Descriptor

New blog domain: kula.blog

2019-03-11T16:48:00.001+01:00

It's still work in progress but new posts will be published on https://kula.blog/

Thanks to 11ty base blog RSS is working from the start so it should be easy to add to your reader if you still use any :)

If now then I hope you can sign up to the newsletter where I'll be publishing new posts and interesting articles to it from time to time.

Please use kula.blog from now on.

New Domain

2019-02-10T10:54:00.002+01:00

I will be moving this blog to a new domain and may lose some of the subscribers if you're using RSS for example.

I hope that the email list can help with the migration (modal + sidebar).

In the future, I also think about sending interesting articles and news about tech and books, but all the things that interest me :) Maybe that's something that you would find worth your effort to subscribe.

>>> Signup page <<<

SJW

2019-01-25T14:35:00.000+01:00

Recently I'm reading a lot less about tech and more about culture and have a bunch of things I can recommend:

On my todo list:

The Coddling of the American Mind

As you probably can guess my view changed in the last couple of years on a bunch of topics. I still consciously try to remain in the middle in the discussions on the topics like `women in tech` as I don't want to be associated with actions of the most devoted believers.
For example, GitHub's ElectronConf postponed because all the talks (selected through an unbiased, blind review process) were to be given by men. is far from How blind auditions help orchestras to eliminate gender bias.

Sprint stretch goals are harmful.

2019-01-24T14:36:00.000+01:00

I was listening to a podcast episode: Command Your Focus and Attention on What Really Matters with Chris Bailey and noticed a really good analogy.

"What allows traffic to move forward at a solid space is how much space exists between the cars on the highway"

What struck me about it is that I can use it to explain team productivity in different organizations I was a part of. Some of them tried to pack the sprint with tasks to the full, but in some, it was ok for the team to pick a 'safe' amount of tasks so that they could be sure to accomplish them.

Before I was thinking about the benefits of slack time that allows the team to work on important things (not only urgent) and to have space for urgent things if they show up. The great feeling of accomplishment when you finish last scheduled task day before the end of the sprint and can tackle additional bug or two or the confidence to spend additional hour making code more maintainable before finishing the task.

Now with the traffic analogy, I think that what whole team or even organization productivity depends heavily on not taking too much work to do. Creating space between tasks allows work go smoothly - otherwise, you end up with hectic fire-fighting and you're always behind schedule.

Another important effect is on creativity. The only way to make sure that humans can be creative is a playful environment. It has to be fun. In my experience, too many tasks and too high expectations create additional stress that kills creativity when it's most needed.

More reading:

https://www.quora.com/Why-is-it-that-in-traffic-jams-some-drivers-leave-a-huge-distance-between-themselves-and-the-next-car - Please take a look at the comments - you can notice people arguing against it as it's not intuitive!

https://www.thekguy.com/surprising-benefits-slack.html

http://trafficwaves.org/

5 things I learned from "How to train your BASH" by Marcin Stożek "Perk"

2018-11-28T14:24:00.003+01:00

Today I've watched a presentation by Marcin Stożek called How to train your BASH and wanted to remember the most interesting points for later.

1. It's best if you follow a style guide

There seems to be only one though Google Shell Style Guide.

2. You can set Bash to "strict mode"

It sets more human-friendly defaults and makes sure you won't miss any errors in your scripts.

#!/bin/bash

set -euo pipefail

IFS=$'\n\t'

3. Shellcheck

Turns out there is a tool called shellcheck that can help you write scripts that you intended to write. Best thing is that it explains what's wrong and gives you suggestions on how you can fix it.

4. You can enable 'debug mode' when running your script

Just pass "-x" to the script eg.

> bash -x ./my_script.sh

5. Use Pure Bash

No one wants to see weird errors when running the script on a new server so it's best if you avoid using external programs in your scripts.

There is much more in the presentation so take a look at it: https://slides.com/perk/jak-wytresowac-basha#/

You want to stay relevant as a software developer for the next 10 years?

2018-11-20T13:07:00.003+01:00

This tweet made me remember my own pondering of the exact same question and how hard it is to guess what is going to be relevant in the future.

A lot of people answer that by saying that you have to always learn but the question is not "Should I learn?" but "What I should learn?".

Unfortunately, no one can predict the future. I would even say that the question itself is flawed.

I'm reading Farnam Street blog for some time already and one of the last posts goes deeper into this problem: What’s Staying the Same?

If you start looking for things that are likely to stay the same you will find a lot of things that are worth learning and they should be relevant a lot longer than the newest fad.

Book Review: Algorithms to Live By: The Computer Science of Human Decisions by Brian Christian and Tom Griffiths

2018-07-10T10:20:00.000+02:00

I liked the idea of the book form the first time I've seen the title and fortunately, it lives up to my expectations. It's not perfect, but it rekindled my curiosity towards math and algorithms. Bringing them to everyday problems makes them much more interesting.

The only thing that I'm missing in the book is the application of algorithms in business settings. For example, when I mentioned this book Mateusz K he has sent me a blog post of Pieter Hintjens: Amdahl to Zipf: Ten Laws of the Physics of People. For example, Amdahl's Law mathematically proves the argument from The Mythical Man-Month that you can't just add new people to the project to deliver it faster because of the cost of synchronization.

I want this post to be a reminder for myself of the most interesting ideas in the book so if you want to savor it yourself just remember that it's worth buying.

Rating: 9/10

Probably the most interesting from this chapter was the rule of 37% that's optimal in the "secretary problem". Given some amount of time for interviews, you should spend 37% of them only on exploring. After exploring you should pick the first option that is better than anything from the exploration phase. In real life, it's more complicated than 37% but the rule to only explore in the beginning was still interesting.

2. Explore/Exploit

The most important question when deciding between staying with something that's working or looking for something new is the timespan. How long will you use it? How long are you going to live? For how long will you be able to exploit this thing?
If you're young or early in your career then you should explore a lot (like small children do) - exploration has a cost but some of the new things may be way better than everything you've seen so far.
When you're old then stick to the things that work, things you already like - it's not worth wasting time and effort to learn things that you won't have time to use.

A simple strategy to use: Minimise regret
The sad thing is we're going to gain regrets even when we choose the optimal strategy. Minimizing regret should get us fewer regrets each year in comparison to the previous ones - so it's the best you can get.
The optimal strategy is to pick new opportunities with maximum gains - or "upper confidence bounds". If you're going to gamble it's best to gamble on things where you can get the most out of it. For me, it was really hard to wrap my head around as it means that when in doubt you should be an optimist.

Another idea I first heard about is "Adaptive" trials. You might have heard about A/B tests where you check two things at the same time to compare results and pick the best option. It's simple but what if one of them kills people? When you test new drugs or medical procedures and see that one option is saving lives and another is killing them is it still moral to continue with the rest of the trial?
In this situation, you can use another simple strategy called "Win-Stay, Lose-Shift". As long as it works keep doing it but switch when it stops working. In medical trials, it means - shift when someone dies.
Overall people explore too much - at least in the lab - but it seems likely it's because we live in the restless world - it constantly changes so you can't be sure if the wrong option from yesterday is still worse or maybe something has changed?

3. Sorting

I've learned that computers were created to sort things. When you search for something - it's still dependent on sorting (imagine best search results being on a 100th page on google).
It's inefficient to search an unsorted list/stack but it's a waste of time to sort things that you're never going to search.
Rule of thumb: do not sort.
There are experiments proving that sorting emails into categories is a waste of time. It's better to search for them when needed.
In many sports (like football or soccer for US readers) the second best team is pretty much random. Only the best team really is the best. The easiest way to make sure that the second and third place are real is for everyone to play with everyone - unfortunately, that takes too much time so not many sports do it.

The best strategy is to avoid direct comparisons like matches and "pecking order fights" to establish the hierarchy and switch to races when something like time can be used to judge who/what is better. Direct comparisons take a lot of time.

4. Caching

Least Recently Used - LRU is the best caching strategy you can use. It's pretty close to clairvoyance and you can use it even for everyday things like what to keep at your desk, nightstand or in the wardrobe. People aren't like computers so by using it strategically you can make yourself make things more often just by placing them near you.
Noguchi Filling System - it's LRU but for documents or just physical stuff. You have a shelf with a list of documents. If you want to add a new document or put one back - always put it on the left. New items and the items recently used end up on the left-hand side of the shelve - the order takes care of itself.
It turns out our memory works exactly the same: Forgetting curve is basically LRU - memories you use often are kept closer - everything else moves into oblivion. Most of the time it's a good thing and it's worth to remember that when you can't remember something it might be the case that it's just far away and searching for it takes time - you may still remember it just fine just the sheer amount of memories makes it hard to find.

5. Scheduling

If you have 100 tasks with deadlines you can decide to do as many as possible or to limit how much they are going to be late - different goals have different optimal strategies.
Personally, I found the idea of Thrashing as applicable to everyday tasks the most intriguing. If you have too many tasks just managing them will take a lot of your limited time. At some point, management itself will take most of your time. It seems a little crazy but everyone can feel it from time to time when you feel overwhelmed before starting a really big project.

The most interesting strategies I've picked up are:

Do the short tasks first to minimize the number of tasks.
People have long context switching cost - try to group tasks into 90min blocks.
Try to group small/easy/short tasks together to avoid switching your focus too often.
If the task takes 2 times longer do it first only if it is 2x more important.
Decide how responsive you want to be eg. check emails once a day and don't break it to avoid context switching costs.

6. Bayes's Rule

To tell how long it will take (or exist or live), you have to decide if you know how long things usually take.
If you know the average e.g. average life expectancy then it is your answer. After the average eg. if someone is older: add a couple more years to existing age at that is your answer (Bayes's Rule with Prior Belief).
If you don't know anything then it's best to assume that you're in the middle of its existence. For example how long a building is going to last? If it's already 30 years old then your best guess is another 30 years (Copernican Principle).

In our lives, we overestimate events that we hear a lot about - if you're watching a lot of News then you're going to overestimate the frequency of mentioned events eg. plain crashes.

Marshmallow test - which I thought was really about delaying gratification might be more about expectations. It's easy to assume children should wait for the second marshmallow but the test assumes that the future is predictable and known. If the child thinks it's likely that you're going to break the promise then it doesn't make sense to wait.

7. Overfitting

I really like the subtitle of this chapter "The Case Against Complexity". Complexity has cost and it might even be damaging. I'm pretty sure it should be taught to IT students instead of Design Patterns.

Given data in the chart (taken from Wikipedia) notice which function (line) get's you closer to the underlying phenomenon. Around -4.5 there is a huge spike in (blue) more complex function caused by overfitting. The polynomial function tries too hard to follow the data - it's no longer useful to describe whetever data points were measuring.
We should penalize complexity. In real life use the Occam's razor - always pick simples explanation. Use Simple Rules even if they aren't perfect as they use less cognitive resources and make your choices easier.

When things change then it makes sense to keep some old genes or traditions even if they don't work that well at the moment. If there's a big change back to the old ways then we won't die off that easily. Conservatism is a little bit like evolution - both are restraints for overfitting.

Use "Early Stopping" when you don't know much about the problem. Only pick one most important metric to avoid overfitting for barely important that can mess the whole model.

8. Relaxation

Questions like "What would you do if you didn't have to work?" are used to solve intractable problems. Turns out some things are just hard and by removing constraints you can get closer to the best answer.

Things to try:
1. Remove the constraint
2. Instead of yes/no try assigning a number between 0 and 1. eg. 0.4 and 0.6 and compare your options.
3. Ask yourself what's really the cost of breaking the rule? How much speeding ticket does really cost?

9. Randomness

Instead of understanding the problem sometimes it makes sense to just sample a lot of data.

It's the first time that I learned about Bloom filter - which is a set that has an error rate of around 1-2% to save time and space of keeping the data.

There was a lot in this chapter about shortest paths and avoiding local maximums with strategies like Simulated Annealing but I'm not really sure how I could use them in practice outside of programming.

10. Networking

This chapter starts with the problem of the Two Generals - which shows how hard it is to coordinate using a medium that can fail. It was a good reminder for me that TCP uses "triple handshake" to acknowledge packets in the networks.
Another reminder was learning about early problems of networks caused by new users and how exponential backoff fixes it by cutting the speed by half when it notices any lost packets. Finally, I understood why it's cut by half - if there are only two users in the network - the first one should leave half of it for the new joining user.
Exponential Backoff can a problem when you want to serve a fast website but it's still a useful algorithm. I was surprised that it's useful in fighting crime - give offenders increasingly long sentences (starting with one day) instead of a bunch of warnings followed by a really long sentence.

Exponential Backoff can be useful to fight with Peter Principle which observes that people in a hierarchy tend to rise to their "level of incompetence". To fight with it organizations use rules like up or out, but it's highly ineffective to let go of people that were highly effective on the lower ladder of the organization. Exponential Backoff suggests downgrading them instead of letting them go. It sounds much more humane to me.

I liked the description of bufferbloat. It's not a good thing that you're router has so much memory - everything gets slower because of it as it keeps packets for too long. Sometimes never is better than late and big buffers are breaking this rule.
In general, it's a case against our devices remembering everything for us - for example emails when you're on vacation. You get overwhelmed by them when you're back. Buffers like those induce "lack of idleness" and raise our stress levels.

11. Game Theory

The biggest gain for me from this chapter is learning about Reverse Game Theory. The easiest way to explains it is with the tragedy of the commons - if something is for everyone the dominant strategy will ruin it. The best way for everyone (paradoxically) is to introduce a penalty from the outside. It's about creating new rules to get the results that everyone wants.

Beware of Information Cascades - if you're looking at others to know what to do - others are probably looking at you to decide what to do themselves.
Stating your contrasting opinion may save the whole group just because it creates contrast.

All in all, it's a really good book. Hope that my notes will be helpful.

Big thanks to Mateusz K for reviewing this post.

Book Review: Masters of Doom by David Kushner

2018-07-02T15:47:00.002+02:00

The book that I wanted to read for a long time. How was it?

I have to say that it's really good. I learned a lot about Id and both John Carmack and John Romero - how their lives looked up before and during the creation of the most know Id Games like Doom or Quake. What was depressing but also making it so much more real is how much politics and infighting happened between all those people working in the industry at that time.

The only downside for me is that there's not much about the technology side of things. There is a paragraph here and there but most of the book is about two John's allowing you to take a look inside their heads. Maybe another downside as not native speaker would be the font - that made it harder than necessary to read.

All in all, I highly recommend it. 8/10

I've spent an hour debugging XSRF error because of ... fetch

2018-05-09T11:08:00.002+02:00

Photo by Tim Gouw from Pexels

Yesterday we were working on a new subpage with a form. Let's use the new `fetch` API to make POST requests seemed like a good idea but we started getting errors from the server:

`HTTP 403: Forbidden (XSRF cookie does not match POST argument)`

Header `X-XSRFToken` was set, we tried passing value in the body, setting token in the template, setting it in the backend only in certain cases but nothing helped. Finally my coworker found out that we're not sending `_xsrf` cookie with the request at all so it generates a new one every time...

Turns out fetch doesn't send cookies by default. It's by design: https://developer.mozilla.org/en-US/docs/Web/API/Fetch_API/Using_Fetch

The fix is incredibly simple: tell fetch to include credentials in the options:

```
fetch("/someURL/", {
method: "post",
headers: {
"X-CSRFToken": token
},
credentials: "include"
});
```

Book Review: The Art of Manliness: Classic Skills and Manners for the Modern Man by Brett McKay

2017-08-19T19:39:00.000+02:00

I'm a a listener of The Art of Manliness podcast for over a year now and just before my last vacation I decided to buy the book. Partly as a support for the podcast and partly to get my own opinion about it (reviews are really all over the place from 1 to 5 starts on Goodreads).

If you read most critical reviews you will find that Bret McKay draws a lot of inspiration from 19th century gentleman. You can think about your grandfather more than your father kind of man being the inspiration for this book. There is a bunch of practical advice on all kinds of topics from shaving, dressing to raising kids but also a lot of trivia and random facts you won't really need so you have to pick yourself.

There is one annoying quirk of this book though, it tries to explain things with words that should be explained with images. I have skipped most of tying knots and ties for that reason. I've noticed that there's a new book called The Illustrated Art of Manliness: The Essential How-To Guide: Survival • Chivalry • Self-Defense • Style • Car Repair • And More! by Brett McKay which I think is a solution to this problem.

The Art of Manliness introduced me to the idea of man's classic elegance and traditions and it makes it a worth reading for that reason but if you find something interesting then you have to dig dipper by yourself.
One of the ideas in the book is to use safety razor as a traditional method of shaving. Because of this chapter I started reading more about razors, shaving and decided to give it a try (I always used electric shavers). I have to admit that I just found it interesting.

Rating: 7/10, but I would recommend it to all man - especially if you're young.

Book Review: Sapiens: A Brief History of Humankind by Yuval Noah Harari

2017-08-14T15:58:00.002+02:00

That's my first review in a pretty long time, but I decided to only review good books - the kind I'm recommending and giving to my friends.

I heard about "Sapiens" a couple of times but I bought the book after learning more about the author and seeing that Cal Newport recommends the book as well. You can read about the author at Yuval Harari Works Less Than You maybe you will get interested the same way I did.

Sapiens starts when humans were hunter-gatherers and brings us to modern times. Most important is his perspective and high level view on agriculture, religions, ideologies and nations. Most striking was how much modern culture we live in is just another kind of ideology and how important it is for people to believe in the same things.

Rating: 10/10. Must read.

I was reading it in english but I liked the books so much that I bought Polish paper edition as well. Polish readers can find shops at: lubimyczytac.pl

Pomodoro Yo-yo Effect

2017-03-09T16:04:00.000+01:00

Recently I started feeling stressed and less productive which forced me to be more conscious of my work, habits and everything around it. Thanks to that I noticed:

I got a lot of notifications from Slack, Messenger and Email.
I felt anxious before starting new tasks.
My focus wasn't predictable, especially on harder tasks.

Don't get me wrong, I felt great most of the time, especially when I sat for hour or two at a time coding. The troubling thing was that it wasn't every day and I felt it was getting worse.

Most of those problems should be fixed by Pomodoro techinique I was writing about before, did it stop working? The thing is, I stopped using it.

A couple of months ago I noticed that staring new tasks, and taking breaks is actually a hindrance. I could easily focus for an hour or more on a task and basically forget about Pomodoro, and feel guild about not following it or not keeping proper count. I've decided that it wasn't necessary anymore. It did it's job and it was time to move on.

Unfortunately the world around me was changing all the time. I think most of it is more people joining our project. It's not a problem in itself but it means more emails and more Slack messages. The problem was that the change was so subtle that it took me a pretty long time to notice and I was sliding into my old distractions habits one small step at a time.

So what do I do now?

I've started by disabling most of the notifications. I still get Direct Messages and announcements on Slack, but it's a lot less distractions. In case of email I disabled all of them. I check email app every 1-2 hours but at least I read 2-3 new emails at a time and it's when it's convenient for me.

I'm back to Pomodoro as well. I was thinking what to do about 25min time limit (optimal would be around 60-90 minutes), but my back ache forced me to look at it from a different perspective. From what I remember from reading Healthy Programmer, every 20 minutes we should let our eyes rest (look at something far away for at lest 20 seconds) and switch between sitting and standing position. I don't have a standing desk right now but I try to at lest stand up and stretch a little.

Of course it's still not perfect, but I already feel it's helping me to get more done.

This whole story makes me think of the Yo-yo effect and I'm curious if anyone has similar experiences?

One more thing: I was fortunate to stumble upon a really good review of Deep Work and Flow at the same time: https://www.robinwieruch.de/lessons-learned-deep-work-flow/. For me it's great as a reminder but I highly recommend if you want to know more before buying those books yourself.

Image source: https://www.pexels.com/photo/biotechnology-bright-chemical-chemistry-207585/

Book Review: Influence by Robert B. Cialdini

2016-11-23T14:36:00.002+01:00

I published a new review on Medium.com: https://medium.com/@krzychukula/book-review-influence-by-robert-b-cialdini-f4575b6786a4#.55i8vpylc

Morning Paper Series About Architecture and The Rise of Worse is Better

2016-10-05T11:11:00.001+02:00

I was reading a really good series of articles on 'the morning paper' that I found myself going back to almost every week. Take a look if you don't know them but this list is mostly for me to make it easier to reference them :)

https://blog.acolyer.org/2016/09/07/the-emperors-old-clothes/
https://blog.acolyer.org/2016/09/08/computer-programming-as-an-art/
https://blog.acolyer.org/2016/09/09/reflections-on-trusting-trust/
https://blog.acolyer.org/2016/09/14/why-functional-programming-matters/
https://blog.acolyer.org/2016/09/15/the-rise-of-worse-is-better/

Small improvements to keyboard layout

2016-07-22T16:00:00.001+02:00

There is a new post on my medium account about new experiment I'm starting and how to setup it for yourself: https://medium.com/@krzychukula/customizing-keyboard-layout-on-mac-572327b031dd#.w6hcwik6z

TLDR:

Remapping Caps-Lock to Backspace and disabling standard Backspace
Switching Right Command and Option(Alt) for writing in Polish

Book Review: Deep Work by Cal Newport

2016-07-05T13:48:00.000+02:00

I've posted the review on Medium - mostly because the editor is much better and some people have problems reading on Blogger.

https://medium.com/@krzychukula/book-review-deep-work-by-cal-newport-97b15ef26265

The Strange and Marvelous World of Keyboards

2016-07-03T19:50:00.003+02:00

Estimated reading time: 7 minutes.

Not long ago while watching one of the presentations of Andrew Sorensen (I recommend The Concert Programmer as well by the way) that he is using a really funny looking keyboard. I started googling it and found out the strange and marvelous world of keyboards that I had no idea existed.

I want to note here some most notable things that I had learned and maybe you can learn something in the process.

At first I was exploring Keyboards, then started exploring keyboard layouts and ended up back in the hardware land at the end.

So let's start with the keyboard that inspired this whole quest: Kinesis Advantage

I found this video review to be most useful to get an idea how it looks and works.

Mini review:
+ It's ergonomic with mechanical keys (that's a good thing)
+ Your thumbs can do something useful instead of just hitting on 'space'
- 'Esc' and Function keys are really small and cheap.
- No arrow keys
- It's really expensive...

Then I learned that mechanical keys are way better than 'standard' keys, but they also cost much more and are 'clicky'. That means that they make a lot of noise while typing.
There are also different kinds of them with different characteristics. It's best to hear and feel it so go to the shop or watch a lot of comparison videos like this one to get an idea:

I realized that one of my coworkers is using the mechanical keyboard and while I was checking it out he told me that he uses Colemak for writing in English. Colemak is more efficient for typing English, but he was switching back to QWERTY for Polish.
From that moment I was really eager to tray it myself and get some experience with it, but I wanted to use the same layout for Polish and English no matter what.

Turns out there is even layout optimized for Polish called klawa.pl with some impressive statistics and really cool positions of keys, but the problem is that I type a lot in English. Maybe 50% maybe more so I started researching which of the English layouts would be best for Polish.

From my small samples, Colemak was actually much better than the newer ones like Workman and I even found layout that enabled me to use only one Alt for Polish letters: https://github.com/michalzajda/colemak

I can recommend a really nice site for learning alternative layouts if you are interested: http://thetypingcat.com/course/colemak Best feature is that it can emulate your new layout without any changes in you OS or keyboard so it's great to learn and get a feel of the layout.

While I picked Colemak I wasn't sure about it and asked the creator of klawa.pl for some help.

Thanks to dooh.pl I get the list of English layouts when used for Polish:

1. KLAWA.PL 56.95

2. DVORAK PL 52.87

3. COLEMAK PL 52.10

4. WORKMAN PL 48.93

5. QWERTY PL 41.93

I didn't expect Dvorak to be so high on the list so I asked one of my friend who I knew started using Dvorak 2 years ago for some feedback.

There is a lot of posts all over the internet claiming that it's great or that QWERTY is good enough but one of them really resonated with me: https://blogs.msdn.microsoft.com/jaredpar/2009/01/12/dvorak-keyboard-is-it-really-faster/. It has a really nice point of view so I would recommend it to anyone thinking about switching to the different layout.

Are you curious about my friend how picked up Dvorak? He gave up on it - basically, it's too inconvenient to work in a company and use a different layout.

After a couple of days with Colemak (and being really slow), I was impressed with the way some of the keys were placed on the keyboard. A lot of words started to be much easier to write - even in Polish. The downside was that it was really annoying to switch to type ";JK". I have no idea why these three were so annoying but I felt that their new positions are slowing me down a lot.

In the end, I gave up on Colemak for now and started reading more about Ergonomic Keyboards and Keyboards in general.

If I start working a lot from home at some point in the future I plan to revisit these layouts but instead of Colemak, I want to go with Programmer Dvorak. This layout makes a lot of sense to me and I can at some point forget about QWERTY all around me it would make a lot of sense.

In the meantime, I want to play more with remapping keys on the Mac and maybe use Karabiner to get some of the benefits of Programmer Dvorak without the chance of the letters layout.

So back to the Keyboards.

There is this one page you have to read through if you want to learn more about ergonomic keyboards by Xah Lee.

The page that will give you most information is How to Chose Ergonomic Keyboards. It's great and you have to read it.

Unfortunately, the problem is when you have a limited budget but still want to get the mechanical keyboard. Basically, you have to pick Microsoft Sculpt Ergo or forget about ergonomics (but pick "tenkeyless" keyboard).

It's really worth reading his post about really small keyboards as well: http://xahlee.info/kbd/idiocy_of_happy_hacking_keyboard.html What I learned is that a lot of keyboards are really overpriced and badly designed at the same time.

If you have to pick unergonomic keyboard it's best to pick one without numbers on the right (tenkeyless) - of course if you're not using them so that the distance between your hand position when typing and the mouse is minimal.

It's best to have all the keys, especially arrows for the keyboard to be useful in all situations.

Space should be as short as possible - it's best if you can click on Alt with your thumbs. Unfortunately, there aren't many keyboards like that and they all are too pricey for me.

So how should I pick a keyboard? If you don't have a budget limit the no problem but for me I can't afford ergonomic and mechanical keyboard at the same time.

Right now I'm mostly thinking about one of the two:

1. CM Storm QuickFire

2. Microsoft Sculpt Ergonomic

Both of them has their own pros and cons:

Sculpt:

+ It's ergonomic

- It's more expensive

- Not mechanical

- Requires USB Connector instead of just blue-tooth

- Comes with the mouse I don't need

CM:

- Not Ergonomic

- Cable

- Looks like it would move mouse more to the right

+ Full size function keys

+ Mechanical

+ Cheaper

+ Easier to learn

I have no idea which one to choose at the moment but I know that I want to try for myself if ergonomic keyboard makes any difference for me and if I like mechanical switches. Picking any of these makes sense.

Later this year or maybe in 2017 I want to see how new ergonomic keyboards are doing and maybe but something more ergonomic. Most notably:

If I forget about ergonomics:

Ergonomic:

http://www.kinesis-ergo.com/shop/advantage-for-pc-mac/
https://www.trulyergonomic.com/store/index.php (Xah Lee uses it)
https://ergodox-ez.com/
https://input.club/devices/infinity-ergodox (but you have to assemble it yourself)
https://shop.keyboard.io/ - I really like the idea behind this one

There is another thing that most of the ergonomic keyboards have - keys are in straight columns.

Nice example here is TypeMatrix:

The thing is that "staggered columns" layout is there because of typewriters:

Can you see how the keys are attached? Each row had to be moved slightly so that it would be easy to assemble it and there won't be any collisions between the keys.

Truth be told it always makes me crazy when touch typing - especially with my left hand.
Consider typing "R" you have to move your first finger up and left from "F". It's crazy uncomfortable and I still remember more that 10 years ago how I thought it was crazy uncomfortable that I have to learn to do that. Unfortunately, I thought that it's probably just me...

Promising:
I really want to check out Keyboardio, but it's so expensive that I want to see some reviewers before I decide to buy one.
What I like about it:
+ Keys in columns
+ Function keys for palms
+ Ctrl-Alt etc. keys operated by thumb
+ Splittable
+ Mechanical
- Expensive
- No F1-12 on arrow keys

I don't care about wood, colors etc. that make it really expensive so if everything goes well I hope to reevaluate it next year.

Another one worth trying would be ErgoDox Infinity or Kinesis Advantage

In the meantime some links for buyers in Poland:

Microsoft Sculpt Ergonomic
Razer Blackwidow - small, cheap and mechanical
CM QuickFire - small, mechanical, relatively cheap

So at the end I would appreciate if you have any suggestions or experiences with alternative layouts (I'm back to QWERTY for now) or can give me some advice if I should go for ergonomic or mechanical for now.

NVM or N

2016-05-18T18:05:00.000+02:00

From time to time I go back to the problem of switching Node.js versions. Sometimes it's enough to just go to the node.js website and install a new version but if you have to switch versions between different projects all the time then you soon realize you need a tool for that.

I know of two such tools:

N has really nice UX and it's really intuitive to work with it. The problem arises if your different projects use different versions of global packages - you have to deal with it yourself.

The biggest advantage of NVM is that it switches global packages with the node version so you really get a clean slate with every version. The biggest problem I have with NVM is that it's UX is really bad and I found myself explaining the same things over and over again to my colleagues.

Because of the global packages, I found myself advising NVM to people last couple of years...

At the same time, I have learned the hard way that global dependencies are causing much more problems than the necessity of NVM. All projects that have used global dependencies were really hard to setup and those dependencies create a lot of problems in the long run.

It took me some time to add those to facts together. Basically switching global packages should not be necessary!

Even for complicated projects it's not really that hard to add all dependencies to packages.json. I've done that multiple times end the cost of slightly longer installation is nothing in comparison to how much easier it is to introduce a new person or to go back to the project after a couple of months.

So what's my advice?

Add all your dependencies to package.json
Try using "n"
Switch to "nvm" if you encounter any issues

Feedback? Are you using something else?

Simplicity Talk from Meet.js Summit 2016

2016-05-17T17:42:00.002+02:00

Video of my Meet.js Summit talk is finally online, so if you missed please let me know what you think about it.

I have learned a lot from feedback at the conference and I have heard some great comments that I feel add a lot of value to my talk.

Important message from @krzychukula: Learn the language well and don't expect frameworks be solution for your laziness #meetjs16
— Mariusz Nowak (@medikoo) March 19, 2016

First, I really like the tweet by Mariusz Nowak. That's something that I haven't mention in the talk because I have assumed that people would do that before going for frameworks. Only when I've seen this tweet I've realized that it's not always the case and I should have emphasized that before moving to new tools and frameworks.

I have advised people to use Editors instead of IDEs but many people told me later that they don't want to miss on refactorings and all the ease of use they provide. To me that was the exact reason why I want to work without them. Later at the conference Michał Płachta he told me why he is using Atom editor himself instead of IDE. It's really easy to 'refactor' 30 places in your code without seeing them, and IDEs make it really easy but can you really know if your change fits nicely in those places? Is your refactoring improving one place you're looking at and making 29 others unreadable mess? If you use editor you have to go to every of those places and sometimes while doing that he realizes that his initial idea wasn't really good. I really like his example, thanks Michał!

All in all, it was great to be there and I want to thanks organizers for accepting my talk!

If you have any good feedback please let me know in the comments.

CraftConf 2016 Review

2016-05-05T10:21:00.000+02:00

Last week I have attended Craft Conference for the first time and I've decided to share my experience.

My favourite presentations:

Dan North - "Embracing uncertainty"
Jez Humble - "You Can Have It All: Software Development at Ludicrous Speed"

Some other worth mentioning:

Adrian Cockcroft - "The Evolution of Microservices"
Marty Cagan - "Great Product Team, Successful Product"
Doc Norton - "The Technical Debt Trap"
Yao Yue - "Toward Better Design and Code"

Most important takeaway:

Most important lesson for me was to be reminded about happiness and job satisfaction. The best way to make sure that something will be done is to let people have fun doing it. For example: if a developer can use this new language or framework she is enthusiastic about it's more likely she will do it just because it will be fun for her. No one likes to do chores, and getting money for that is not going to help much.

Other interesting themes/ideas:

Microservices - they are everywhere, and there are lots of tools that I haven't even heard about. A lot of innovation happens in this area but on the other hand, most companies don't need microservices and they actually might introduce more problems than benefits.

Pair Programming - it was inspiring to see about 90% of the attendees to raise their hands when asked if they program in pairs.

Relay/Falcor/om.next - I was sceptical but it might make sense to speed up development if everyone is a full-stack developer and maybe if different clients have really different needs - so that one REST API would be really inefficient. I still think most companies can do just fine without them.

Full Spectrum Developers - the main idea is that developer should know something about everything in the business. Some UX, finance, design, operations, management, estimation etc. At the end of the presentation with this idea, there was this really good question: how do you convince business people to be "Full Spectrum"? It was a brilliant question, but it showed up to late so it's something I think about after the conference.

Quality - a lot of talks were about the quality of what we do. Most speakers were convincing us that we should always create hight quality code (some even presented proofs). Yao Yue even told us to refactor while doing business features and just add this artificial time to the features. Some may disagree but in my opinion she's right. Most of the time business don't really know what it means, and what are the consequences of cutting corners. Most professions have government imposed acceptance criteria with inspectors and tests - if we want IT to stay how it is we should make sure to create quality code and be moral in the code we write. I recommend reading: Being A Developer After 40 by Adrian Kosmaczewski.

Thanks to Mateusz Kwaśniewski for reading drafts of this post and letting me know about Craft Conf this year.

Hero's Journey in Presentations

2016-04-20T15:59:00.000+02:00

Artur Skowroński recommended me: Want a Better Pitch? Watch This. by Andy Raskin and I think it is so great and applicable to all kinds of talks that I want to help it get more popular.

So the plan of presentation is going to follow these steps:

Enemy - Problem
Why now? - call to action
Show how the world will look like without the problem
Problems
Evidence that you fixed the problem

The image at the top is because it reminds me of Hero's Jurney. It has fewer steps but there is a lot of similarities. Especially #3 made me realize that grand vision of the future is something that I should have used in my talks.

The result at the start of the presentation makes a lot of sense actually. Everyone already knows that I speak about something because I probably fixed it and if you tell about it at the end then it might be hard for some people to see the bigger picture in the solution you're going to present.

How to Learn when Life is Short

2016-01-20T09:46:00.000+01:00

I haven't really think this through but I have read really interesting articles recently and I just wanted to write for myself later.

Life is Short by Paul Graham

It's just that, learn from people older that you what is most important, and that life actually is short...

Instead of Goals or Resolutions, Try Creating Rules by Leo Babuta

In the time of New Year Resolutions, it's good to read something that is proposing different way of achieving things. Still I see much more value in habits than setting goals. But...

The Easy Way To Learn Hard Stuff by Per Harald Borgen
It's something that I read from time to time and I know it works sometimes for me. The things I learn then are really learned and something Jakub Wasilewski has told me many times to do. Still it's uncomfortable and for some reason, I have failed to use this method more than once or twice. Writing it here so maybe I can think about it more and figure out how to do that more often.

Towards an understanding of technical debt by Kellan Elliott-McCrea
The most interesting idea that the idea of technical debt is a wrong analogy to the programming. The article hooked me at "Code choices that suck the will to live" as I'm really interested right now what it really means when you get that feeling. Is it really the code or maybe my confusion and lack of domain knowledge? It's somewhat connected to the first article "Life is Short".

That's it for now.
Please let me know if you have a view on any on them.

How Music Works by David Byrne

2015-08-07T20:17:00.000+02:00

I have expected a book about music theory, but it's something completely different.
You can learn about history and origin of music and how music industry evolved to its current state. It was intriguing to read about amateurs and how that music was played locally before recording technology emerged.

Book motivated me to finally learn guitar and explained a lot about the music industry.

Rating: 6/10

Book Review: So Good They Can't Ignore You by Cal Newport

2015-08-06T20:10:00.000+02:00

One of the best books I have ever read.

Do you feel that you don't really have the passion for your work?
If you have passion then maybe you feel that you don't have any particular mission?

Long story short this book will liberate you.
Your gain your passion with work and effort you put into something you do - preferably that pays your bills. If you like something or you have no other choice and you have to do something you still can make it your passion. In my opinion, that's liberating. What's the alternative? For me, I would need to be sad that I can't make reading fantasy books my job...

Mission:
After watching Brett Victor Inventing on Principle, I was thinking about my mission... What should it be? Cal Newport puts that into another perspective. First you have to gain expert experience and skills to have a mission. With enough knowledge in your field, it will be obvious next step to do. Keep build your skills and mission will show up eventually.

The most important idea from the book: Deliberate Practice.
Find something hard in your field and try to understand it. For me, it's learning more algorithms and languages out of my comfort zone. For now I've decided to read "How To Design Programs" but I want to learn a language with static typing after I go back from holiday.

Cons: some of the xamples are really extreme and book is really repetitive - I don't mind it as I treat it as a chance to better learn the ideas, but it's easy to find commenters annoyed by it so it's worth to mention.

Rating: 9/10

The Now Habit A Strategic Program for Overcoming Procrastination and Enjoying Guilt-Free Play by Neil A. Fiore

2015-08-05T19:48:00.000+02:00

Most interesting ideas:

Be mindful about your self-talk.
Do not think about yourself as lazy - procrastination is an escape when you're overwhelmed or are scary about finishing.

Unschedule
Put your free time and thing like commute/lunch/dinner etc into the calendar. What is left is your time for work but:

You have to start working for 30min and you can put something in the calendar after you have finished 30min work.
Do not work more than 20hours a week and 5hours a day.

Interesting that I read it to read how I can do more to discover that I already work too much. Guilt-free time for play is probably the nicest idea from the book.

Rating: 6/10