and how do you want to transform them.

Why style RSS feed?

But why would you want to style RSS feed? Nobody sees it as it is, everyone uses some kind of reader, right? Well, yes, but actually no.

Big corporations do what they can to fight the open web, so most popular browsers conveniently forget what is RSS. For example, Firefox forgot what is RSS back in 2018. So if you try to open RSS feed in Firefox now you'll see this:

Actually, it's not even that bad. At least, Firefox shows you the file structure instead of dumping raw XML on you.

Some browsers, like Vivaldi, actually support RSS (Vivaldi even has its own RSS reader). Vivaldi also can style RSS feed without any XSLT. Here's how raw RSS feed looks like in Vivaldi:

Looks pretty neat even without any XSLT.

How to style RSS feed

To style RSS feed we need to prepare a custom file with extension xsl, say, rss.xsl. Then we need to edit our RSS feed to tell our browser where to look for this file. Like that:

<?xml-stylesheet href="rss.xsl" type="text/xsl"?>

We should put it in the beginning of our XML, so it would look like that:

<?xml version="1.0" encoding="UTF-8" ?>
<?xml-stylesheet href="rss.xsl" type="text/xsl"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
... etc, etc.

Styling the feed

The actual XSLT file should have structure similar to this:

<xsl:stylesheet version="3.0">
  <xsl:output method="html" version="1.0" encoding="UTF-8" indent="yes"/>
  <xsl:template match="/">
    <html lang="en">
      ...
    </html>
  </xsl:template>
</xsl:stylesheet>

Inside the <html> tag you can place any HTML or CSS code you like. You can even use JavaScript, but please, don't do that, have mercy!

The simplest XSLT magic might happen like that:

<?xml version="1.0" encoding="utf-8"?>
<xsl:stylesheet version="3.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
                xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/"
                xmlns:itunes="http://www.itunes.com/dtds/podcast-1.0.dtd">
  <xsl:output method="html" version="1.0" encoding="UTF-8" indent="yes"/>
  <xsl:template match="/">
    <html xmlns="http://www.w3.org/1999/xhtml" lang="en">
        <head>
            <title><xsl:value-of select="/rss/channel/title"/> Web Feed</title>
            <meta http-equiv="Content-Type" content="text/html; charset=utf-8"/>
        </head>

        <body>
            <xsl:for-each select="/rss/channel/item">
                <p><xsl:value-of select="title"/></p>
            </xsl:for-each>
        </body>
    </html>
  </xsl:template>
</xsl:stylesheet>

Not particularly impressing, but it works.

Our point of interest is inside <body> tag where we use XSLT to walk through every <item> tag inside RSS file and display it with simple <p> tag.

It's not extremely useful but should give you the idea how to display items from RSS feed. But let's turn them into actual links and prettify a little. Let's rewrite the contents of a <body> tag like that:

<h1>All posts</h1>
<xsl:for-each select="/rss/channel/item">
 <section>
  <h3>
   <a target="_blank">
     <xsl:attribute name="href">
             <xsl:value-of select="link"/>
     </xsl:attribute>
     <xsl:value-of select="title"/>
   </a>
  </h3>
  <p id="pubdate"><xsl:value-of select="pubDate"/></p>
  <hr/>
 </section>
</xsl:for-each>

Now this is something actually usable.

You can take it from here. Add a <style> tag and unleash your creativity. Add more info on the page, for example, explain what this page means and how to use RSS feeds for users who aren't actually familiar with the technology.

How to add XSL support to BSSG

I use Stefano Marinelli's amazing BSSG static site generator to build this site. As far as I know, it doesn't support XSLT out of the box, but it's written in bash, so we can easily add it.

Just open scripts/build/generate_feeds.sh file around the line 86 and add a line about XSLT file. Like that:

# Create the RSS feed header
cat > "$output_file" << EOF
<?xml version="1.0" encoding="UTF-8" ?>
<?xml-stylesheet href="rss.xsl" type="text/xsl"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">

That's it.

Here's how my RSS feed looks like now. You can just open it in your browser, but I'll post here a screenshot because when you start playing with CSS, it's hard to stop. Who knows how it would look in the future!

Remember the step about unleashing your creativity? Yeah, I skipped that...

Credits

This post is built upon Ricardo Lopes' post «Styling Your RSS For a Better Web» which in turn is inspired by:

• Pretty RSS Feeds by Brandon Rozek, which was inspired by
• pretty-feed.xsl by Matt Webb, which in turn was inspired by
• How to style RSS feed by lepture.

I don't know a thing about XSLT and still managed to build half-decent RSS feed with just the vim and a couple of good articles.

Do better.

Create better Web.

This week was quite busy: a lot of work, a lot of texts, a lot of reading. Quite a cliche, but I'm reading Wilhelm Wegner's «Rome» (so yeah, guys are always thinking about Roman empire, I guess): huge tome that is hard to get away from.

I've spent quite some time with Dillo — it's an interesting web browser for The Olden Web. I've been playing with its plugin interface but in the process I broke something and still can't figure out what. Now some plugins aren't working and Dillo can't connect to its plugin server (it's a separate process). Even Bookmarks aren't working. That's truly strange.

Anyways, there's been a lot of tidying up on my desktop and not a lot of time to surf the Web. But still I stumbled upon some interesting things.

Articles

• Emacs Writing Experience

Randy Ridenour shared his story on how he came to writing in Emacs and why there is NO such a thing as «Emacs writing experience». I like reading such personal stories about adopting software.

• Why You Shouldn’t Use “Markdown” for Documentation

Eric Holscher explains why you shouldn't use Markdown for documentation. That's quite an old text from 2016 but the points are still valid. That being said I personally never thought Markdown's problematic areas as such, and I use Markdown for translating books. Still, it was an interesting reading.

Indieweb

• TuM'Fatig

That's definitely one of the best sources for everything Slackware and OpenBSD.

• is OpenBSD 10x faster than Linux?

No, it's not, but if you carefully choose your tests it can be. Quite an interesting and funny post.

I have tested it myself and OpenBSD is faster on my weak laptop with 1.6GHz CPU than Linux desktop with 3.2 GHz CPU. There's a link to explanation in the article.

Movies

• Freud's Last Session (2024)

This movie set in September 1939 (re)creates (somewhat) imaginary meeting of dying Dr. Siegmund Freud with the very Clive Staples Lewis. They meet when the world is plunged into chaos of World War II. Darkness comes, and two gentlemen, dying and young, argue about existence of God.

I wouldn't call this movie a revelation or life-changing experience. It's just a good movie that is illuminated by brilliant performance of Anthony Hopkins (Dr. Siegmund Freud) and Matthew Goode (C. S. Lewis; good Lord, Matthew Goode is a shapeshifter — just compare his rendering of Lewis with his role in an amazing Netflix police drama Dept. Q!) Also, I would mention that it's a movie that is a feast for an eye — contemporary interiors are recreated with such meticulousness that you just want to move there and live for the rest of your life.

That's it for this week, I guess!

Not only that, most of the Internet now is just an app on your smartphone. You can't control it, you can't do anything about it.

The Internet becomes insanely complicated and unpredictable place.

In this post I try to outline some of my thoughts about a state of Digital Doomsday, why it could happen and what should we do about it.

What is a Digital Doomsday?

Digital Doomsday is a state when the Internet is no more. There could be different scenarios why we can lose the Internet (or access to the Internet). Let's review some of them starting with the least plausible.

• Don't Look Up!

Imagine huge asteroid falling on Earth and bringing death and destruction. Big enough rocks destroy almost all life on Earth quite regularly (in a geological perspective, of course). Although we usually think it's quite unlikely that a big rock size of Manhattan shall fall on Earth in the next Tuesday.

• How We Got Along After the Bomb

Not an impossible scenario, not at all! Right now Doomsday Clock shows 89 seconds to midnight. FYI, only 10 years age we were 3 minutes from midnight. If you watch news at least once a month you can see an ominous pattern: not a lot of peacemaking, quite a lot of warmongering.

• Sabotage of submarine communication cables

Optic cables that lie on a seabed are extremely vulnerable. Just look at this map of SCC's choke points:

Source: Securitization of Physical Cyberspace Infrastructure as a Nexus in U.S.-Russia Relations: The Case of Submarine Communications Cables

They are vulnerable.

Modern Internet is too centralized, too dependent on CDNs to survive a large scale attack on SCCs done right. There are no good guys in this game: any country or major geopolitical force is capable of doing that.

How possible is this? I have no idea. But looking at the last ten years, there's been a lot of things I had no idea about, and lo! they happen like the most obvious thing and nobody cares.

• Toxic Internet

All of the above scenarios, plausible or not, presuppose that you wouldn't be able to use the Internet, because it is no more (on a more or less global scale). But what if you wouldn't WANT to use the Internet? What if the Internet would be too hostile or too dangerous a place? In other words, the Internet may become toxic.

Toxic Interwastelands

Do you remember «The Right to Read» by Richard Stallman? It doesn't matter what you think about him personally, this story is brilliant. It was published in 1997 and having read it circa 2004 I thought to myself: «Ok, it was funny, but barely plausible». Little did I know back then. Little do I know now.

Nowadays the right to read is being slowly revoked. If you use modern Amazon Kindle for some reason, you don't own your ebooks anymore — you only license them. Also, publishers can modify and censor the books you thought you own following some current cultural and political agenda.

The Internet is just four sites, and they're not even sites anymore, they're generally an apps on your phone. These apps are a walled gardens, you can't even access public info posted there if you have no account (why don't you? It's free and it only takes a minute!)

All-pervasive AI follows you everywhere trying to replace you in your everyday tasks (and failing miserably).

In UK you can't even watch porn without confirming your real identity (if it isn't a reason enough to stop watching porn, what is?) Other countries watch how it goes with sincere interest, no doubt. European age verification app might actually ban any Android phone not licensed by Google.

Microsoft designed a Pluton to further «enhance security» by storing your encryption keys and other sensitive data in a black box implemented right inside your CPU. It's for your own good, obviously (actually, it's not the first time they're doing something like this, they have already tried the same trick back in 2002 with Microsoft Palladium, but then they failed).

And while we're at it, good ol' Microsoft also screenshots everything you do on your computer and sends it somewhere on the Internet, including passwords and credit card data.

In many countries you HAVE TO own and use a smartphone to actually exist and use public and governmental services. In my country most banks don't even have any web interface whatsoever — only an app that just wouldn't work on your jailbroken Android device.

This is what we have NOW. It's not a dystopia, not a Stallman's fantasy from almost 30 years ago, not a spooky doom-and-gloom fib, it's a reality. And we obviously ask ourselves what will happen tomorrow?

Honestly, I have no idea. But there is a chance that at some point you might realize that you just can't use the Internet anymore because everything you have is several apps filled with mindless and mind-numbing «funny» videos while everything else is forbidden. Net neutrality is already shaky as hell. Sooner or later your Internet may become a feudal landscape, only with TikTok and fast pizza delivery services.

May be at some point you'd think you had enough.

Are there any silver linings?

Of course, there are.

Right now we have a lot of open-source OSes, we have sane initiatives to lessen the dependency on huge tech companies with proprietary tools, we have Fediverse, we even have local and open-source AI.

Things look bright, don't they?

Fediverse stats aren't promising, to be honest...

No matter how bright are the skies, we still need plan B.

We need to develop and invent new ways to bring people together (be it Fediverse or some obscure privacy-first mesh networks). But we also need to figure out how to survive on our own and grow our own vegetables (figuratively and literally).

Which means we have to plan our digital prepping.

As I see the problem, we need the data to survive (and quite possibly to rebuild the civilization) and tools for general computing.

What about the data?

Books, articles, knowledge bases, you name it. There are great projects that can help you to build your own «offline data center», like Kiwix.

With Kiwix you can download the whole Wikipedia (and not only that) to your device. Kiwix works pretty much on any OS and works fine.

We also need physical books for redundancy and ease of use (and physical media in general — for example, movies on DVD and BluRay, and music on CDs).

I'd recommend to check out thrift book stores close to you or on eBay. There are also places like Abebooks where you can buy precious books like «Programming Perl» or «Advanced Programming in the UNIX Environment» for a couple of bucks.

What about the tools?

We need a serious and reliable OS on our lonely journey, so it won't be Windows or MacOS, obviously. Back in the days Windows would be number one choice because it worked everywhere, had a huge software library and supported everything you can throw at it. But now even Windows 10 won't work properly without an SSD, not to mention RAM and CPU requirements. And actually modern Windows is more like a surveillance machine with extra features like Notepad and Calculator. So Windows is out of the question.

Let's try to outline some criteria for our OS (and software) of choice.

1. It should be open-source.
2. It should be easy to maintain.
3. It should be possible to work on its source code (because there probably wouldn't be anyone to do that except yourself).
4. It should have low system requirements and support most of the hardware.
5. It should have a lot of software available.
6. It should have a good documentation (and/or porter's and developer's guides).

So that kinda narrows it down a bit. Basically we have to exclude all rolling Linux distros, overbloated distros and stick with LTS releases.

Below are some of OSes that I'd choose. It's not an ideal and only choice you should have. It's just what I would choose. For example, there are BSDs like NetBSD or DragonflyBSD that I didn't put on the list because I just don't know enough about them.

Slackware

Slackware has a great reputation as one of the oldest maintained distros out there (actually, one of the oldest distros ever). It comes with a huge variety of software out of the box and even has 32-bit version (as of version 15.0).

Slackware uses simple init system based on rc-files and generally sticks to the KISS principle and «user knows better». It always does what it's told and doesn't try to overrule or restrain you.

Out of all Linux distributions Slackware is considered the most BSD-like.

Generally, even on an island in an ocean Slackware is more than enough to get you started: it comes with all the software you might need. One of the important Slackware «features» is the fact that there's no dev-packages. Every software package comes with all required development libraries, headers and docs, so it's very easy to compile new software for Slackware: you probably already have all the dependencies.

And while we're at it, Slackware doesn't do dependency checking, but there are a lot of community tools like sbotools that can do that for you.

Debian

Debian doesn't require any special introduction: it's a very old distribution that values security and stability. In some areas it might even be considered «a vanilla Linux», because it's so popular, simple and heavily documented.

One of the selling points of Debian has always been a huge number of software packages available. Sadly, in recent years Debian does not provide ready DVD/CD images for this software, you have to build them yourself using a jigdo utility, but that's hardly a problem when you prepare for a harder times. Most importantly, you can build «source» images that contain all the necessary source code.

FreeBSD

FreeBSD has so many advantages that I'll just outline them briefly and leave a link to a great article by the famous vermaden «Quare FreeBSD»: complete system (unlike Linux which is just a kernel plus a bunch of packages), simplicity, documentation, init system and, most importantly in our case: ZFS.

ZFS is a survival grade file system that does everything you might need from a file system: self-healing, snapshots, reliability, sending/receiving datasets over the network, even Boot Environments (basically, it makes really hard to kill a FreeBSD). ZFS is a first-class file system in FreeBSD, unlike Linux that has incompatible license and doesn't have some great ZFS features like Boot Environments (also, doing ZFS-on-root is still a huge lottery on Linux).

And in a situation like Digital Doomsday you REALLY need a good file system.

OpenBSD

I love OpenBSD for its extreme simplicity and straightforwardness. I don't know how they do it every time, but OpenBSD developers somehow manage to solve problems as they are instead of creating a whole lot of abstractions and generalized solutions: if you need to connect to Wi-Fi, you just do ifconfig; if you need to mount a drive, you do just that.

OpenBSD is a secure and robust operating system that comes as a whole package. The only serious downside of OpenBSD (for me) is lack of a good file system.

FFS2 is a «good-enough» file system, but don't raise your hopes: it works, but it doesn't come even close to ZFS in any respect.

Solène Rapenne in her article «What if Internet stops? How to rebuild an offline federated infrastructure using OpenBSD» brings up some good points why OpenBSD might be an OS of choice to rebuild the Internet.

It's pretty easy to organize your local repository for OpenBSD. For example, all of the amd64 packages for OpenBSD 7.7 take only ~80 Gb in size, so its quite manageable. But, as Solène mentions, OpenBSD ports distfiles is what we really need.

So, if we try to assert how many disk space we need for all ports on OpenBSD 7.7, we'll get this picture:

me@desktop:~/ports$ find . -name "distinfo" -type f -exec grep "SIZE" {} \; | sort -u | awk '{total+=$4}; END {print total}'
118157658347

If my unix-fu isn't terribly wrong, it means around 118 Gb. Again, not that much, and we'll have all the source code we could possibly need along with OpenBSD-specific patches.

Some concluding thoughts

Some people might say that it's a some kind of defeatism and we should rally up and fight for our digital freedom instead. That might or might not be true: not everyone is inclined to activism in any shape or form.

But anyway I think that having a last line of defense in form of just your personal computer that is capable to function fully autonomously is a necessary step. Because if you don't have this last line of defense already prepared, what will you do when all other lines would fall?

Self-hosting was just a phase one.

Prepare for Digital Doomsday, for it is nigh.

The idea of a challenge is to use some old hardware, software or technology to find out what you can do with it nowadays. Practice shows that there are a lot of things you can do. Just look at the initiatives like EndOf10 that incentivize people to use free and open source operating systems instead of throwing away perfectly fine hardware just because Microsoft said so.

This year Old Computer Challenge is «There is no challenge». So you're basically free to use whatever you have or want.

What I got

For this challenge I decided to dig up and repair my first PC. It wasn't a monster machine even in 2003, but it served me well for almost ten years. Here are the specs:

• CPU: AMD Ahtlon XP 2000+ (1.6 Ghz, single-core, i386, Anno Domini 2002)
• RAM: 640 Mb of DDR2
• HDD: Some old 80 Gb Western Digital drive from ~2004

Some capacitors on the motherboard, of course, gloriously went to Valhalla some time ago, but all in all it was a perfectly fine machine if only a little dusty.

I tested several OSes on it, including those that I owned but never had a chance to test: Suse 10.1, Mandriva 2006, ASPLinux 10. But in the end I decided to test OpenBSD. And because we are in the world of Normal Operating Systems, I used the latest release, OpenBSD 7.7.

First challenge: how do we boot?

Well, that's not a simple question. This PC has USB ports, but BIOS doesn't know how to boot from them. The only available options were CDROM and floppy drive.

Thankfully, there is such a great tool as Plop Boot Manager. It does one simple thing: you put it on a floppy drive or CD/DVD, boot from it and then choose what do you want to boot next (HDD, USB, etc.) And it's only 500 Kb!

OpenBSD 7.7 on a 23-year old machine

This might sound ridiculous, but the most recent OpenBSD works just fine on a 23-year old machine. Right after installation without any tinkering and tuning it boots in 88 seconds up to xenodm. I think that's great. The mere fact that the most recent and complete modern OS just easily boots on an almost quarter of a century old hardware deserves a shout out in praise of OpenBSD developers. Most modern Linux distributions do not even provide 32-bit installation options.

Sadly, after some use it turned out that most of my RAM is ridden with bad blocks that caused hard crashes. I pulled RAM out leaving only 256 Mb.

And you know what? OpenBSD still worked just fine.

It doesn't care. It just works.

There is no challenge

Funny thing is, I didn't even needed to delve into this technonecromancy because I do Old Computer Challenge every day for many years.

My own machine that I use every day for all my work (i.e. it basically pays my bills) qualifies for an OCC, because it's a 2011 machine with a pretty good 3rd generation i7 CPU from 2012 and some 8 Gb of DDR3 RAM. All PCI-E ports are fried, so I can't even use a dedicated GPU, getting by with CPU videocore. And I'm fully content with it. It does everything I need.

It's not that I don't have money to build some new and modern machine. Actually I did some time ago, but just gave it away to someone who needed it more. It's just... this one works fine, why do I need more?

There is no challenge.