zerokspot.com

ScriptConf 2017 in Linz

Sun, 29 Jan 2017 20:11:31 +0100

Christian Heilmann giving his keynote about the state of the community.

Thanks to Netconomy, some friends and I had a lovely trip to Linz last Friday for ScriptConf, a one-day JavaScript conference. The event started at 1pm, which was extremely convenient for us as we could use the morning - and the beautiful weather then - to drive north from Graz, grab a quick lunch and be there just in time for the talks to begin! Definitely a nice advantage if you’re living in driving distance! But never rely on that as I learnt the hard way 2 years ago 😉

Anyway, back to ScriptConf! This was the first conference organized by this crew but you wouldn’t have noticed it. The schedule was near-perfect, the venue excellent, the catering great; in short: The organization was top-notch! Perhaps the only aspect I didn’t like (but that’s a matter of opinion) was the session length of 40 - 50 minutes. Personally, I prefer 20 - 30 minute-slots but that’s just me 😊

Usersnap, one of the sponsors, had also a beautiful plush alpaca everyone made selfies with in order to get a chance to win an iPad … or simply because it was so cute! Sadly, it didn’t want to accompany us back to Graz 😞

Now with all the meta out of the way, let’s get to the main attraction: The talks:

Christian Heilmann presented the opening keynote about the state of the JavaScript community using buffets as a metaphor. Criticizing some parts of the community he presented a basic etiquette for everyone inspired by how everyone should normally behave at a buffet and how to apply that here.
Next, Raquel Vélez from NPM gave an overview over how npmjs.com evolved over time: Like how they came to pick React and Hapi and the story about search on the site 😉
After a quick coffee-break, Nik Graf gave an introduction about the server-less architecture with a focus on orchestration and how the infrastructure around the containers are defined.
In case you were not up-to-date with Ghost’s road to 1.0, Hannah Wolfe definitely changed that. So many great stories about why certain things were done in a particular way!
Before dinner-break, Felix Reiseberg from Slack gave probably the best talk about Electron I’ve seen so far! And all that with live-coding!
Wanted to know more about how to profile your code on V8 with some additional knowledge about internal caches? Franziska Hinkelmann had you covered!
Last but not least came Mathieu Henri with a live-coded demo-scene’esque audio-visual presentation. Party before the party!

The moment the talks make their way online, don’t do cherry-picking. Just watch them all 😉

This amazing day ended with a nice little party were everyone got one last chance to relax, talk to some of the speakers and collect stickers (Big thanks to Raquel for the custom NPM stickers ❤️) before heading home. As with everything else, even this final schedule-entry was perfectly organized.

Selecting an OpenPGP key in MailMate

Sun, 22 Jan 2017 19:35:31 +0100

For pretty much all my not-work-related mail MailMate has become my favorite tool. It is fast, lightweight, supported OpenPGP through GnuPG 2.x, and has decent shortcuts. I also really like that even if the author hasn’t the found time to put a GUI around a feature’s settings, there are settings.

In a recent example I wanted to make sure that for one particular e-mail address a certain OpenPGP encryption key was used. MailMate doesn’t have any kind of GUI for managing keys but you can still choose which one to be used for which e-mail address through the ~/Library/Application Support/MailMate/Security.plist file. In there you basically just create a simple mapping like this:

{
	map = (
		{
			address = "<e-mail address>";
			userID = "<key id>";
		}
	);
}

You can learn the details deep down in the hidden-preferences section of the MailMate manual. Perhaps the only slightly complicated part here is learning the value you should put into the userID field. The example in the documentation uses the old long-form key ID, which is a bit complicated to get from most GUIs. Here it’s much easier (and also recommended by the gpg2 manpage) to go with the fingerprint (--fingerprint option) without spaces. As a fallback (for whatever reason) you could also use the 8-char key ID normally printed right in front of your key when using --list-keys.

Fixing macOS's CharacterPicker madness

Wed, 18 Jan 2017 20:45:25 +0100

Ever since updating to macOS Sierra I had a weird issue: the character picker (“Emoji & Symbols”) behaved erratically. I could only open it every couple of minutes and then it might appear out of nowhere whenever I focused a native text input field. Today I was close to simply re-installing the whole machine as I was also facing some other issues but decided to check first if the behaviour could be reproduced in a fresh account.

But before that I went through the list of third-party kernel extensions to see if anything might be relevant. kextstat | grep -v com.apple is really nice for that 😉 Just don’t kextunload any extension directly 😉

After removing some old extensions of apps I had previously uninstalled the manual way (and restarting as I definitely kextunloaded things I shouldn’t have), I tested if I could still reproduce the issue. Yep, CharacterPicker was still drunk. So on to creating a new dummy account and another reproduction-attempt.

Turns out, it was gone. Plans for operation “Nuke and rebuild” were cancelled and I started sifting through ~/Library. I even moved that whole folder but “Emoji & Symbols” was still hanging in the Activity Monitor. I dug a bit deeper and learnt that the CharacterPicker.app (which is just localized with that name) was launched or at least controlled by a background service:

launchctl list | grep Char
605	0	com.apple.CharacterPicker.FileService

Whenever I had that issue, that service would die a horrible death after a minute or so with the exit code -9. But during that time there was another service called com.apple.CharacterPaletteIM.<some number> visible in that listing. Googling for com.apple.CharacterPaletteIM on DuckDuckGo lead me to a support article by apple titled “Mac OS X Leopard: Keyboard Viewer, Character Palette does not appear”.

Hm… there are most likely no traces of that OSX on my current machine simply because I had a completely different backup-strategy back then. Anyway, I opted to follow the steps mentioned there after inspecting the mentioned folders first:

sudo rm /System/Library/Caches/com.apple.IntlDataCache*
sudo rm /var/folders/*/*/-Caches-/com.apple.IntlDataCache*

The latter wasn’t relevant as I didn’t have any matching files, though. After a reboot, everything was working fine again 😄 And yes, there was definitely some “Migration Assistant” in the past of my current setup. I’m just not sure if I used it for Sierra or El Capitan.

The end of App.net

Sun, 15 Jan 2017 01:07:35 +0100

Sadly, it looks like the end for app.net is near. According to this post by Dalton Caldwell and Bryan Berg the site will be shutting down in March. It has been in maintenance mode since 2014 after failing to get enough adoption and therefore financial funding for new features to be developed.

Now, it looks like even the funding for keeping the system up and running has run out and so it will be shut down. Given that I myself haven’t had a paid account for quite a while, I feel hardly in a position to complain. That being said, esp. in the light of what platform Twitter has become over the last years, any alternative should be welcome and worth saving.

As Dalton wrote in his post, the biggest issue of App.net was probably the chicken-and-egg problem. I really enjoyed the service but Tweetbot was simply the far superior mobile app and so I posted less and less and eventually quit altogether. Looks like I wasn’t the only one.

But that wasn’t the only reason why I stopped using the service. It didn’t solve another of Twitter’s problematic aspects: It being centralized. I haven’t yet seen any good solution here, but perhaps something like diaspora might in the end attract enough users to become more mainstream.

Perhaps the only good thing that’s coming out of all this is that we might see all of App.net’s code eventually and for the foreseeable future on Github. I’m still curious what the developers have been up to over the last years as their respective Github profiles look rather empty 😦

Goals for 2017

Sat, 14 Jan 2017 23:44:35 +0100

This year it took me even longer to come up with some goals for the new year 😊 2016 was a bit of a weird ride and 2017 set out to be even more of an extraordinary year for me judging by its first two weeks. As such I thought I’d try to change my focus for my goals a bit this time around:

Get a life

This is probably the hardest topic of anything in here to write about. Over the last couple of years I’ve prioritized work- and OSS-related topics over pretty much everything else, probably in part due to feeling a bit like an imposter whatever I do. This also put sleep extremely high on my priority list with the obvious consequences.

This year I will try to move things around a little bit. Slightly less sleep, slightly less being obsessed with the craft, more time for my friends (I love you, folks!) and for getting to know new people!

I’m not yet sure how this will involve or affect my moving plans/ideas but an effect it definitely will have.

Less grumpy

From time to time I tend to be a bit too grumpy when something annoys me (and quite a lot of things do). This is not helpful as it distracts from looking for a solution to what made me grumpy in the first place. For this reason I’ll aim at being less grumpy this year 😉

Deeper knowledge

You might have noticed some TIL posts appearing here from time to time. I want to do more of these as they help me digging really deep into certain topics, far below the point needed for actually resolving the original issue I had.

This will hopefully also lead me more down the OS, security, and hardware layers again.

Even more writing

Somehow related to that I want to do even more writing than last year. This isn’t limited to my blog, though, as I do quite a lot of documentation related work. Something that I’ve come to enjoy more and more ever since attending my very first WriteTheDocs event in 2015.

I’m not yet sure how I’ll integrate that into my career planning as I’ll probably still be a software developer but perhaps also with something like a “technical-writer light” add-on. For that I’d also love to look around for how formal education is done in that area and at least read some relevant books there.

Becoming an Angel and/or contributing more

Pretty much what I wrote last year. At least for 34C3 now I’m going to have a DECT phone as it arrived last week 😉

And regarding contributing: I want to look around more actively for things I could help out with. This isn’t limited to OSS or related fields but perhaps also some social initiatives. I’m not yet sure what exactly I want to do here, just that I want to do something; and if it’s just donating blood more frequently. Putting this up on my yearly-goals list should help with the motivation.

There are also some other topics that might become relevant over the year but that I don’t really see as primary goals:

Attending more diverse conferences. This is something I’ve started about two years ago and which will reach a new peak this year with me even skipping DjangoCon Europe.
More Linux in my life again. This doesn’t mean that I will abandon mayo’s completely but I want to slowly move in a direction where I could actually migrate away from it again 😊

Let’s see in 11.5 months how many of this I will have been able to achieve, but this time I at least want to aim high 😊

The Three-Body Problem

Sun, 08 Jan 2017 10:47:15 +0100

While I wrote in 2015 that I’d love to be part of a book club again, for some reason I didn’t really read all that many of the Sword & Laser assignments since then. At least I punished myself for that by listening to their podcast and hearing all about those awesome stories I was missing. In order to resolve this, in 2017 I’ll try to read at least most of their Scifi assignments starting with the one for January: The Three-Body Problem by Liu Cixin (in the English translation by Ken Liu). The fact that I’m already writing a review only a week after I picked this book up should already tell you something 😉 Turns out, this was one of the most interesting reads I’ve had in a while!

First, the story takes place in China and spans all the way from the Cultural Revolution to the early 21st century. I think this is the first book I’ve read taking place during that era and in that region so I was already hooked! The early parts of the book follow Ye Wenjie. Her father was killed during the first days of the Cultural Revolution and she is sent to a labor camp. Later thanks to being one of the last real experts in her field she gets the chance to work on a top-secret government facility tasked with the search for extraterrestrial life.

The latter half follows Wang Miao, another nanotech-scientist in Beijing who gets pulled into a huge mystery threatening the very basics of scientific exploration. While the story jumps between these two eras the time-line is still mostly linear. Once Wang is introduced, the story bits in the 1960s up to the 1980s are explored as accounts from some characters living in that era or through protocols.

The book mostly deals with classic Scifi questions like what would happen if a civilization learnt it wasn’t alone in the universe. That being said, the way this topic is handled is quite interesting as the focus here is on how a culture could be manipulated from the outside. While this may sound like something out of the X-Files, it certainly isn’t. Rather, what would it take to bring every scientific progress to a halt?

What also drew me in was the to-the-point writing style. Very little time is spent on decoration, still the world-building works quite well. And there is tons of building going on here! Yet the author doesn’t shy away from going hard-scifi in some places. Especially during the last chapters lots of time is spent on exploring topics like n-dimensional space. I cannot really write more without turning this post into a huge spoiler but I’ll say it this way: I simply couldn’t stop reading! Sadly, the book ends in a cliffhanger which is also pretty much the only unfavorable aspect that I could make out 😊

So far the series consists of three books. I’m pretty sure I will also read at least the second entry but not right away. First I should really make some progress with my pile-of-shame 😉 But if you haven’t picked up this book yet, please do so!

2016 retrospective

Fri, 06 Jan 2017 21:10:37 +0100

The absolute highlight of the year: My trip to Portland.

2016 was a bit of a weird year for me. I always had in the back of my head that I should finally move away from Graz but at the same time I wasn’t sure if I actually wanted to. This is something that will probably keep being an issue for the next couple of months 😉

Conferences and traveling

But in general, I think it was a good year. I, once again, went to tons of great conferences, albeit not as many as in previous years:

No idea, why I didn’t write about DotGo esp. since this was probably the most enjoyable event of that series so far for me!

That being said, the absolute highlight of the year was my trip to Portland for WriteTheDocs US and PyCon US in May. This city now has a very special place in my heart and I cannot wait to return in a couple of months.

PyVideo & contributing to OSS

There I also started contributing to the PyVideo project. Sadly, over the last months I’ve done far less than I set out to but I hope I can find some time during my trip to Brussels for FOSDEM (and ideally in the time before that) to work on a couple of things that have bother me for some time 😊

In general, I’ve tried to be more active when it comes to contributing code to various projects. Whenever I see something that is bothering me, I try to write a report and ideally even a PR for that. I’m still struggling with keeping a balance between contributing to OSS and my other hobbies (esp. reading and listening to podcasts) but I think I’m starting to see a light at the end of that particular tunnel.

Learning French on Duolingo

During the year I was working really hard on my Duolingo streak. Except for one or two accidental interruptions I managed to keep the streak going throughout the year and finished the French course right before New Year’s Eve. I’m not yet sure what I’ll do as a next step, though. I definitely want to avoid the situation I faced right after my final high-school French exam all those years ago, when I forgot everything within a month because of neglect.

Perhaps I’ll start reading French news or Wikipedia entries alongside German and English ones. As for Duolingo, I will most likely not continue practicing there for now, but perhaps I will return for a different language in the future 😊

Technology

Technology-wise this year has probably been the year of containers and Docker for me. By now I’ve moved pretty much every project I’m actively working on over to Docker (if possible) and am integrating it more and more into my local development workflow.

I’m not yet completely happy with how configuration is handled in those projects, though, as having plain-text passwords flying around in environment files is sub-optimal to say the least. There I’m currently looking into Vault.

With regards to other tools and technologies, not all that much has changed in 2016. Emacs is still my editor of choice after a quick sight-seeing tour in VSCode-land. Go, JavaScript, and Python are still my preferred languages and I don’t see that changing anytime soon. Sadly, Django hasn’t really resurfaced for me but React is becoming more and more important as the tool-chain around it matures.

Content ownership and privacy

Content ownership and privacy were also becoming even more important for me this year. With the road both Twitter and Facebook are both on, I’ve decided to prioritize my blog. That was also the reason why I spent half of 33C3 re-integrating a search engine and yearly archive into it after moving to Hugo earlier this year. Everything important to me will be reflected here (alongside these other channels).

I’ve started to rely on messengers like Threema and Signal (with WhatsApp as a last resort) more and more. That being said, I’m really not happy with the way this whole market is fragmented and how mediocre some of the provided UIs are. At least I could avoid things like Telegram so far.

My use of GnuPG has also gone up quite significantly. There are still far too few people I communicate with this way, but that’s another issue 😉

The world and everything else

Soldiers on the streets and in the airports of Brussels and Paris while I was there 😔
Trump, Brexit, right-wing populism, terror attacks all around the world… 😞
Austria has a (former) green president 💚
Big thanks to Jonny Cruz for the most uplifting tweet of the year and to WeRateDogs for at least one smile per day! ❤️
Overwatch ❤️
I managed not to play a single minute of Pokémon Go 😉

How about last year’s goals?

Last but not least, I want to review the goals I had set for myself last January:

Become an angel: As I already wrote in my post about the 33C3, I sadly didn’t angel this year. That being said, I just ordered a DECT phone so not having one will definitely no longer be an excuse 😉 ❌
More blogging: Well, last year I managed to write 60 blog posts! ✓
I didn’t move out of Graz last year. In the end, that was a good decision, though. Especially during last months of the year I got to know a few really nice people 😊 ❌
Re-learning French: Just in time for the new year I’ve managed to get through the core Duolingo course material. ✓
Contributing more: I’m not sure how I should measure this one but to me it feels like a success. Quite a few pull-requests were made, I tried to help here and there with documentation, and I donated more than during 2015. ✓

3.5 out of 5 (as I don’t really want to count me not moving as a complete failure) doesn’t sound too bad 😄 Now I just have to come up with some goals for 2017…

33C3: The 33rd Chaos Congress

Sun, 01 Jan 2017 08:57:56 +0100

It’s really hard to write about an event like the Chaos Communication Congress simply because it has so many facets. It pretty much boils down to you how awesome an event you make it for yourself. For me it is simply an amazing event with great people and just the best way to end a year!

And that despite me not doing all that much there, and perhaps even being somehow out of place. There’s just so much to look at, so many awesome communities presenting what they’re working on over the year, and let’s not forget the great talks and parties … ok, no parties for me since I’m not really a party-person, but I’ve heard they’re amazing 😊

If you want to get a great overview of what 33C3 was in general, please take a look at this great post by Elliot Williams over on hackaday.com.

Getting things done

Just like last year I wanted to work on a couple of projects. Originally, I had planned to do a bit of styling work on the pyvideo.org website and help Ulrich with moving pyformat.info to Lektor. Instead, I spent most of the first two days adding a search index to this blog using bleve and VueJS. Giving these two a try had been on my list for a really long time, so I don’t feel any shame here 😉

Right on the first evening we found a really nice and remarkably quiet table in the first floor of the CCH which was simply to inviting not to get any work done 😉

Talks

This was probably also the place were I saw most of the talks during Congress thanks to a really stable internet connection (compared to other events) and great work by the VOC (Video Operation Center). That being said, I still haven’t seen that many yet. Something I’ve started to correct the minute I got home on Saturday afternoon.

During the event my personal highlight was probably the “space-night” which consisted of five talks on the third evening in the huge Hall 1. From learning what ESA was working on to getting a quick tour on how large the known universe is, there was something for everyone with even the slightest interest in space exploration. I really can’t wait watching many of the other talks already available for streaming and download.

Everything else

What makes the Congress every year so special is simply the openess and welcoming atmosphere. You see something exciting? Just walk over to the person next to it and they will give you a tour!

Sure, there are some not so pleasant folks (I’m looking at those of you who critizise people simply because they don’t have GNU Linux as their primary OS), but they are kept in check and balance by the vast, vast, vast, …, majority of nice people there.

During day zero (the day before the Congress actually starts) I also stumbled accidentally through the kids area (which was still under construction) which basically consisted of a whole floor. But don’t get the impression that that’s the only area safe for kids. Except perhaps for some of the party areas I cannot think of any place were children (young and old 😉) weren’t welcome and looked after. While many other events (e.g. in the Python community) try to make it possible for parents to bring their children, Congress has IMHO perfected that.

Things left undone

Sadly, there were also quite a few things that I wanted to do but simply didn’t, for one reason or another:

Being an angel: Angels are at the heart of what makes Congress feasible. Hundreds of volunteers doing small or big jobs (all important) like picking up bottles, helping out the bars, doing session-chairing, … After last year I absolutely wanted to contribute more than just a small donation on top of the ticket price. But sadly, at first it sounded like you’d need a DECT phone in order to become an angel. Later, I learnt that wasn’t necessarily true but registration was already closed 😞
Visit more assemblies: These are the areas dedicated to hacker spaces from all over the world to exhibit their creations. There are also assemblies focused on specific technologies like Python, Ruby, or Rust where you usually find folks running local meetup groups. Sadly … probably the only one I visited was Aaaaaaaaaaaa by various Austrian (and neighboring) hacker-spaces. The model train was simply a great point to meet 😉
Side events: No parties (for reasons see top), no lockpicking workshop.

Well, hopefully, I can get all that off my list next year.

That being said, we found a really nice restaurant near Sternschanze with good beer and I also had korean food for the first time in my life. And it was delicious! Oh, and we managed to take a group picture of everyone from Netconomy attending 😉

All in all it was once again a great event and I’m really looking forward to whatever the CCC plans for next year. A big thank you to all the organisers and angels who helped make this such a great time. Sadly, Congress won’t be in the CCH for the foreseeable future anymore, as that building is being renovated.

Books in 2016

Thu, 29 Dec 2016 12:30:29 +0100

In the last Sword & Laser episode of the year Veronica and Tom mentioned a nice post by Iain Bertram on Goodreads where he analysed all the books he read during the last year. While he focused on the authors’ and main protagonists’ gender, I wanted to take a look at the reading formats and genres I had been reading throughout the year. Goodreads itself already has a very simple overview page but I wanted to also see some other metrics here.

Genres read in 2016

After a bit of spreadsheet magic here’s what I’ve learnt: I’ve read 20 books, with about 30% of them being technical and another 30% being science-fiction. I’ve even read 2 fantasy books. Most of these books (50%) were PDFs purchased on platforms like Inkshares, O’Reilly, or even the respective author’s website. The average book had ~263 pages with the longest one ending after 720 pages (the usual suspect).

Format distribution in 2016

I’m actually pretty happy with the mix of genres and the lead of non-Kindle books this year. Inkshares, especially, has become an important place when looking for new things to read. I’m pretty that won’t change in 2017.

Sadly, I’ve also started to amass quite a pile-of-shame with Cooking for Geeks and Learning Swift having been on my “in-progress list” since January and June, respectively. I’m still trying to make my way through them 😉

1-pixel Moon and the Solar System

Mon, 19 Dec 2016 22:12:54 +0100

An amazing journey through our solar system at scale!

Just while watching The Expanse (and thanks to BoingBoing) I stumbled upon this amazing little website by Josh Worth called If the Moon where only 1 Pixel. Here our known solar system is depicted on a scale where our moon is represented by just one pixel.

The whole journey from the Sun to the outer edge of the system is filled with clever comments and observations about our view of the solar system and the universe in general.

Among other thanks it visualizes extremely well how close Mercury is actually to the Sun and how much empty space is coming after Saturn… and pretty much everywhere else. No wonder it took so longer to discover anything beyond it.

And my personal highlight? The lightspeed-button in the lower right corner 💝

Automatic mode selection in Emacs

Fri, 16 Dec 2016 16:53:15 +0100

Depending on which file you’re opening in Emacs, you normally want to have different behaviour or different highlighting patterns to be loaded. This is something where you’d normally use the auto-mode-alist variable to bind a file name pattern to a major mode:

(add-to-list 'auto-mode-alist '("\\.js\\" . js2-mode))

Now I was in the situation where I wanted to use js2-mode pretty much everywhere close to JavaScript except for files that contained JSX. There I prefer web-mode or js2-jsx-mode for now. Turns out, the whole mode-selection process is extremely customizable and is explained in detail here.

At first I thought I might add file-local variables to each JSX file. That’s not really a good option, though, as it would force my mode-selection onto basically everyone else working on such files:

// -*- mode: web -*-

For what I had in mind, it should be enough to just take a look at the first line of a JavaScript file. If it contains something like import React ... web-mode should be used. That’s pretty much what the magic-mode-alist allows.

(add-to-list 'magic-mode-alist '("^import React" . js2-jsx-mode))

That’s obviously not really bulletproof (as that import might be in a complete different line or I’d be dealing with an CommonJS import), but for the most part it should do what I want 😊

During playing with these settings I finally also got a chance to give Wilfred Hughes excellent refine command a try. It made modifying the various something rather long alists extremely easy!

I also learnt about the normal-mode command which allow you to reset the mode-selection in a buffer back to what would be picked automatically.

Alert resolution timeouts in Prometheus

Sun, 11 Dec 2016 12:09:28 +0100

I’m currently using Prometheus for quite a few services and esp. in combination with Grafana and AlertManager it has proven to be an extremely handy tool. For instance, we usually have alerts for every single major component of a service. If this component becomes unreachable, alerts are sent to a specific Slack channel.

What bothered me, though, was how long it took for the resolution message to arrive. By default, it takes 5 minutes; luckily, though, you can customize this in AlertManager’s global settings:

global:
  resolve_timeout: 20s

This would set the timeout to only 20 seconds, feels much more usable to me given that most of our check intervals are somewhere in the 5-15s range and the alerts are set to something between a 10-20s range. I’m pretty sure I will be tuning this setting in the future but for now this should do 😉

I have no idea why I didn’t see this setting in the documentation the first time around, but now I’m glad I looked again (albeit using a detour through the source code 😉)

A Night Without Stars

Thu, 08 Dec 2016 17:26:37 +0100

It took me quite a while but I’ve finally managed to finish the second part of Peter F. Hamilton’s Chronicles of the Fallers duology. This post will contain spoilers so if you want to avoid these, here’s a quick tl;dr:

Same as with the first book and big parts of the Void trilogy this book takes place outside the Commonwealth on the world of Bienvenido, which is to some degree comparable to early 20th century Earth. So, no utopia 😉 Most of the story is all about the distrust between the people with and those without biononics and their fight against the Fallers introduced in the first book. And thanks to them no longer being inside the Void, there are no more mental super powers. That being said, there is new hope on the horizon due to some unexpected help from the Commonwealth…

If you enjoyed the first book or any of the Void ones, I’m pretty sure you will like A Night Without Stars!

Spoilers ahead

So, now to the details:

The first book ended basically with the planet being tossed out of the Void. The story here picks up about one hundred years after that when people have finally acclimated to a universe without providing its inhabitants free special mental abilities. Due to this change the powers that emerged at the end of the last book have tried to cling to their influence as much as possible, basically simply converting one regime into another. The post-Slvasta world felt very close to the governments of the Captains.

The whole world is controlled through a military system that somehow feels completely unchecked, being judge, jury, and executioner at the same time. On the receiving end of this system is the Eliter shadow-society. Basically Commonwealth-like Advancers they are treated with a lot of suspicion and hate by the successors of Slvasta and his revolution.

This whole inner-human conflict feels extremely real. The authors goes into quite some detail on what kind of “advanced interrogation techniques” are available to the military and police when dealing with Eliters.

Most of the first half of the book focused on that. The second half tries to catch up with bringing the Fallers into focus again. Personally, that felt a bit rushed. It all went from “Eliters are evil!” to “we are overrun by Fallers” within a couple of pages.

One the other hand, esp. these last 50 percent of the book where awesome. I especially enjoyed the trip to the southern pole and the planet-hopping afterwards. It’s just pure scifi with all the technical marvels you can expect from a Commonwealth book. The last 30% alone made me give the book a ⁵⁄₅ rating on GoodReads 😉

Just the final minutes of the conflict could have lasted a bit longer. Personally, esp. the end of Jenifa was a bit underwhelming. Sure, there were only two options: Either she would come around or go all-out mad. Still…

That also shows how well crafted all characters were. From Florian, Ry, Chaing, Essie, to the Warrior Angel and even Roxwolf. They all felt very real and full of stories. Heck, who wouldn’t sit in the prime minister’s chair if they got a chance and mess around, right Corilla?

There were simply all shades of grey here. Nobody was purely good or purely bad. As much as Jenifa was misguided and borderline-crazy, she still seemed to just wanted the best for her people. It was also great to see perhaps the most “evil” character not being in some kind of leading position. Usually, the antagonist is the head of some organisation, while in this case Jenifa is mid-tier at best.

This book also did a good job finishing up the story-arch. As interesting as Bienvenido was, I fell like everything relevant has been said about it and its solar system. IMO revealing so much in so few pages at the end of the duology was perfect, but at least all the major story-lines have been concluded.

I can’t wait for what will come next 😊

Apple Watch: First impressions

Sun, 04 Dec 2016 21:26:32 +0100

About two years ago, when Apple released their first smartwatch, I was pretty sure I wouldn’t get it any time soon. I had recently been burnt by Android Wear and didn’t think Apple’s competitor could be that much better justifying the big price difference.

Then, in September or October I started too look into smartwatches again. The concept was still very appealing to me since I’m on a bike basically every day and having a notification and navigation interface available without having to get my phone out at every corner was great with Android Wear. What drove me away from that was mostly quality concerns on every level. The hardware was mediocre as was the software.

So I read up about the new newer Pebble Watches (also does that had just completed the latest Kickstarter at that time) and Apple Watches. And it would have been quite a hard decision were the new Pebble Time 2 available now. I had also seen the Series 1 Apple Watch thanks to a friend and therefore knew mostly what I was getting myself into and for that and a handful of other reasons opted to go with Apple’s product.

A little over a week ago I finally got my Apple Watch Series 2. Naively, I opted to install all the apps I had on my phone also onto my watch. That made for a really bad first impression as while the watch kept running for at least 18 hours, it drained my phone within 6 hours (which normally lasts for 2 days). I hadn’t thought about that but once I had removed all the apps that I didn’t want on the watch anyway, thinks improved significantly.

Now, the watch’s battery lasts for about 4 days and doesn’t affect the phone at all anymore. After 2-3 days I also finally started to understand the activity meters and am really enjoying them. They kind of feel like a more wholistic approach to the step-counter by Fitbit I’ve been carrying around for the last couple of years. There are some articles out there that try to explain the differences but for me it simply boils down to “stand up” meaning standing up and moving about 😉

Weirdly enough, I haven’t used the Overcast remote app all the much so far, but it’s pretty much what I wanted so far. That’s also a nice summary of the watch so far. It is extremely well-made on the hardware side while the software is good. There are some things I’d have liked to be handled differently mostly related to the mail app, but only minor things.

In general, though, I’m very happy with it. A rugged version would be awesome, though 😉 I thought about getting one of the cases available but none are optimised for the Series 2 so water all of a sudden becomes an issue again. Well, perhaps next year 😊

TIL: Directory modification times

Sun, 27 Nov 2016 08:13:48 +0100

Recently, I needed a way to check when the content of a folder had last been updated. If it hadn’t seen any changes within the last couple of days I wanted to remove it.

One solution for this would be to recursively check the modification time of every single file in that folder but I also got curious about what actually defines the modification time of the folder itself.

As this is something that can be different from file-system to file-system, let’s look at some of the more popular systems for macOS, Linux, and Windows. All of them are basically handling the mtime of a folder in pretty much the same way:

The mtime of a folder is only modified if a file directly below it is added, renamed, or removed. Changing anything about the content of any of the files or their attributes won’t change the mtime of the directory that contains them (Source).

To verify this I’ve created a little Python script and tried it on my local machine (macOS), a Linux box, and a Windows box.

There still seem to be some differences between the various file-systems, though, esp. regarding the amount of precision they provide.

For HFS+ (macOS) it was pretty easy to find some details in the specs that specify the basic behavior:

contentModDate: The date and time the folder’s contents were last changed. This is the time when a file or folder was created or deleted inside this folder, or when a file or folder was moved in or out of this folder. See HFS Plus Dates for a description of the format.

attributeModDate: The last date and time that any field in the folder’s catalog record was changed. An implementation may treat this field as reserved. In Mac OS X, the BSD APIs use this field as the folder’s change time (returned in the st_ctime field of struct stat). All versions of Mac OS 8 and 9 treat this field as reserved. See HFS Plus Dates for a description of the format.

[…]

HFS Plus Dates: HFS Plus stores dates in several data structures, including the volume header and catalog records. These dates are stored in unsigned 32-bit integers (UInt32) containing the number of seconds since midnight, January 1, 1904, GMT. This is slightly different from HFS, where the value represents local time.

Compared to more modern systems the mtime et al. are only exposed in seconds, while others are exposing nanoseconds. That also shows when you look at the st_mtime_ns value in Python (inspecting one of the source files of this blog):

>>> import pathlib
>>> pathlib.Path('32c3.md').stat().st_mtime_ns
1473583000000000000

ext4 (Linux) on the other hand offers nanosecond precision through some additional fields which are then merged on the API layer for Python et al. to use:

>>> import pathlib
>>> pathlib.Path("README.md").stat().st_mtime_ns
1457287154109435711

On Windows NTFS seems to handle modification times in a similar way to ext4 and HFS+ according to the Description of NTFS date and time stamps for files and folders document, but I couldn’t yet find anywhere in what precision that value is actually stored. It is exposed as nanoseconds according to File Times and also the little Python snippet returned a result similar to ext4.

While looking into all that I also learnt something about how the different file-systems handle things like creation time … a whole other can of worms, though (some details).

Personally, I’m just glad that mtime doesn’t seem to come close to the configurable nightmare that atime is.