The following are a 'stream of consciousness' (micro *log style).

ℹ Chronologically, the main sections are separated by days, with the most recent at the top. The individual sub entries are ordered by minute, again with the most recent at the top (of its respective day section). If you are actually trying to read any of this and feel you are missing context, just scroll down a little and start at an earlier subsection and work your way up. 😉

ℹ Timestamps match the local time in Oslo, Norway [CET/UTC+1 during winter] unless otherwise stated.



Ok, I am still tweaking the formatting of this. I decided that I over simplified the links and the titles where buried, so I made them a little clearer and easier to see. I also decided to use a simpler 'dinkus' as the previous one was a little big and dominating.

Improved link titles; simpler 'dinkus'


Ok, I tweaked the formatting, so that the URLs are as simple (raw) as possible, which has the benefit of making copy and pasting easier as you can quickly select a complete line via a triple click on most clients. In addition, I changed the level two and three headings so they are both less prominent, and removed one space of indentation from, bullets, quotes and preformatted text.

Further simplifed example (compare to yesterdays below)


17:12 [✍ 17:46]

I have been experimenting with converting my .gmi posts into plain text for use on a Gopher hole I have been thinking of setting up. I would be easy enough to just post the *.gmi files but for the fact that I need to deal with wrapping, UTF-8 characters in my posts and also I think the presentation of raw versions of the links could be better.

I wrote a little "quick and dirty" shell script that reformats things using 'sed' and 'fmt'. The biggest problem is actually the lack of proper UTF-8 support in many (most!) Gopher clients. Reformatting of .gmi files themselves to "look good" (well… better) is actually quite trivial due to the simplicity of the format.

So far I am dealing with the extra UTF-8 characters by replacing them with similar characters or text descriptions, or just deleting them altogether (when I arbitrarily decide that their primary purpose was decoration or unimportant). Already it is a pretty massive list (just using characters found in my posts) and has the potential to grow and grow. I would either need to maintain this going forward (for further characters I might want to use) or find a some similar solution that someone else has already created and use that.

Example test .gmi

The above converted to plain text



Here in Norway we celebrate Julaften as the main day of the winter holidays, on the 24th (today). Wherever you are in the world, if you happen to read this enjoy Jul, Christmas, Saturnalia, festival of Sol Invictus, Winter solstice, or whatever you like to call it (and however you like to celebrate it). Happy Holidays! 🎄🎅



By a combination of my latest backup, memory and a couple of fortunate screenshots I took yesterday, I think I have pretty much all of this Flight log back as it was.


From minch,

You need git!

Yeah, yeah… I know. 😒


Ok, I didn't lose too much but this is still very annoying to backfill! 😡


Ok, this suddenly all became VERY meta. I just deleted this entire Flight log (by mistake) and my closest backup was from yesterday 😲. I shall try and fill in the posts as best I recall them and then tack the backup that I have on the end.

12:01 [✍ 2021-12-25]

Recently I have been discussing symbols for editing with minch.

I want something that is both concise and easily understandable. I have settled on '✍', followed by a date/time and (optional) explanatory text, inside square brackets '[ ]'.


Here are some examples of what I mean.

Here is my original text. [✍ 2021-11-30] Some additional thoughts.

## Additional thoughts [✍ 2021-11-30]

Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. 


=> gemini:// Link 1
=> gemini:// Link 2 [✍ 2021-11-30: I found another handy link]

I went back and changed the highlighting of edits in a couple of old posts.

Linkification in Gopher clients 🔗 (A new intro)

Trying Gemini ♊ (New sections, with one referenced via '§')

I see that minch also adopted this style in one of his updated posts.

Sonic the Hedgehog on the C64 - minch (Added a link) [✍ 2022-01-22: updated link, thanks for the tip Alexey]



I was just thinking about minch's quote in relation to Gopher (which also first appeared in 1991) and Gemini. He could have just as easily written.

From the moment [you start exploring] it's 1991 all over again, except it's 30 years later and you're [using Gemini]!

[✍ 2021-12-23: tweaked the fake quote (above)]

It totally works, right? 😉


Now minch is starting to write the content I would expect from him, rather than just meta stuff about Gemini on Gemini.

From the moment the game starts up it's 1991 all over again, except it's 30 years later and you're on a C64!


Sonic the Hedgehog on the C64 💨 [✍ 2021-12-23: Updated link—new domain]


Enjoying chatting with minch on how he plans to use Gemini going forward. The index of his capsule already looks pretty nice.

[✍ 2021-12-23: Updated links]

minch's 'Retro' capsule

Why Gemini? — minch

Choosing a Gemini Server — minch


It started to snow 🌨 again. Yes… that feels more like it!



Besides my favoured 'Information Source character' (ℹ) [§ 2021-12-18 19:25] for break-out explanatory text, some other characters that could potentially be used for the same kinds of purposes include '💡', '💭' or perhaps the more traditional '☞'.

Commonly used to represent ideas (as over a head in a cartoon), thinking, and learning […]

Light Bulb (Emojipedia)


[The Thought Balloon is] Used to represent thinking, or thoughts, […]

Thought Balloon (Emojipedia)


The typical use of the pointing hand is as a bullet-like symbol to direct the reader's attention to important text, having roughly the same meaning as the word "attention" or "note" […]

Index (typography) - Usage examples (Wikipedia)



In the last few days, in an attempt to make my posts clearer I have read (and thought) a lot about stuff like: dinkus (and flourished section breaks); footnote symbology; information/note symbols; pilcrows; section signs.

Now you could certainly argue that:

However, to be honest, it was interesting going down this route of discovery. I learnt a lot about symbols and many things I have seen before suddenly make much more sense to me. I also have a better feel for how text was handled pre-web. I have a certain fascination for alternative ways stuff was (or could be) achieved. Hence Gemini, Unicycling and half my other interests.

So no matter how I continue with regards to Gemini, this was all worth it.


Yes, I got a friend to start a capsule. What do I win? 🎁

This is what happens when you have conversations with your friends about niche technologies and you dive into the rabbit hole... Thanks Ruari!

Why Gemini? [✍ 2021-12-23: Updated link—new domain]



And now I am thinking about Section Sign (§) to reference a headings.

Section sign (Wikipedia)


I reckon '§ Heading name' is reasonably intuitive. I have edited my 'make uninstall' post and added in a couple of these and I feel it is quite understandable.

'make install', uninstall help (howto remove manually installed software)

[I also updated some entires below to add in a few.]


Using a Pilcrow (¶) actually seems like a pretty good way to manually reference within a document, similar to same page 'anchoring' (think `<a href=#foo>foo</a>` with `id=foo` on the web).

in academic writing, it is sometimes used as an in-text referencing tool to make reference to a specific paragraph from a document that does not contain page numbers, allowing the reader to find where that particular idea or statistic was sourced. The pilcrow sign followed by a number indicates the paragraph number from the top of the page.

Pilcrow - Modern usage (Wikipedia)


Using '§ ¶5' or (perhaps even just '¶5') to point the reader to the 5th paragraph is pretty handy.


If you are using daggers († & ‡) for footnotes and need more than two, you can also use combinations and multiples of each, e.g.

ℹ When many footnotes are used, it is generally more practical to use consecutive numbers or other symbols to identify each footnote.

Some other symbols, that primarily serve another purpose but have been used as footnotes (according to Wikipedia), include §, ‖, ¶, Δ, ◊, ↓ and ☞.

Note (typography—Wikipedia)





◊ 'LOZENGE' (U+25CA)



While searching around on this topic, I have also seen the usage of 'playing cards' symbols as a way of doing footnotes.

Beckhoff Information System - Automatic generation of footnote symbols


♣ 'BLACK CLUB SUIT' (U+2663)



ℹ Personally, I would avoid using unicode 'superscript letters' when I have lots of footnotes because there are only a small number of true superscript letters in Unicode. It is far from a complete set. There are some that look like superscript letters but they are mostly meant for other very specific uses. Additionally some letters have nothing that even 'resembles' a superscript equivalent (i.e. 'q'). This means that if you use a bunch of them you have to deal with the potential for font switching issues that I mentioned [§ 2021-12-20 11:34] but for the letters it is in my opinion often visually worse in many cases. It is also semantically worse, because while the superscript numbers come from two different sections ('¹²³' are from 'Latin-1 Supplement' [U+0080 → U+00FF] and '⁰⁴⁵⁶⁷⁸⁹' are from 'Superscripts and Subscripts' [U+2070 → U+209F]), they are all intended as actual superscripts. Because of missing letter characters and/or things that could potentially display oddly, you may even find that you want to skip over certain letters (which is potentially quite confusing).


There is also a 'triple dagger' that has been used historically to indicate another level of notation. However the actual unicode symbol for this is much more recent (Unicode 11.0 [June 2018]) and as such it does not have a lot of support with many fonts, thus meaning you cannot really count on it displaying correctly. For this reason I would avoid this character for the time being.



Some other characters that are quite handy include superscript numbers for use as footnotes.











ℹ The characters '¹²³' are part of 'Latin-1 Supplement' (U+0080 → U+00FF), while '⁰⁴⁵⁶⁷⁸⁹' are part of 'Superscripts and Subscripts' (U+2070 → U+209F). As such the first three may display slightly differently from the latter group because some font sets will only contain the first ones. This can cause font switching. If you are concerned about this and have lots of footnotes, consider other forms of footnotes, e.g. the pure ASCII '[1], [2], [3] …' style found in many plain text documents.

Unicode Block 'Latin-1 Supplement'

Unicode Block 'Superscripts and Subscripts'

Alternatively, the two 'dagger' characters can also work if you only have a couple of footnotes.

† 'DAGGER' (U+2020)

‡ 'DOUBLE DAGGER' (U+2021)

ℹ Traditionally the convention is to begin to use daggers only after you have first used asterisk (*) as the initial footnote but that would be problematic in Gemtext because if you start a line with '*' you have a [Gemtext] bullet point. You could potentially work around this by use of a space before the '*', or perhaps a 'zero width space' [§ 2021-12-19 22:38]. However it is probably best to just avoid '*' altogether (for this usage in Gemtext) because a user might still interpret them as bulletpoints if their client does not render using the '•' character. Additionally asterisk is also used informally in plain text (and officially in many simple formatting languages like Markdown) as the start (or end) of emphasis. You don't want the reader to have to think, "What does '*' mean this time?". For these reasons I would consider asterisk a poor choice for (Gemtext) footnotes and skip straight to daggers if you want to have a couple of footnotes.

A dagger, obelisk, or obelus † is a typographical mark that usually indicates a footnote if an asterisk has already been used.

Dagger (mark)—Wikipedia



Here is another interesting character '​'.

Can't see it? That is the idea. It is a 'zero width space'. Why would you want this, you might ask. Well some websites (e.g. Twitter) are a little too aggressive at turning text into links, if they consist of alphanumeric characters and a single 'Full Stop' (. [Period 🇺🇸]). Thus they would turn '' into a link that points to ''. This can cause confusion, with some people trying to click on these 'links'. However place a 'zero width space' after the '.​' and it looks correct, without automatic 'linkification'.

ℹ This trick is not perfect. If you were to 'copy and paste' a file name tweaked with a 'zero width space', it would not match the real file name, since it contains an extra character. You need to be careful with this one, so as not to lead to other types of confusion.


Ctrl+a, Ctrl+c in the 'text area' to get one in your clipboard


I think I might write up an article on useful Unicode Characters (including Emojis). Recently I have been using asterism (⁂) and Information Source characters (ℹ) in my writing [§ 2021-12-18 19:25] but there are plenty of others that are handy. In the older days of Twitter, when tweet length was particularly restrictive I made a lot of use of ellipses (…) and interrobangs (‽) (why use more characters than you need?). I also frequently use '¤' (denotes an unspecified currency). Norway has no 'Norwegian Kroner' symbol and typically people append 'kr' (one character longer) or 'NOK' (longer and possibly redundant when context has made it clear I am speaking about Norway).

Some others I tend to use include (in no particular order): bullet (•) (and white bullet [◦] for sublists), em dash (—) and various arrows (←↑→↓).

I also use 'down tack with circle below' as a kind of mini ASCII art for a unicycle, since no unicycle Emoji exists. Though since this is not its intended usage, and support is missing on many systems, I realise that is non-ideal from an accessibility standpoint. 😔

⁂ 'ASTERISM' (U+2042)





• 'BULLET' (U+2022)


— 'EM DASH' (U+2014)


↑ 'UPWARDS ARROW' (U+2191)




(Hmm… It seems I am halfway to a post already 🤔)


I just stumbled across a very interesting article on 'Web3', Blockchains, NFTs & DAOs. Very nicely written. I highly recommend it.

The Third Web - tante



On second thought, I have decided that the Information Source unicode character 'ℹ' (U+2139) perhaps works even better (than square brackets '[ ]') for these additional notes.

A lowercase letter i, enclosed in a rounded square or circle. It is used in apps or websites to indicate the availability of help or further information.

ℹ - Information Source Emoji: U+2139 - Unicode Character Table

Unicode Character 'INFORMATION SOURCE' (U+2139)


I am still (almost constantly) editing old Gemlog entries. Most of it is related to formatting. Since Gemtext is quite restrictive, I really need to think about how best to present things.

Most recently I switched to using asterisms '⁂' as 'flourished section breaks' or 'dinkus' (similar to how one might use '<hr>' in HTML).

Using ⁂ for section breaks (Wikipedia)

I also decided to make use of square brackets to present the additional, background notes in my article on uninstalling software.

Square brackets are often used to insert explanatory material […]

Uses of [] (Wikipedia)

Square brackets [ ] are used to enclose a section of writing or printing to separate it from the main text.

See section: 'How to use square brackets in writing?'

'make install', uninstall help



I decided to expand on my previous short reply.

Re: Do You Ever Feel Stuck In A Loop?


I have to say, it is rare that I relate to a post so completely

Do You Ever Feel Stuck in a Loop?

What have I done for the past few weeks?

• Running.

• Working.

• Watching series and films on Netflix and Disney+.

• Procrastinating everything on my projects list, because the first and most important item is boring.

and later…

This shouldn't be a problem; after all no person's value is measured in achievement and no time is better spent than the one you enjoy. It does, however, produce a certain amount of anxiety to want things done but not wanting to do them. I feel stuck, running in circles and trying to escape reality by watching almost anything Netflix recommends to me.

Change a couple of these, like so:

and this could have been me.


I do still hope that more Gopher clients add support for extracting URLs from pure text documents to make navigation easier. Having played with this feature further in the VF-1 client, it is just as nice as I imagined.

Here are two that do that.

VF-1 (Github)

VF-1 (Download via Gopher)

GemiNaut (Gemini/Gopher [Windows] client–Homepage)

GemiNaut (As above, via Gemini)



It is so annoying when you think you have a great idea, that should have been obvious to everyone, and then you realise it was obvious to everyone and YOU are the one who is slow on the uptake. 😳

I am stupid



And now one more! It feels a lot easier to reach out to people in Gemspace than I would have expected.

Wirth's Law


I feel like I am slowly working out this (Gemini) community and have already had very pleasent exchanges with a bunch of people here, including:






Nice to meet you all! 😃


And… both the kids are now at home again. "Working" will be interesting again. The school for the eldest shut because of outbreaks and the youngest we decided to keep home until we all get properly tested.



Cool, finally comments on my gemlog! One step further to being a proper part of the Gemini community. 😃

Linkification in Gopher clients


I really have a thing for constantly edited and re-editing my *log entries. I have always been like this. I should really learn to either control myself or actually wait until I am done until I publish!


I wrote up a few quick thoughts about handling links in raw text files on Gopher.

Linkification in Gopher clients


Ok, I just discovered this and I have to say, it is very awesome!


Two things happened recently that depressed me somewhat

The end result of both of these is that I am not doing a unicycle commute through the snow. 😞

Thoughts on unicycling in the snow

Last year I started to do a "commute replacement" when it was primarily home office because commuting is one of the highlights of my day. So I would cycle off towards to woods for 5km or so and then circle back and arrive "at work" (home). Maybe it is time to starting bringing this practice back. Indeed, since they are a bit colder, there might even be some snow left in the woods. 🤔



Looks like there will be further locking down here in Oslo as the case load is going up rapidly. Press conference in 1½ hours. Will it ever end!? 😞



I just checked Google Maps. It claims the following times for travel from my place to the remote post office (these are one-way times):

(I unicycled and it took me 50m round trip. Not bad considering all the snow)


Until recently it was 200m (656′) from the front entrance of my apartment to the local "post office". Not a full post office but a mini one in a supermarket, opposite where I live. Any package that was too big for our post box, would end up there. It was quite convenient.

A couple of months back this mini post office shut down. From this point on new parcels were delivered to another mini post office that is 1.5km (0.9mi) away. Not too bad but it is up a steep hill, and it is also a non-obvious (to me) choice because there is actually another mini post office that is only 480m (1575′) away. Why had they assigned our new post office further away, when there is one so much closer?

Even more strange, once (just once) we did get post delivered to the nearer mini post office. At the time I thought, "Oh good they have corrected it. From now on we will not need to walk so far". Then, the next parcel received went to the one farther away again… 🤔

Yesterday notification of a new parcel arrived. The note informed us that it was sent to yet another post office. This time, 4.9km (3mi) away. WTF is going on?



I took an extended route into the office this morning so that I could pick up a couple of presents for my youngest for Christmas, from the gift shop at the Oslo Technical Museum.

It was pretty empty there. There were some new Corona regulations announced last night and it is cold and snowy out. My unicycle looked a little lonely in the bike racks. Poor little guy! 😜

Unicycle parked in the snow

I also took the opperatunity to stop at a local coffee and donut shop, to pick up some 'healthy' breakfast for me. 😉

Donuts and coffee

My route on Strava



Strava is having issues with their APIs and uploading activities. I hope this resolves soon. I am irrationally impatient 😉

Strava Status page


I really would like a "TallBike" but given that my wife already fears for my death on the Penny Farthing, plus I have no build skills and paying someone to build a custom one for me would cost the earth, I guess that is not happening anytime soon. 😉

Tall Bike (Wikipedia)

Wolfgang Seibel's 9ft Tallbike (Strava)

Wolfgang Seibel (Strava)



I once wrote a script that automated extracting a Widevine binary out of a ChromeOS ARM image. This is handy for people using Vivaldi (or other Chromium based browser) on 32bit ARM Linux, as there is no other valid source for this file, which is used to handle DRM on sites like Netflix.

Script to extract Widevine from ChromeOS

In that script I do a couple of things in a quirky "me" kind of a way. Nothing so obvious as using backtick subshells but still things I suspect others would do things differently. Anyway one day I bought a Pinebook Pro (a small, cheap ARM notebook) and on the copy of Linux that came with it (I cannot recall the distro) was Chromium. Interestingly this Chromium worked with Netflix. Curious I looked around and found a script, that not only used the same trick as me (getting a Widevine binary from a ChromeOS image) but it even had several of my quirks. It was not an exact copy of my script, but clearly they had seen mine and "borrowed" the heart of it.

I never actually specificed a license for the script above but it is simple enough and I am pro open source, so it is fine. Nonetheless it might have been nice if they pinged me first or at least gave me some credit.


Following on from yesterday about how I like to do wierd things in shell scripts, because… well… just because. I suddendly remembered Gemlog post on uninstalling software compiled from source.

Under the 'Alternate versions' section you will find a version of my tracking script using backticks `` for subshell variables.

#!/bin/sh -eu
m=`expr \`expr \\\`date +%s\\\` - \\\`stat -c%Y "$1"\\\`\` / 60`
find /etc /opt /usr -mmin +`expr $m - ${2:-1} | sed 's/-.*/0/'` -mmin -`expr $m + 1 + ${2:-1}` ! -type d

Almost nobody does stuff this way anymore because escaping nested subshells is such a pain and all modern bourne shells are perfectly happy with the easier to work with "$()" convention. But I still like to do stuff like this though for fun.

Software Uninstall (Gemlog)


Ok, I have decided to try keeping a basic journal (i.e. this page you are looking at) on Gemini to try posting some of the stuff I usually have on twitter and see how that feels for me. To give myself a little pre-content I copied over a handful of my tweets from yesterday (below). So I suppose this is the first "real entry". 😉

Time will tell if I keep this up.



I do like doing weird shit in scripts actually. I have this snippet in of mine to check if `strings -o` is outputting in Octal or Decimal

[ "$(printf 'popular\novel' | strings -o | sed -n '/v/s/ *\([0-9][0-9]*\) .*/\1/p')" -eq 10 ]

If it is true, then it is Octal.

Indeed that is actually the problem with `strings -o` and the reason for checking it if you are relying on its output. On some systems you get a result in Octal and other times Decimal (`-t` is not always present). As an example, GNU strings returns offsets in Octal, while BSD returns them in Decimal. I actually don't remember right now what BusyBox does.

Could I test in a better, clearer way?… Yes

Would it be as fun?… Probably not 😜


Excuse me while I delete the contents of this file, while leaving the (now empty) file in place

$ >somefile

(…actually… that is pretty handy… and also shorter than `touch somefile` to make a brand new zero byte file… 🤔)

At a previous job, during the interview I was asked how to quickly/easily do various things on the terminal to test my basic *nix knowledge. One of those questions was about creating a new, empty file and I suggested using touch for this. Now I wish I had said that… much neater.


From now on I will write all my shell examples in the most unexpected (but valid) way possible, just to screw with people. I was already often doing this with tar commands but now I shall take it up a notch.

Let's just check the approx size of the most recent text file:

$ ls -hr *.txt -lt


Once I realised that was valid, it then occured to me that you can even do this:

$ <somefile>outputfile someprogram

It is just… wrong 🤪


Ok, I have used *nix for a while but I must admit TIL that this is possible:

$ <somefile someprogram

I have always done it:

$ someprogram <somefile

The top one just looks so weird to me 🤯

P.S. Ha… I started writting this at 13:37!


🔝 Capsule index