IBM Plex Mono is one of my favourite monospace fonts to have come out in the last decade. And it got an update in May:
– stylistic alternate for /l in Roman styles
– stylistic alternates for /comma and quotes
– glyph /overline (U+203E)
– support for Venda language (10 additional characters)
– support for Arabic transliteration (10 additional characters)
– support for Hebrew transliteration (1 additional character)
– support for Sanskrit transliteration (18 additional characters)
– combining marks /macronbelowcomb (U+0331) /circumflexbelowcomb (U+032D) /brevebelowcomb (U+032E)
– ASCII control pictures (U+2400-U+2421)
I’m one of those weird people who use different fonts in different applications. IBM Plex Mono is what I use in Kate.
ZFS includes inline compression, meaning files are compressed and decompressed on the fly. While you might expect this to reduce IO performance, it’s been proven to improve throughput by reducing the amount of expensive reads required, especially on older spinning drives.
Inline compression is wonderful. I don’t need to rotate logs to compressed files. I can temporarily expand virtual disks to raw images without taking up the entire maximum space of the disk. It’s also great for thwarting zipbomb attacks!
Save for highly-compressed files like video, I’ve always left the compression algorithm as the default, which for years has been lz4. But I have a few specific datasets for which more aggressive compression can be done. For example, I have one dataset containing my collection of reference PDFs and ebooks I rarely access, and have large blocks of text that are great candidates for lossless compression.
I interrogated this dataset holding my books with the requisite zfs commands:
# zfs get compression zten/books
==> NAME PROPERTY VALUE SOURCE
==> zten/books compression lz4 inherited from zten
# zfs list
==> zten/books 42.7G 601G 42.7G /zten/books
So about 42.7 GB of books compressed with lz4.
I created a new dataset, this time with the ubiquitous gzip and its default compression level:
# zfs create -o compression=gzip zten/books-gzip
Then transferred the same files across, and ran the same commands:
# zfs get compression zten/books
==> NAME PROPERTY VALUE SOURCE
==> zten/books-gzip compression gzip local
# zfs list
==> NAME USED AVAIL REFER MOUNTPOINT
==> zten/books-gzip 41.6G 559G 41.6G /zten/books-gzip
We’ve saved about 1.1 GB of space, though the files can still be accessed in exactly the same way. Neat.
Next, I thought I’d try zstd, or ZStandard:
# zfs create -o compression=zstd zten/books-zstd
Then transferred the files across, and ran the same commands:
# zfs get compression zten/books
==> NAME PROPERTY VALUE SOURCE
==> zten/books-zstd compression zstd local
# zfs list
==> NAME USED AVAIL REFER MOUNTPOINT
==> zten/books-zstd 42.7G 516G 42.7G /zten/books-zstd
Interesting, we’re back to where we started with lz4.
That’s the post everyone. Tune in next week when I compare different coloured wall paints. This one is a slightly different shade of beige compared to this one.
☕︎ ☕︎ ☕︎
Compression is always a balance between performance and space (and quality, in the case of lossy systems). How much space you save using a specific algorithm or tool has to be weighed against its practical use. It’s no good having wildly high compression ratios if it takes minutes to load.
In my experience, it’s hard to go wrong with any of ZFS’s compression options. gzip does give better compression, and the difference between loading a PDF from a dataset using gzip over lz4 was imperceptible to me. But then, I’d expect a more demanding workload to render this difference more stark.
In our current world of “AI” vapourware and rising prices, it’s also been interesting to see a resurgent interest in compression (and deduplication) again, as we all try and make better use of what we have. Or at least, get more storage with as little negative impact as possible. ZFS to the rescue, again :).
I’m a day late for a Music Monday, but I had to share this!
Drumeo is such a fun channel for learning drums, and the drum theory behind popular songs. My favourite series involves the team playing the melody and vocals for a drummer, then asking them to produce a drum line. The twist is the drummer has never heard the song before, and often its a genre in which the drummer has no familiarly.
I only just got around to watching Kristina Rybalchenko’s video from seven months ago, but hers might be my new favourite. Her crash symbol grooves were sublime, as was that frantic ending. She nailed it!
I don’t know how else to put this, but I’ll never not be impressed by the inferential abilities of creative people. My beautiful late uncle Dave used to be able to render entire songs on his trumpet having only heard it once, and it always blew my mind. Being able to intuit what’s harmonious and wonderful without needing a score or specific instruction is real talent… and sadly, one that wasn’t hereditary!
I first learned of Om through his news site GigaOm, and his appearances on the Cranky Geeks show back in the day. I remember being surprised at the civil comments GigaOm stories attracted, which I attributed to the quality of the journalism and lack of clickbait. In later years I especially followed him for his photography.
I was lucky to have a few conversations with him on Twitter too back in the day. One in particular I remember was scheming with him to teach his American friends about cricket. I even came close to meeting him when I briefly worked in San Francisco in the late 2010s, but a work issue pulled me away.
Last Friday I talked about a flaky network switch that was wreaking havoc on our wired home network. A few of you sent thoughts, which I wanted to share.
Why wired?
Jeremy, whom I’ve asked if he has a blog I can link to:
You said you have Wifi networks. I’m curious why you use wired ethernet at all.
If a device supports wired Ethernet, I’d rather plug it in. Maybe I’m old school like that. It also keeps the Wi-Fi spectrum free for devices that need it.
Our Token Ring network also wouldn’t support Ethernet Wi-Fi (at least, not without some fancy encapsulation). Tokens famously hit walls rather than pass through them. If someone ever develops a wireless MAU though, I’m all ears.
I had similar issues once with an 8-port metal TP-link dumb switch in my house. It turned out to be the fault of the 12V 1A power supply it was plugged into. That was 10 years ago and it has been working fine since then, no packet drops or oddities observed (other than every few months having to jiggle my bad 8P8C crimps elsewhere to get gigabit negotiated, woops).
(The less said about my crimping abilities, the better).
The traditional gigabit metal dumb TP-link switches have insane MTBFs in their datasheets (more than decades IIRC?) but they don’t seem to give any lifetimes for the smarter models like your TL-SG116E. I can’t assume the switch chips inside are the same since they need to be more programmable, but I thought my half-data-point could still have a chance of being useful to you.
Alternatively, you may have already poured cooking oil into all of the jacks >:D
That’s something I hadn’t considered. I have so many spare 12V bricks, I might give it a try on an area of the house that isn’t “prod”!
Ghost issues
Hauke, from the beautiful Seeheim:
I can so relate. Still trying to figure out why Retrospect backups
(and bulk tcp transfers in general) keep stalling on my Quadra 650,
requiring a hard reset.
And I loved the “Heinzelmännchen”.
My dad is also from Hesse, not Cologne where our beloved Heinzelmännchen come from. But they was a fixture of my childhood; he would claim they were responsible for fixing things in the night while we slept. I should invoke them more often :).
Snow Leopard was a version of Mac OS X released in 2010. It was billed by Apple as having “0 new features”, with a focus on refining the system, improving efficiency, and fixing bugs. Those of us using Macs at the time welcomed the news after living with updates that seemingly broke as many things as they introduced with each release.
This idea is so powerful—and so longed for—that it’s escaped containment among the Apple crowd. I’ve seen everything from Linux distro to phone updates referred to as Snow Leopard releases, when their vendors cite stability and bug fixes over new features. Likewise, people plead with their vendors for a Snow Leopard release when they feel quality has slipped.
This could be the first time I’ve ever downgraded an Apple OS.
As we speak I’m downgrading the Snow Leopard partition on my trusty old Core Duo 2006 MacBook Pro back to Leopard (and while I’m at it, I’m finally getting around to upgrading the FreeBSD partition to 8.0!). I kept waiting for updates to fix the severe stability problems in the Finder, my FireWire 800 ExpressCard and my iMovie HD plugins, but I’ve decided to give up for now.
This was putting it mildly. When I tried it again a few months later, I still had so many problems that I downgraded again, then leapfrogged it for Lion when it was released in 2011. Jeff Johnson has a great post discussing all the problems Snow Leopard introduced and had to have patched. It was a mess.
Still, whether Snow Leopard really was this bastion of stability and polish, the myth surrounding its development and stated purpose obviously resonates today. People invoke its memory because they’re broadly craving stability and quality of life improvements, something almost everyone has wilfully discarded as they chase the next rung of enshittification.
Apple’s marketing team hit on something so popular that it has staying power sixteen years later.
We had weird issues on our home network over the last month or so:
Large files being transferred over Samba and NFS would time out midway through, and rsync transfers would also fail with “Resource temporary unavailable” messages.
Clara would be dropped off our Minecraft server after an indeterminate amount of time playing.
KDE on Fedora would constantly spam me with “Connection established” alerts multiple times an hour.
Our Apple TV would pause a video with a spinning wheel of death for a moment, before resuming.
Now I’m no fancy pants network engineer, but something was wrong. It wasn’t wrong enough for me to seriously investigate though; the cobbler’s child walks barefoot, and I have enough stuff to do at work. The issues also seemed to resolve themselves as well, something I chalked up to the Heinzelmännchen doing their thing whenever something failed.
But eventually last night, the final straw was invoked to break the camel’s back. I was transferring a large repository with thousands of small files, and the damned network kept stymieing my efforts. Exasperated, I disconnected the network cable out of my ThinkPad, and… the transfer finished. Granted it was much slower, but it did the entire rest of the transfer without issues whatsoever.
It was at that point I made a critical—and some may say blindingly obvious—observation: I was now on one of the Wi-Fi networks. So were our issues confined to our wired Ethernet network?
The home office in our flat is in the corner of our loungeroom area, and underneath our tables sat a TPLink 16-port Gigabit Easy Smart Switch. This is connected with an uplink cable to our little ex-Sophos NetBSD npf router in the study via a wall conduit. We originally bought this switch for our old apartment when we didn’t have a separate room for hobbies, so all my retro PCs were connected to the switch along with our regular desktops and laptop port expanders. To say it was a rat’s nest of cables at the time would be an understatement.
The switch looked… fine? It had all the Das Blinkin Lights one would expect, and observing it for a few minutes didn’t show any dropped ports. I unplugged and reconnected every Cat6 cable, connected them to different ports, and even swapped out every cable from the switch to see if it was just a loose connection or an errant cable. I then tried a large transfer, and the issues persisted.
Out of curiosity, I rummaged around my parts boxes and found an classic Belkin 8-port Gigabit switch I’d bought to match my Power Mac G4 QuickSilver:
I unplugged the TP-Link from power, disconnected all the Cat6 cables, and connected them into the Belkin. As mentioned above, the number of ports on the Belkin were more than sufficient in our new home office arrangement, now that the old computers are in the study.
It’s been just over a full day now, and… zero issues. Devices connect instantly. My KDE Fedora desktop has stopped spamming me. Minecraft and other games run uninterrupted. But most importantly of all from a quality of life perspective: I can transfer files across the network without timeouts or dropped connections. Yay!
Yes, this is blindingly obvious stuff, but it’s incredible to me how much I’ll live with an inconvenience when I’m busy or otherwise occupied, when the solution is literally a hardware swap. It also amuses me that the all-metal TP-Link box was SOHO kit, where I suspect the Belkin was made for residential use.
Josh of Espresso Tonic, and Joel of JoelChrono, are excellent bloggers you should read. They’ve both written recently not just about portable devices, but of their form factors. Their discussions on how such design has impacted their use, enjoyment, and productivity with such devices got me thinking how it’s impacted me as well.
There are so many places I could go with this, but today I’ll focus on three types of devices. I sense a series.
Handheld game consoles and tablets
Josh spends much of his post talking about how much the form factor of the Nintendo 3DS brings him enjoyment. If you’re not familiar, the 3DS was a handheld console released prior to the Nintendo Switch, and came in a clamshell form factor. It’s received renewed attention of late, thanks to Nintendo’s regressive decisions in the Switch 2, and an expanded modding community that’s pushed the hardware well beyond its initial capabilities.
I can relate to this. I haven’t posted about before, but Clara and I bought Ayn Thor handhelds recently, after my sister and brother in law convinced us. Naturally I got the one with the beige case and rainbow buttons, because I’m a gentleman (cough).
I’ve had a Switch Lite, and I’ve tried similar “slab” style handheld game devices, but they’ve never gelled for me. I suspect it’s for a similar reason why I love the idea of tablet computers and touch screen laptops more than actually using them. I don’t just reach for smaller laptops because of their keyboards, it’s a fundamental difference in their design that makes them more enjoyable and pleasant for me to use.
Regarding game devices, there’s something awkward to me about using controls on the same plane as a display. Positioning the display for optimal viewing results in the controls being at odd angles to use comfortably. With a hinge, I can independently orient the input and output devices exactly as I need them. This is especially important when on aeroplanes and trains, where I may be constrained for space on all sides. I’m a tall lanky person with arms that stretch from here to Sunday, so being able to keep my arms folded while holding a smaller device with a hinge just works better than a wider, narrower one with buttons on each side.
There’s also something… fun about foldable devices like that. It’s as though I have a little book in which all these adventures are awaiting. The folding position also protects the screen(s) from damage when the console inevitably gets thrown into a backpack, shoulder bag, or well-proportioned chino trouser pockets.
People love their iPads, ROG Ally consoles, and Steam decks, so clearly the limitation is with me. But I hadn’t appreciated until recently just how much the form factor plays a role in whether I’ll be catching up on my Japanese RPGs on the train, or doing something else. I get the feeling this will be a theme on this post.
Cameras
I love OM System (ne. Olympus) cameras. I’ve talked here at length about the Olympus Trip cameras my parents gave me when I was a kid, and how the Olympus OM-D E-M10 MarkII was among my favourite electronics purchases of all time. So much so, that when OM System came out with their OM-3, I deliberately saved and bought it when we last went to Japan.
Those who know their cameras understand that OM System hardly have the best specifications for the price (the bang for the buck, as PC Magazine reviewers used to say). You can get a full-frame kit from the likes of Pentax, Nikon, or Canon for less money than an OM System camera with a much smaller Micro Four Thirds sensor (albeit one with stacking). Likewise, the pancake lenses I insist on using are sharp, but only manage f/2.8 at their widest. A tourist with their Sony Alpha and a Zeiss lens will almost certainly run circles around me, especially in lower light and at greater distances.
But that’s not the point; at least, for me. Olympus and OM System cameras have a specific form factor that I’ve yet to find from any other brand. Their cameras feel crafted in a way that’s difficult to articulate; not unlike a hinged screen (AAAAAAH!). The dials are beautiful and precise, and are just such a joy to use. If you’re a mechanical keyboard fan, you know exactly what I’m talking about. Once you use controls like this, going back to the mushy buttons and dials with play on any other camera feels like a downgrade.
I have taken more photos, and had more fun doing so, because I carry around an OM-3 with a tiny lens. If I had a “better” camera with a “better” lens, it’d likely stay at home.
Storage
I mentioned yesterday how I found the size of MicroSD cards troubling, but that’s not the whole of it. They’re also just a pain to use. The raised lip on the edge of those tiny cards is wholly unsuitable for grabbing onto, and the design of the card slots into which they’re placed seemed to regard ergonomics as a mechanical afterthought. And as I said on that post, I have excellent near-sighted vision and dexterity, so imagine how people who don’t must feel.
As an example, the Tangara MP3 player I use only has one major design feature I find frustrating, and it’s entirely down to the MicroSD card slot. Putting the card in is a pain, and pushing on the card to pop it out requires surgical precision and a steady hand. There’s only so much the designers could do with a surface-mount card slot, but it would been awesome if it had a regular SD card slot instead. It would make transferring the latest FLAC album a breeze, rather than something I end up putting off.
I was looking back to see if you were looking back at me, to see me looking back at you. I-I I-I was… I was looking back…
I’ll admit, this is why I loved the Iomega Zip Drive when I was a kid. How’s that for a tangent? There was something oddly fun about carrying around a 100 or 250 MB disk in my school bag, then popping it into the drives of the school’s Dell Dimension towers, and hearing them whir to life with all my stuff on them. Zip disks were bigger enough compared to regular floppies that I could have an entire year’s worth of school work and assignments on them. Then I’d get back home, pop the disk in there, and update that automatic Iomega catalog program.
I actually… miss this more than I expected. That tactile experience of these chonky disks felt satisfying and deliberate in a way other disks didn’t, to say nothing of USB keys or an OpenZFS volume mounted over a VPN with NFS. Would I trust the latter system over a disk that I’m sure Clicked to Death or flipped bits just by looking at them the wrong way? Absolutely! But that’s not what we’re discussing here.
I had another of those I’m living in the future moments today when I was at an electronics retailer, and I saw a 1 TB MicroSD card. I couldn’t afford it of course, but its mere existence shook me in a way I wasn’t expecting.
One terabyte.
On a thing the size of my thumbnail.
Heck, even this picture of a card from my preferred vendor is likely larger than the card itself, depending on your viewport:
The 486SX tower my dad brought home before I started school had a one hundred megabyte hard drive. This thing had one million megabytes. Yes I know about the whole megabyte/mebibyte thing; don’t spoil my fun. Point is, that MicroSD had about ten thousand of those hard drives in it. Kinda. And rather than filling a shipping container, I could lose it between the lounge cushions.
That’s actually something I’ve worried about a lot lately. More of my devices have switched to MicroSD cards. Leaving aside durability and integrity concerns of data stored on SD and MMC cards, MicroSD cards are… scary. SD cards were about the right size. Even XD, RS‑MMC, and MiniSD cards had enough surface area to comfortably grip. MicroSD cards are easy for me to lose track of, and I have good near vision and dexterity. If you don’t? As my New York friends say: eyy, forget about it!
MicroSD manages to be too big and too small. Imagine burning down the library of Alexandria, only this time you sent it up the vacuum cleaner. It’s the stuff of nightmares.
I suppose people were saying the same thing about my dad’s one hundred megabyte hard drive, having come from punched cards and those unusual 3-inch floppies the Amstrad CPC used. You’re telling me you could fit hundreds of Apollo-era guidance computers on something as small as a paperback book?! That’s nuts!
