Notices by Elias Mårtenson (loke@functional.cafe), page 2

So, I've been playing around with the GPX file format and PostGIS. My goal is to build a tool where I can run my own server that collects GPX tracks from bike rides, and match them against predefined routes. Like a self-hosted Strava if you like.

I started this because the enshittification of Strava has reached a tipping point and I cannot in good faith upload anything to them anymore.

Once I get this working, I want to add the ability to share rides and favourite segments via the Fediverse. Not entirely sure how far you can take that, but that's something to think about later.

Anyway, it seems I can use PostGIS to match a ride against a set of segments, at least when using a small set of test tracks and some synthetic data that I've generated.

Is there anyone here that is familiar with PostGIS who would be willing to advice me as to whether I'm actually doing this right? I can't help but think that my successful experiments are actually pure luck.

@clacke it's certainly possible that it went the other way. They're too similar for it to be a coincidence, no?

What does enshittification look like? Well, most of you know what it looks like, but here's another example.

Today I was biking in an area I had never been to before, going on a bike route that takes the route of an old railroad line.

I needed to go back where I came from, so I asked google maps for a bike route back. It recommended a 43 minute route that began with going in the opposite direction, and then snaked its way via small side roads to where I needed to go. The bike road I was on wasn't even properly marked.

I then when to OsmAnd (openstreetmap for android) and it gave me a 30 minute route that started by turning back on the same road I was, back for about 10 minutes and then pretty much straight back.

I think I know why this is. Google Maps no longer exists to make navigation easier. For users, it's become little more than a vehicle to provide connections to businesses. Taking it a bit more abstractly: Its real goal it's all about location-based advertising, and bikers don't make them any money.

@clacke same here. I wonder if the american comes from someone mishearing bubble tea?

@clacke it's called the British style, and I guess it's used in former colonies. It's interesting that Singapore has then the right way around.

Fortunately it's ready to switch them when you rent a bike. I recommend you do that since your instinct will take over in a panic situation and you can easily go over the bars.

@amszmidt @emacsomancer you can actually do that in vim too. All the necessary infrastructure exists in vim to do that. The reason people don't do it much is partly because vimscript is not something a sane person wants to program in.

But it's also because vim users are in feneral more aligned towards your view on what an editor is supposed to be. No one really wants to build applications on it like they do in Emacs. Rather, they prefer to launch vim from the application that needs to edit something.

So while there are technical reasons it's mostly a cultural thing.

@amszmidt @emacsomancer but it lives in Emacs, using the GNU Emacs infrastructure like an operating system.

Is GNU Emacs a great platform to build applications on? A proper Lisp operating system would be nicer, but it's still better than Electron.

Swedish had a word "dygn" that refers to a 24-hour period. It's extremely useful and I am often very annoyed that this word downy exist in English.

English users the word "day" most of the time and just home there won't be any miscommunication, or they use cumbersome versions like "24 hours".

It also leads to very silly expressions like "I'll be on vacation there for 3 days 2 nights", instead of just being able to say "I'll be there for 2 dygn".

@amszmidt @emacsomancer Not totally, and unless you are very careful when you define the words "editor" and especially "lisp machine", all of us who lament the state of modern operating systems will keep getting misunderstood.

Emacs is certainly an 'ersatz' Lisp Machine if you define 'lisp machine' as a thing that runs some kind of Lisp and contains a user interface on top. That completely ignores everything that makes Lisp machines interesting, but we have to acknowledge that people doesn't know this.

Every time I listen to american accents being spoken

@tfb Absolutely nothing at all, until you need to say the name of the Audi engine: CCZC https://mymotorlist.com/engines/audi/cczc/

@clacke a hint of natural flavours. A hint?! (goose chasing the company) What the fuck are the rest of the flavours then?

@tfb @clacke speaking Swedish with German grammar sound work in the Netherlands.

@libreoffice I paged down page, after page, after page, and with maybe 3 exceptions, every. single. fix modifies code comments, change the character case of html tags.

Here is one of hundreds of examples: https://github.com/apache/openoffice/commit/8c3681a51823f1c4a400590c771b000ee885c29d

What on earth is going on? Surely a human isn't assigned to make a few of these changes every single day? Or maybe the contract allows them to afford this?

@libreoffice I don't understand the openoffice project.

They claim to maintain it, and if you look at the repository, there are updates all the time, but pretty much every single one of them are reformatting, some typo fixes, changes to formatting of documentation, etc.

You have to scroll way down before you find a single fix that actually changes the code.

https://github.com/apache/openoffice/commits/trunk/

Just read the commits. It's so very strange.

Why are hey even spending the time doing this? The only explanation I have is that they have some contract where someone is paying them money for it to be updated, and they do the absolute bare minimum to avoid breaking the contract.

This, unfortunately, has had an extremely harmful effect on the open source community, who are best represented by libreoffice here.

@larsbrinkhoff @ramin_hal9001 @kentpitman the old unexec model is gone from emacs now though. It wasn't compatible with some changes to glibc.

@skinnylatte And to be clear to anyone not familiar with Singapore housing, this is supposed to be the cheap apartments.

Condos are much more expensive.

@allison @ezio yes, 2.5.1 was the first version that was production grade, especially with the later patches.

I do remember that to get that Solaris 2.0 beta to run we had to dig out some really old hardware in the lab. Thankfully we had shelves full of it.

@allison @ezio Sorry, I didn't. Me and a few other colleagues did play around with getting old stuff that we found running, like a version of Solaris 2.0 beta that we found on a CD (it was terrible). I also remember coming across a CD with Sun Nextstep.

But I don't think I ever heard about Spring.

@allison @ezio Specifically, they named the new product Solaris 2.x, with the underlying operating system named SunOS 5.x. The BSD-based SunOS versions were retroactively renamed Solaris 1.x, with the underlying OS called SunOS 4.x.

This is indeed very complicated, and most people just kept calling the old system SunOS and the SysV-based one Solaris.

Source: I was working for Sun around the time when Solaris 2 because usable.