Coding Chats

Podcast Host

Coding Challenges

Writing weekly coding challenges that help software engineers level up by building real-world software applications.

Chief Challenge Writer

Trivial Business

Providing interim software engineering management (Head of Engineering/VP Engineering/CTO), software development, consultancy and mentoring to clients on a contract basis.

Example projects:

■ Providing on-site training and mentoring to various organisations. Topics included NLP and Machine Learning. Delivered courses: Introduction to C++; Advanced C++; Python for Scientists; and Linux Essentials.
■ Interim CTO redefining the IT infrastructure and software systems of a business turnaround.
■ Interim CTO of an angel backed start-up.
■ Leading the development of a large scale website for the BBC.
■ Developing a bespoke CRM system for a publishing franchise.

Chief Everything Officer

Developing Skills

Writing a newsletter of weekly tips to help you level up as a software developer.

Chief Teaching Officer

Acast

Acast is a podcast advertising company. They help creators and advertisers connect.

AWS, CDK, Typescript/JavaScript and node.js

Software Engineering Manager

Met Office

The Met Office is the United Kingdom's national weather service, providing critical weather services and world-leading climate science, helping you make better decisions to stay safe and thrive.

AWS (Cloud) and Python for handling large meteorological data sets.

Contract Software Engineer

Sky is Europe's largest media company and pay-TV broadcaster.

Building serverless solutions to support Sky’s broadcasting activities using Python 3.8, Java, AWS and the Serverless Framework.

Cirrus Logic

Cirrus Logic is a fabless semiconductor supplier that specialises in analog, mixed-signal, and audio DSP integrated circuits.

Managing a team developing Data Engineering Tools to gather, process and distribute data to support the development, testing and deployment of machine learning algorithms using Python, Conda, PyTest, Numpy, Pandas, Dask and Dask-Distributed.

Manager of Data Engineering Tools

Starting to build a new team whilst designing, developing and deploying software tools to manage big data for the development and evaluation of biometric algorithms.

Principal Software Architect

Software development for High Performance Computing using Python, SciPy, Pandas, NumPy, Dask, Dask-Distributed and C.

Principal Software Engineer

Morningstar

Morningstar Real-Time Data provides real-time. low latency financial market data feeds.

Leading a globally distributed team of 40+ software engineering and QA staff developing and supporting real time low latency market data feed handlers and supporting infrastructure for several hundred financial exchanges. 
Using C++, C# and Python on Windows.

Head of Software Engineering

Helping you become a better software engineer through coding challenges that build real applications.

I am on a mission to level up one million software developers!

I’m doing that by: 

1. Helping you improve your coding skills through coding challenges that have you building real applications.

2. Helping you improve your software development and soft skills with posts and articles about software development.

You can find the coding challenges here ➡ https://codingchallenges.fyi

You can follow me here on LinkedIn for the posts and articles!

I also post some longer articles on my Developing Skills newsletter here ➡  https://developingskills.substack.com/


My BACKGROUND:

I'm a software engineer and sometimes a manager of software engineers. I've worked as both a senior individual contributor (Staff+) and a senior manager (VP Engineering, Head of Software Development).

I've worked across many tech stacks. I've spent most of the last six years building distributed systems with Python and Flask deployed on AWS using CloudFormation and some CDK. 

Recently there has been some TypeScript/JavaScript, React and Next.js on the frontend with Node.js and Express on the backend. Deployed on AWS using CDK.

For most of my career before that I worked in C++. I'm coming the full circle now and currently learning Rust.

I have a passion to help individuals, teams and departments reach their full potential.

There are too few and too many software engineers.

There’s thousands of unfilled software engineering roles right now. There are also thousands of software engineers that can’t get a role right now.

It’s because there’s a HUGE skills mismatch in the software development industry.

There has been a huge boom in people coming into software development and training as web developers.

Yet the skills companies list as difficult to hire for are: AI and Machine Learning, DevOps, Data Analysis, Cyber Security and Cloud architecture.

There’s a simple fix; hire for fundamental knowledge and potential. Then train people in specific technology.

With good fundamental knowledge specific technology is relatively easy to learn.

John Crickett

Stay away from people who try to convince you that a programming language doesn’t scale.

✅ Shopify have built massive scale on Ruby.
✅ Stripe have built scale on Ruby and JavaScript.
✅ Instagram supports 2 billion daily users on Python.
✅ Dropbox was built on Python.

Programming language doesn’t matter. It’s architecture that is the biggest lever when you’re looking to scale a systems, not languages or frameworks.

I've never worked anywhere that does TDD.

Yet I've often done TDD.

How is that?

It's very simple. TDD is a technique for writing code. When I write my code, I usually use it, it's an approach that works for me.

It doesn't matter whether or not I'm in a team or organisation that does (or doesn't) do TDD.

I don't need permission, I don't need the team to buy-in, I don't even need the team to know that I do TDD.

When I write my code with TDD I just do these three simple steps:

1. Write a test for the next bit of functionality I want to add.
2. Write the functional code until the test passes.
3. Refactor both new and old code to make it well structured.

When I push the change to the repo, no one else needs to know how it was written and most of the time they just don't care.

So don't use not working on a team that does TDD as an excuse not to do TDD. If you believe it will help you write better code, just do it. 

The most vocal and confident software engineer is probably not as clever as they think they are.

They’re probably not as good at writing software as you think they are.

Unfortunately, far too many people are overconfident in their own ability. Many people don’t understand a topic well enough to know little they know. When that is combined with the human tendency to equate hubris and arrogance to talent, there is a vicious cycle.

People think they’re better than they are and if they are vocal and confident enough, others mistake their overconfidence for talent, and tell them how talented they are, reinforcing their overconfidence.

So next time someone asserts confidently that there is only one programming language or framework that will solve your problem, only one way to use version control or only one way to make large changes to software, look for a second opinion from someone who isn’t so certain.

Confidence does not equal competence.

“The whole problem with the world is that fools and fanatics are always so certain of themselves, and wise people so full of doubts.” 
- Bertrand Russell.

I’m launching a podcast for software engineers! 🎙️

It’s called Coding Chats and it’s going to focus on software engineering for the 99%.

Because these days software engineers are presented with a rose-tinted view of the world - learn to code, learn to ace LeetCode and blitz the system design interview, land a high paying job at a Big Tech company or rapidly growing start-up, make great money and speed-run your career. 🚀

But that’s not the full story. It’s the highlight reel for social media. 🎥

And it only applies to the 1% of software engineers who work in big tech.

Coding Chats will be focusing on the 99%.

I’ll interview guests from novice software engineers to CTOs. From small companies to enterprises, from bootstrapped startups to VC backed startups. Asking them about their career, the challenges they have faced in their role and how they have overcome them.

We won’t tell you what to do, we will share with you what they did, and what they learned from the experience.

The first episode will be available on April 25th.

Until then, here’s the draft intro and ‘jingle’. Please let me know what you think?

P.S. who would you like me to interview and what would you like them to talk about?

Apparently I’m now the number one ‘IT & Tech’ influencer on LinkedIn in the UK.

The 19th across all categories on LinkedIn in the UK.

And 34th globally for IT and Tech.

All according to Favikon.

Thank you to everyone who has supported me, engaged with my content and in particular to all those who have supported and championed Coding Challenges.

I’m learning a lot from you all! Thank you! 🙏

Be careful with the word "But".

It seems inoffensive. But it’s not.

Consider this; if someone reviewing your code says: “This is a good approach, BUT I don't like the way you've written this class, you should consider these changes....”

What do you remember?

The last part.

Why? Three simple reasons:

1. Our brains focus on negativity,
2. Therefore we remember the last words more,
3. And everything after the "But" has more power.

Instead, change the “but” to an "and":

“This is a good approach, and I'd like to suggest some changes to the way you've written this class..."

For more on this, check out the Developing Skills post linked in the comments below.

Python isn’t suitable for enterprise software is a myth.

The reality is Python plays a huge part in enterprise software and big tech.

Here’s what I’m basing that off:

1. Personal experience - over the last decade I’ve built enterprise software in Python and trained developers and quants at banks and hedge funds to write Python, so they can build enterprise software using Python.

2. Python is a core language for J.P. Morgan’s Athena program and Bank of America’s Quartz program. There are around reported to be around 5,000 developers using Python at Bank of America and around 10 million lines of Python code in Quartz alone.

3. The pinnacle of enterprise software is the mainframe. The pinnacle of mainframe suppliers is IBM and they offer support for Python 3.11 on zOS.

4. Not only is it used in enterprise software, it’s used in big tech. For example, here’s a few big tech users you might have heard of: Netflix, YouTube, Google, Dropbox. Netflix uses Python for everything from operations management and analysis through to security and networking. Dropbox built must of it’s service in Python - millions of lines of it.
5. It’s used at Intel, Salesforce, Datadog, NASA, Pixar, Spotify, Pinterest, and Instagram. Nasa use it to process data from the James Webb Space Telescope as well as a number of other things - check out their list of open source. Spotify runs a number of it’s backend systems in Python.

That doesn’t mean you should use Python for your system. It does mean you should make an informed decision rather than believing myths.

We can develop quicker if we don’t write tests!

I’ve heard it so many times and it’s never been true.

In my experience, automated tests have proven their worth time and time again – saving me and the teams I’ve worked with – so much time. Time not spent debugging code, time not spent supporting customers who’ve suffered from bugs released into the wild. Time that can now be invested in building a better product.

I’m shocked when I come across a modern software development team that either doesn’t write automated tests or worse still turns off the few automated tests, they have so that they can ‘code quicker’.

It doesn’t work that way. It is always slower to develop good quality software without tests. Or as Michael Feathers eloquently puts it in the great book ‘Working Effectively with Legacy Code’:

“Code without tests is bad code. It doesn't matter how well written it is; it doesn't matter how pretty or object-oriented or well-encapsulated it is. With tests, we can change the behaviour of our code quickly and verifiably. Without them, we really don't know if our code is getting better or worse.”

TDD is the best way to build software.

And for those who confuse TDD with automated testing, I'm talking about writing the tests before writing the code, aka test-first.

Here's why I think TDD is best:

1. Many defects in software stem from a lack of clear requirements. To write a test, we need to understand the requirements. To write a test first, we need to understand the requirements before we start coding. That's a good thing.

2. Designing interfaces and abstractions is hard and getting it wrong makes our code hard to change. By writing the tests before the solution you need to think about the interface and abstractions from the point of view of the person who will be using them.

3. Testing should be focused on the behaviour, not the implementation. Tests that focus on the implementation may miss errors in the behaviour and can be brittle as they're too tightly coupled to the implementation. These tests a PITA as you then need to change them when you change the code, which means you can refactor quickly and easily.

4. With a test suite that tests behaviour, you can change the implementation safe in the knowledge that as long as the tests pass, the behaviour of the code remains the same. Refactoring or even re-writing becomes cheaper.

What do you think?

P.S. I know people who write great software with a test-last approach, I think that's perfectly doable, but harder and requires more discipline. 

My framework for learning a new programming language:

Step 1 - Order the best introductory books.
Step 2- Read the official tutorial whilst waiting for the books to arrive.
Step 3 - Read just enough of the tutorial/books to be able to build my first program.
Step 4 - Start building some real-world software (I use the Coding Challenges for this).
Step 5 - When stuck, go back to Step 3 and repeat.

Works every time.

I’m launching Coding Chats - yet another podcast for software engineers! 🎙️

These days software engineers of all levels are presented with a rose-tinted view of the world - learn to code, learn to ace LeetCode and blitz the system design interview, land a high paying job with a Big Tech company or rapidly growing start-up, make great money and speed-run your career. 🚀

But that’s not the full story. It’s the highlight reel. 🎥

And it only applies to the 1%.
I’ll be focusing on the 99% (those of us who are interested in more than just careers in big tech).

I’ll interview guests from novice software engineers to CTOs. From small companies to enterprises, from bootstrapped startups to VC backed startups. Asking them about their career, the challenges they have faced in their role and how they have overcome them.

We won’t tell you what to do, we will share with you what they did, and what they learned from the experience.

The first episode will be available on April 25th. 🗓️

Until then, who would you like me to interview and what would you like them to talk about?

If it takes 6 months for a developer to become productive in a new role, something is wrong.

And a lot of the time it’s the onboarding provided by the organisation, the hiring manager, and the team.

Here’s what a good onboarding should look like:

👉 A welcome email sent several days before their start date with a plan for their first day.
👉 Laptop and relevant kit sent to them before their first day if remote. Ready and waiting for them if their first day is in the office.
👉 All the accounts they need set up in advance, with passwords written down clearly for them (of course these should all need to be immediately reset by them).
👉 A first day welcome meeting as soon as they’re online/in the office. Ensure that this is at a time you can both make without stress.
👉 A peer buddy assigned to them for the first week (sometimes more) who can help them out. I often ask the most recent joiner to do this as they have most recently been through the experience themselves.
👉 A clear plan for their first weeks with clear expectations.
👉 An clear onboarding guide that gives them all the information they need to become a productive member of the team.
👉 Some work set aside for them to complete and commit on their first day and during their first week - creating some small wins for them and some good initial momentum. Make them feel like they’re adding value as soon as possible.
👉 A series of meetings planned over the first week with their manager, their team and people from the surrounding teams that they’ll need to know.

If you provide all that and the developer joining the team has some experience as a developer they’ll be a productive member of the team in much less than six months.

If you’re not providing that, why not?

You probably use Git. 95% of software developers do!

But do you know why it’s called Git?

Or what made it different from the version control systems before it?
Git was created in 2005 when Linus Torvalds wanted a distributed system but none of the available free systems met his needs.

In a great example of how performance should be considered for software, he gave a concrete example of the performance problem to be solved; a source-control management system needing 30 seconds to apply a patch and update all associated metadata. Then he explaining that this would not scale to the needs of Linux kernel development, where synchronising with fellow maintainers could require 250 such actions at a time.

Therefore he specified that patching should take no more than three seconds. He added a couple more requirements: CVS was an example of what NOT to do and; have very strong safeguards against corruption, either accidental or malicious.

As a result the software development community, ended up with a fast, distributed version control system which was free. It has since become the de facto industry standard.

Oh and the name, Linus Torvalds described git as "the stupid content tracker" and has said that the name (depending on your mood) can mean:

👉 Random three-letter combination that is pronounceable, and not actually used by any common UNIX command. The fact that it is a mispronunciation of "get" may or may not be relevant.
👉 Stupid. contemptible and despicable. simple. Take your pick from the dictionary of slang.
👉 "Global information tracker": you're in a good mood, and it actually works for you. Angels sing, and a light suddenly fills the room.
👉 "Goddamn idiotic truckload of sh*t": when it breaks.

Another bit of random, knowledge I’ve acquired in writing my Coding Challenges newsletter! There’s even a challenge that walks you through building your own git client.

Did you know that more embedded software is written in Python than Rust?

I was amazed to learn that.

I was even more to learn that Python is the third most common language.

The top four are:

C (~70%)
C++ (~23%)
 Python (~5%)
Assembly (<2%)

From my digging around it seems that Python is very popular on IoT devices.

Sometimes writing tests doesn't pay off, slows down development and accrues technical debt.


A CTO wrote that yesterday. They are wrong.

Their argument seems to be that:

1. Bad tests don't provide a good ROI. Which is true and a flawed argument. The solution is not to avoid testing, it's to train your team to write good tests.

2. Poorly thought out tests are too closely coupled to the implementation, making it slower to change the code. Which is also true and also a flawed argument. The solution is not to avoid testing, it's to training your team to write tests against the interface not the implementation.

3. Bad tests are technical debt. I disagree here too, bad tests are deficiencies in internal quality caused by a skill issue.

By rejecting good practice because of a skill issue you doom the product to ever decreasing quality and the people to never having the chance to grow and improve their skills.

To solve a skill issue, train people. 

Be a leader.

Be a CTO!

Why you should consider using old tech to build your new project.

It’s called the Lindy Effect.

And understanding the Lindy Effect will change the way you chose a programming language and technology stack forever.

In 1964 Albert Goldman wrote about Lindy’s Law, referring a a delicatessen New York where comedians gathered to discuss recent show business happenings. In the article Goldman described a belief amongst the observers that the amount of material a comedian has is constant, therefore the more often they perform, the shorter their career will be.

Similar ideas were put forward by Benoit Mandelbrot and Nassim Talleb. The latter suggested in his 2012 book ‘Antifragile: Things That Gain from Disorder’, referred to the idea as the Lindy Effect noting that:

“If a book has been in print for forty years, I can expect it to be in print for another forty years. But, and that is the main difference, if it survives another decade, then it will be expected to be in print another fifty years. This, simply, as a rule, tells you why things that have been around for a long time are not "aging" like persons, but "aging" in reverse. Every year that passes without extinction doubles the additional life expectancy. This is an indicator of some robustness. The robustness of an item is proportional to its life!”

We can see this applying to technology and software development, all the most popular programming languages have been around for decades and the longer they are around, the more they are used, the longer they’ll be around and the more support there will be for them.

So next time you’re choosing a programming language or technology stack for your project, think of the Lindy Effect and pick a language and technology stack that’s going to be around for at least as long as your project.

P.S. I guess COBOL and Fortran are going to around a lot longer than we thought!

The correctness of code can be verified through tests.

I saw this claim recently, and I disagree.

Software testing is usually done empirically. We feed various inputs into the software and verify the output is correct.

If we can do this exhaustively, we can prove correctness. But realistically for anything we build in the real-world it’s not possible to test exhaustively.

Instead we select a subset of inputs that will test all, if not most of the execution paths. Unfortunately no matter how well we do this, we’re not proving the code is correct.

The only thing we can prove this way is the presence of a bug. All we need is a single example test case that fails.

If we don’t see a failing test case we don’t know if the code is correct or if we’ve just not tested it with input that triggers an error.

In short, as Dijkstra wrote:

“Program testing can be used to show the presence of bugs, but never to show their absence!”

Is this the best software development interview question ever?

“What happens when you type a URL in the address bar of a browser and hit enter?”

I think it might be. It’s so open-ended, a complete answer could hit everything from GUI programming through the operating system to device drivers, Bluetooth, network stacks, network protocols and all the way back up to rendering the returned web page.

There is something in the stack that every developer, no matter what their speciality is can drill into.

If you’re into GUI development you could talk extensively about the keystrokes and rendering of the URL, then handling the enter keypress. If you have deeper knowledge you could talk about how the OS handles GUIs. You might even go further down and talk about device drivers. BIOS and interrupt handlers.

A Bluetooth expert could dig into what happens between your wireless keyboard an computer for every keypress.

A network specialist could talk about everything from Wifi, through BGP, DWDM, OTN, HTTP, TLS, DNS resolution.

There’s something for everyone.

What do you think? Is there a better question and if so, what is it?

Taplio

John Crickett's Linkedin Analytics

John Crickett

What is John talking about?

John Crickett's Best Posts (last 30 days)

Want to drive more opportunities from LinkedIn?

Taplio

.css-1jgreyo{background-image:linear-gradient(284deg, rgba(39,43,72,1) 0%, rgba(33,47,183,1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}John Crickett's Linkedin Analytics

John Crickett

What is John talking about?

John Crickett's Best Posts .css-qqfgvy{font-size:var(--chakra-fontSizes-md);}(last 30 days)

Want to drive more opportunities from LinkedIn?

John Crickett's Linkedin Analytics

John Crickett's Best Posts (last 30 days)