DEV Community: The Stoic Developer

Line-Perfect Comparison of TodoMVC Apps

The Stoic Developer — Sat, 07 Sep 2024 23:28:42 +0000

Like todomvc.com, but focusing on making the code as similar as possible.

Also trying to keep DOM code with the HTML. I'm not sure how useful this is, but some frameworks offer some added sugar for HTML.

Consider component-party.dev for more detailed comparisons.

I started this for learning, and I'm not sure I'll work on it very often. I thought somebody might benefit and maybe improve on it.

EDIT: Oh, I "just" forgot to put the link!

https://github.com/icetbr/line-perfect-todomvc-comparison

How to Live a Meaningful Developer Life

The Stoic Developer — Thu, 09 Sep 2021 12:08:38 +0000

I have many years of experience as a developer and enough money to live by. I want to make something useful and fulfilling with my time and skills.

What I expect with this post

to learn from your feedback and experiences
to motivate developers into setting their hearts on something more than just retiring and ignoring all your years of experience
maybe to connect with some of you, so we can work together to realize some of these ideas
to push me out of my safe zone and giving these topics more attention

Some things about me

I'm 38 years old
I have 17 years of professional experience
I live a frugal life, only need money for basic needs
I like helping out developers less experienced than me
I like to code
I have a good job, but I'm feeling burned out of being just a cog in the machine
I don't have kids and have a lot of free time, whenever I'm not working overtime
I like to explain things using bullet points :-)

What I'm brainstorming about

Working for free

contributing to open source
it is OK to do charity / social work
but money is a measure of success

At the beginning of my career, I never thought about giving back to the community. Seventeen years ago open source was not as hot as it is now, but mostly I was just worried about learning what was required to get the job done. Later, I never seemed to have a bug I needed to be fixed so bad that I would spend my time fixing it.

I'm open to giving my time for free for charity projects, but I never found one where my skills are needed. I'm not a "create me a website" kind of guy.

Other than these cases, there is a saying "when you work for free you are worth as much as you are paid". The less somebody pays, the less they appreciate you.

Leveraging my talents

I know best practices, unit testing, agile, many tech stacks
do I want to create niche websites and live-off ads?

Nothing wrong with gaining money through ads, but it doesn't sound satisfying to me. I am curious about how much I could earn though.

I think most of what I learned in all these years are things that I like, but perhaps I'm just used to them. I enjoy the problem-solving and creativity aspects of software development. Looking through thousands of lines of log and fixing other people's bad code, no.

Teaching the next generation

the basics: not enough intellectual challenge
advanced stuff: only if in a work environment, not a classroom
online classes, ebooks: too much of these already

I've tried teaching people with very little skill, and it was a drag. Unless you're dealing with an exceptional learner or maybe a friend or family member, it is not worth my time.

As for more advanced stuff, this is something I've always done. When somebody doesn't know something, I "teach".

I feel there are enough quality places to learn online already, I don't see myself doing something better right now.

Following my passions

sharing about my experiences in the tech industry
becoming a fictional book writer, one of my lifetime dreams
writing about philosophy and living a good life

This might sound too generic, but programming is one of my passions. This is basically the blogger's path. Learn something, share something.

Naturally, my life isn't just about programming. I'm a lover of knowledge, therefore philosophy, and I find reality too boring, hence I like fiction.

Founding a business with like-minded people

I'm a tech guy, bureaucracy/marketing are not things I aspire to do
hiring experienced developers is hard, focus on teaching junior developers
look for people who love programming
no employees, only partners, everyone is a real owner, profit share
focusing on the local economy, no cutting costs by outsourcing

I don't want to spend time on things I might be able to do, but don't like. Partnering up is the way to go, as much as I like flying solo.

I've been a part of many hiring processes, the best programmers are usually already well-employed. Helping to shape the career of young developers is in line with my teaching goals.

I know this sounds a bit naive, but I want to find people who like code as much as I do. Variable/bonus compensation is nothing new, but I want to make clear I'm not just after employees, I want colleagues and to make new friends.

I'm not against outsourcing, but profit is not the end goal in this stage of my life.

Becoming an angel developer/co-founder

finding promissory projects with people struggling with the technical aspects
helping with the MVP, team formation, and then moving on
keeping myself available as a long time consultant
putting my foot on as many projects as I can
maybe I will fall in love with one idea and stay full time

This seems easier than the previous idea. I would get a taste of all the stages required to launch a successful, or not, business. This might be a lot of work though, but it seems like a valuable experience.

Don't commit to just one project, trying out the field until something resonates with me. And if it doesn't, not a problem, it is still aligned with my goals of helping out other people in the industry.

Finding a better job

some people are leaving Google/Facebook to work on their own
- so this might not be an ultimate goal in life
money in, money out: the more money I make, the more people I can help
don't just make more money at the expense of my soul

I still have more juice in me. Although I mentioned people leaving the companies many of us consider the top in the industry, I'm sure some plan to die working there, because they love it.

I have a very narrow experience switching companies, for one thing, it is a very laborious process. Perhaps I should give another one a go before giving up traditional employment.

Will I be leaving my job?

Probably not right now, but for the first time in years I'm back into considering it. I have many reasons to keep working where I am right now:

the team is great, and they count on me
the company is amazing overall, they have the best of intentions
I'm in a good position, being one of the seniors earns me some respect
I'm loyal to the bone, in 17 years, I worked for 2 companies, and I just left the first one due to cutting costs
yes, it is more comfortable to stay where I am in maintaining a big and complex software I know everything about. This also makes me feel more needed here.

But I do want to have a plan mapped out in case things continue to not go my way. My health is at stake here, working overtime for too long messes up my sleep, and then everything falls apart after that.

I will keep expanding on the ideas I mentioned here. I already contribute to open source, and I am trying to write more in order to share what I know.

It remains to be seen if I will delve into the entrepreneur route or just go for another job if I ever quit my current one.

Thanks for listening to me.

Developer UX comparison of javascript testing libraries

The Stoic Developer — Tue, 27 Jul 2021 19:33:16 +0000

Hi, over a year ago I made this project for personal use, with intention to share the results once it was good enough. Since this might never happen, here it is for those who want to help improve it, give me some tips or just might otherwise make something useful out of it.

https://github.com/icetbr/comparing-testing-libraries

What I'm looking for in a testing experience

fast single test run
- "flicker free" watch mode, that is I hit "CTRL + S" and VScode's terminal shows my test result without even blinking
pretty print string comparison diff
- I use almost exclusively deep equal comparison in all of my tests
async support
no special syntax
- it should be easy to change between testing libraries
- makes for a more consistent experience when reading other people code with different test frameworks
clean stack traces
- I only need one line of stacktrace to find my error, I don't want it to be the 5th of 10 lines
clear terminal before run
bail on error
- if the change I made broke hundreds of test, I don't need to see all of them
easy toggle serial/parallel tests
- unit run in parallel, integration in serial

My impressions

this is by no means a complete comparison, there is way to much to talk about each library
all of these libraries are awesome and i could use any of them for my needs. Having the choice, I'm being very picky In order to choose one
less maintained libraries usually have deprecated dependencies, it would be ideal to avoid them
- this include many of the tap reporters
popular libraries benefit from an active community, which translates to better assistance, more examples, plugins, etc
the library performance is not as simple as it look like
- for one off tests, some need to be run in watch mode because of the costly startup time (Jest)
- for the whole test suite, some need to run in parallel mode to really shine
- some are fast regardless how you use them (zora, tape)
I couldn't find any tap reporter that I like as it is
- the best one, tap-difflet needed to be merged with tap-dot for less verbose outputs
big projects like Ava and Jest have some interesting features, like running only the tests affected by the changed code
- currently I don't need these

Jest

initial configuration: hard if not using defaults
- needed to include testEnvironment
- huge performance cost otherwise (~80% on cold start tests)
- needed to include testRegex
- didn't recognize my command line pattern
very active development
too many lines of useless output when in watch mode
very user focused, readability in mind (ex: many useful assertions)
bail doesn't work for tests in the same file (bug)
problems identifying test files (ex: camel case userTest.js vs user.test.js)
polluted diff result, impossible to have inline diff
ridiculously slow cold start
Jest doesn't always try to run test suites in parallel
- weird errors when improper mocking
expect doesn't accept error messages
asymetricMatchers produce output structure different then equals

Mocha

very active development
my flicker free dream come true!
questionable choices: tests don't force exit when done
stack trace clean level: 1 (minor details)

Ava

very active development
no support for nested tests
parallel by default, but can use --serial cli
annoying messages in watch mode
the slowest of all watchers

Lab

somewhat active development
best (not by much) equal diff
makes sense to use if you're using hapi (I am)
stack trace clean level: 2 (some internal calls)
flicker speed: has some delay

Tape

no support for async (use tape-promise or other)
need a tap reporter
special syntax (t.test)

Zora

interesting idea, it is "just" javascript
fast no matter how you run it
paralel tests by default, it takes extra work to make then synchronous, bad for integration tests
weird integrations with nodemon that makes it sometimes hang
the tap reporters I like are old and unmaintained
- has its own reporter, but only if using its own runner
special syntax (t.test, await test, and others)

Benchmarks

I started using npm scripts, but they have a overhead when first invoking them
bash scripts give me more flexibility
NOT using a SSD if I remember correctly

Watcher

results based on observation using the fastest of the watchers
10 is instant feedback


mocha	10
zora	9
tape	8
jest	7
lab	7
ava	6

Cold start

the time to run 10 tests from the ground up
don't be fooled by the numbers, this is not a huge issue
- most of the time people run tests in a batch, not like this
checkout [zora's][https://github.com/lorenzofox3/zora] more interesting yet still not a huge issue benchmark
if you have an interesting benchmark, please send me the link
ava and jest have an aditional large start cost when first run
./perf.sh to run all


zoraSingle	0,37
zora	0,52
tape	0,70
zoraReport	1,12
tapeReport	1,26
mocha	1,72
lab	3,60
tap	5,40
ava	8,06
jest	9,52