smiletolerantly

joined 7 months ago
[–] [email protected] 40 points 2 months ago (6 children)

Sure!

The two sensors and apps of the same name I'm complaining about are the Freestyle Libre 3 and Dexcom G7. IMO, absolute bare minimum for what is required in an app of this kind is:

  • get the glucose level from the sensor and show it in the app, incl. a history graph (even the old dedicated handhelds did this)
  • play an alarm when that value is below/above a low/high glucose threshold

The Freestyle Libre 3 does this, and absolutely nothing more. They ported the software from their dedicated device to Android and called it a day. Frustratingly, this means that you don't even get your current glucose in the notification area, or the lockscreen, or anywhere else. You have to open the app. You can set the alarm thresholds arbitrarily, but only get 1 high and 1 low alarm setting. Disabling these means you need to go two levels deep into the settings menu. And you WILL be doing that, constantly. Why? Imagine this: You get woken by an alarm for high glucose. You check the app, and see that you've just barely crossed the threshold, but have been hovering below it for the past hour. You take insulin, but of course, that takes time to act. Since sensor measurements are a bit jittery, you can count on the glucose level to dip back below, then back above, below, above,... the threshold for the next 1-2 hours. The app will blare an alarm at the highest volume your phone is capable of every. single. time. Your (and potentially your partner's) sleep will be interrupted IDK how many times, unless you completely disable the alarm. Be sure to remember switching it back on in the morning, though! :) The fix would be incredibly simple: either allow muting alarms for a set period of time, or don't play an alarm if it has recently been played and the amplitude of glucose measurements is small.

On the flipside, if you miss an alarm - say you are speaking in a meeting so you swipe away the blaring alarm - you will not be reminded again. My high glucose alert is/was set rather low (150 mg/dl) because that's the point where it makes sense for me to take additional insulin. If that alarm goes off and you dismiss it, it will not go off again (unless you first dip below, as discussed). Doesn't matter if two hours later you are at 160 or 380. Sure, it is also my responsibility, I am not denying that. But again, adding this feature would have been hugely helpful and so, so easy to implement.

Let's take a quick sidebar and talk about hardware. The Freestyle Libre 3 is an AMAZING piece of hardware. It's incredibly tiny - about two pennies stacked, both in diameter and height. Still, it sticks to your skin very well, and measures and sends your blood glucose to your phone via Bluetooth every minute for two weeks straight before the battery gives out. As if that weren't enough, it's also so well-made that in the almost two years I was using it, I did not have a single unit that was defective or ran out of battery in less than the promised two weeks. It's fantastic.

...but sending glucose info every 60 seconds is also a drain on your phone's battery, especially if your app isn't, uh, made well. Take a guess as to where this is going. Ready? Wrong, it's worse. If I would go to bed with 100% battery, I would wake up with 40% left. I needed to charge my phone 4 times a day. Since switching to the Dexcom G7 (imperfect as it is - we will get to it, I promise), that is back down to once a day.

Apart from draining the battery, the app is also slow, and gets really slow the longer you have it installed. This mainly shows itself in the glucose graph being slower to render over time. Deleting the apps data, i.e. starting fresh, makes everything run reasonably fast again. Over the course of a couple of months, a couple hundreds of megabytes of app data accrues, and the app starts to crawl instead of run.

The next part is pure conjecture on my part, but I am still going to share it. My theory is that every time you open the app, the entire graph history is computed from scratch. That would explain the amount of data <-> slowness relation, and probably at least partially account for the battery drain.

In short: Freestyle Libre 3 (the hardware) is fantastic, Freestyle Libre 3 (the app) does the bare minimum and sucks at it.

You can not use it with xdrip+, except by installing a cracked version of the original which exposes the current glucose level through a webserver on your phone, and an additional app which grabs the values from there and passes them on to xdrip+. This is error-prone and does not exactly help with the battery drain.

On to the Dexcom G7. In the next comment, because apparently there's a 10k character limit on comments.

[–] [email protected] 171 points 2 months ago (13 children)

OK, this is only tangentially related but it has been on my mind lately and I need to rant:

I am T1 diabetic. Over the last decade, a LOT has happened to improve my life, especially in regards to no longer needing to check glucose levels with blood, as glucose sensors you wear on your arm have become ubiquitous.

It started with a dedicated device that you needed to hold up to the sensor to get a reading (much nicer than pricking your finger) to that sensor being able to notify the dedicated device of high/low glucose values (yay! Sleep through the night, knowing you'll be woken up if something is wrong) to the sensor now constantly streaming glucose values to your phone.

Which is fantastic.

In theory.

In practice, there are two companies making these sensors (OK, there's a couple more, but they suck way more and are much less commonly used).

And both of their closed-source apps suuuuuuuuck. They do the bare minimum and nothing more. (Actually, it's worse than that. Ask me if you want to know. It's its own rant.)

Then there's xdrip+, a FANTASTIC app made by diabetics for diabetics. Instead of just showing you "this is your glucose" and sounding an alarm, once, when it's required, you can (just off the top of my head): Set an arbitrary amount of alarms with their own behaviors, which can be configured to vary by time of day; show the glucose everywhere (notification, lock screen, home screen,...); mute alarms for a custom time; do not sound an alarm if you're trending in the correct direction fast enough; do not sound the alarm multiple times if your are jittering around the threshold; notify other people automatically in case of emergency; and roughly 1000 things more. The app is well maintained, and of course open source.

Can you guess what the problem is?

That's right, manufacturers disapprove of using this app. For the worse one of the two sensors mentioned, the community reverse engineered the communication and it is now working perfectly with the app. For the better sensor, they can't and won't due to fear of legal repercussions.

It's my health. And I need to decide between worse hardware and useless software.

There's no technical reason for this. I dream of the EU passing a law that requires manufacturers of wearable medical devices to publish the comm protocols and to legitimize use of third party software.

Rant over.

[–] [email protected] 3 points 2 months ago (1 children)

It was pretty great, wasn't it?

Although I must say. I eventually landed on neovim. Steep, steep learning curve, but now I would not switch back again.

[–] [email protected] 4 points 2 months ago (3 children)

I gave it serious consideration when the death of Atom was announced and I was unsure where to move on to.

Looks like in the meantime a lot has been done (as far as I remember, TreeSitter and LSP weren't built in back then...? Not sure though), but the lack of a plugin system is still killing it for me.

TBH it looks like it has 75% of the features you want from a codeditor, which is much more than the use-case for Nano, but no way to go the remaining 25% of the way.

[–] [email protected] 14 points 3 months ago

No.

I've been using it for 10 years. Back then, it just started out as a chat app with group support - just like Wahstapp, but free (yes, WA used to cost money) and way better than SMS.

My entire social circle switched to it, and has been using it ever since. Why? Because to this day, it's easily the best chat app, feature wise. Literally every time WhatsApp or Signal or Threema add a shiny new feature, Telegram has already had it for a while.

Since Covid however, there is a huge stigma attached to it, and I do get why. It's sad, really. I wish there was a 1:1 clone of Telegram's chat features, minus the Channels (or whatever they are called).

[–] [email protected] 0 points 3 months ago (1 children)

I hate the labelling of Telegram (and also, to be fair here, WhatsApp) as social media platforms.

That's like saying email is a social media platform because mailing lists exist.

[–] [email protected] 20 points 3 months ago (1 children)

The "anti-woke" (or whatever they call themselves) hate is absolutely ridiculous, and anyone who takes it seriously is an idiot.

That being said though... The show was just bad. Nonsensical or missing character motivations, a thin excuse for a plot, and a huge disappointment compared to what it could have been.

I really hope the reactionary hatred didn't factor into the cancellation, or at the very least that Disney doesn't take away from this that they should hire fewer women and people of color. I hope they take away to invest more into better writing.

[–] [email protected] 5 points 3 months ago

Yep, same here.

Samsung TV cannot reach the internet. Nvidia Shield is running Flauncher as the default launcher. I watch all my content through Jellyfin and SmarttubeNext.

I have not seen an ad in years.

[–] [email protected] 4 points 3 months ago

Yeah this is severely lacking in terms of theoretical compsci.

[–] [email protected] 2 points 3 months ago

I know it's been three weeks, but thanks for telling me about this! I might actually do this, for the projects here and there which aren't packaged into nixpkgs (yet).

[–] [email protected] 1 points 3 months ago (1 children)

Thanks for the suggestion, I must say though, I am very happily de-googled :D

[–] [email protected] 1 points 3 months ago

Thanks for the recommendation! Looks like a great option. Actually, the p2p aspect prompted me to have another look at the Jitsi docs, and lo and behold, there's an option for that, as long as no more than 2 people participate in a chat.... (The reason I'd prefer Jitsi is actually just that NixOS comes with options for jitsi out of the box, for Miro I would have to introduce containers into my setup :D)

view more: ‹ prev next ›