Dabbling in some years old matlab code is quite the blast from the past. Good thing octave can run my analysis code just fine. The simulation itself would take forever though, because octave doesn't support `parfor` it seems. They have a parallel package, but I just hope I'll never have to touch the simulation code again - fingers crossed.

Also, talking about tying tools together: jc and jq are pretty cool. jc can parse the output of some common cli tools and convert it to json, and with jq you can run filters and transformations on json.

Can we stop pretending that tying several little "do one thing well" tools together doesn't introduce its own additional complexity, as soon as we don't just talk about piping commands into each other? As every approach it has its strengths and weaknesses ... mind-blowing!

I finally understand my old vim/latex setup again 😂 After hours of frustration it is kind of fun again to work with it. But I'm getting too old for this shit, I don't have time any more for tools that give you 100 ways to shoot yourself in the foot until you find the right way (TM) of doing stuff.

After several years, having to touch some old documents again. I have a love/hate relationship with it. Is there an up-to-date best practices guide? Regarding tools and packages. I'm using latexrun with lualatex and biber - no idea why the last two, I guess someone recommended it back then ....

I can finally automatically deploy my website. My repository is hosted on where I register a webhook that notifies my webhoster to pull the repository on updates: nmehne.de/posts/2020-07_source

But of course, this was all only an elaborate procrastination scheme ... back to whatever I was supposed to do in the first place ...

Trying to "usb boot" from a 2 v1.1. I don't get what I'm doing wrong. Supposedly you can put the /boot/bootcode.bin file on the sd card and put the real os on a usb drive. The rpi should read bootcode.bin and then use the usb drive. But I only get lit up red+green LEDs which means it didn't work. Not sure what I'm missing - do I need to specify somewhere which partition it is supposed to use for the boot, or does it automatically detect it?

I'm on my fifth smartphone in nine years. Trashed two (one after 11 months, the other one after three years), sold off another two. Not particularly sustainable. Spent roughly 100 to 120 EUR/year for smartphones (including peripherals and spare parts) and need to get at least another 1.5 years out of my current phone to reach the same level. It's a used one, so hopefully the battery will last me that long, because replacement batteries tend to suck ...

biking to work 

First time going by bike to work. Didn't really take longer than going by public transport. The daily car influx into the city feels absurd, when you see it from the outside. Will bike there more often from now on. Wondering how miserable it would be to go when it's raining ... I was quite exhausted on my way back towards the end and conditions were pretty much perfect.

So, the "webhook" concept seems to exist outside hooks and must be configured via the code hosting platform. Supposedly supports webhooks but I don't get how you can configure it ...

every time: hm, I guess I can put these two commands in a shell script. Better use variables ... How does variable expansion work again? Err, better exit early on errors and unset variables. set -e or sth? Oh, better add a version compatibility check. Uh, what's the if syntax? Hm, what were the weird comparison operators again? Argh, need the absolute value for the version check. #&@-$&$-@¥£™√¢×{~\}=|

Just found out that my t440p has a socketed CPU ... kind of cool, people seem to upgrade it to 4 cores @ same wattage. Not sure if it makes economically sense to do it, though, because an i7 4702mq is currently 100 EUR+ on ebay. But are there new laptops that are sold with socketed CPUs? Anything "ultra book"-like seems to come with soldered CPUs which feels like it's 99.9% of the market ...

So if I start again, I want to be able to easily rebuild the server configuration if something goes wrong. What is an easy (to learn, to run) way to do so? From hear-say I guess I'd need sth like ansible? Would that be a good start or too complicated a solution for such a tiny project?

Rather entertaining talk that shows the importance of frictionless data analysis: How a File Format Led to a Crossword Scandal - Saul Pwanson: youtube.com/watch?v=9aHfK8EUIz

Found via lobste.rs

Julian :arch: :linux: boosted

Thinking about getting into again. I used to run a rpi2 and it was powerful enough for my needs. Is there any SBC nowadays that 1) has full upstream kernel support (I don't care if graphics are not fully supported) and 2) boots from usb so I don't need flaky sd cards?

So I don't quite get mastodon, yet ... may have missed sth in the docs. It feels obvious to me that I should use # searches to find interesting content in the federated timeline. I can do one-off searches (""), but that seems cumbersome to do explicit manual searches. Isn't it possible to create custom timelines? Like, maybe I want to see toots for "" and "" in the same timeline. Or what am I missing, how do you find interesting federated content?

Just came across this sentence on peertube's wikipeda page: "Each server hosts a torrent tracker and each web browser viewing a video also shares it.."

Uh, if I understand that correctly it is dangerous for me in the current legal system I live in. I don't know if the stuff on there is legally uploaded, so by watching a video I may be re-distributing it illegally which is potentially very expensive in Germany. :/

Julian :arch: :linux: boosted

"If you are in academia, share your publications, share your teaching materials, share your code, share your data, share your time. If you have the opportunity, please help to facilitate more open exchange of talent and of knowledge."

#OpenScience #OpenData #science #data #research #share #exchange #idea

Show more

Fosstodon is an English speaking Mastodon instance that is open to anyone who is interested in technology; particularly free & open source software.