This post first appeared on my blog at www.mikenikles.com. Photo by myself, Sulsseeli in Switzerland
CPU, Memory & Storage on a laptop
I'm coming up to two decades since I started my career in software engineering and it's an understatement to say things have changed... 👴
The choice of integrated development environments, or IDEs for short, as a junior software engineer were somewhat limited. I personally was too new to have an opinion and often used what co-workers used to develop Java applications - Borland's JBuilder.
You better had a beefy machine to run this thing smoothly 💪. Over the years, IDEs continued to be fairly CPU & memory intensive. To support a more or less productive work style, it could result in purchasing a new laptop every two or three years.In parallel, developing software required an increased storage capacity - initially because of *.jar file dependencies in the Java world and more recently due to node_modules directories that can quickly take up large amounts of disk space as illustrated in the following, academically absolutely correct (not) image:
This too led to a need to purchase better hardware. That all was fine "back then", but it's 2020 as of the time of this writing and I don't see a need to pay a lot of money to purchase a top-of-the-market laptop.
☁️, ☁️, ☁️ - Everything is in the ☁️
As the software world embraced cloud environments to deploy and run applications, one aspect seemingly was forgotten... IDEs.
I'm a Chromebook user - have been for a year and a half and will continue to be (AMA if you're curious why)! While I can install VS Code - my editor of choice - on a Chromebook thanks to its linux support, I couldn't find an answer to "Why?". Why would I develop locally when in the end, the applications I work on run in the cloud? Why would I want to carry my laptop everywhere I want to write code?
There have been solutions to run an IDE in the cloud for some time. I remember Cloud9, then experimented with code-server before I settled on www.gitpod.io.
Are You Ready-To-Code? Start Instantly. Anywhere.
Gitpod launches ready-to-code dev environments for your GitHub or GitLab project with a single click.
Gitpod is based on the online IDE Theia, which very recently launched version 1.0 🎉!
Why Gitpod?
A core principle of mine is productivity. Anything I can automate, I automate. Anything I can get rid of, I get rid of. Anything I can outsource, I outsource. If I can drop my local development environment, I do so. As soon as you work in a team, all your effort of setting up a local development environment instantly multiplies by the number of team members. I'd much rather not have to deal with that for every new team member that joins and have them started with a fully functioning development environment within minutes. Who's worked in a team where the most recent developer who joins updates the "Development environment setup" wiki page only to find it outdated by the time the next team member joins? Be honest with yourself when you answer that 😉.
Let's illustrate life without and with Gitpod:
Now the Gitpod marketing team's magic... That entire circle is irrelevant!
Please read on if you're a skeptic by nature.
A URL to start the development environment
Here's how to get started with Gitpod and any Github or Gitlab repository - up and running with 0 clicks (but 1 press of the Enter key 😉).
- Copy the Github or Gitlab URL of your preferred project.
- Open a new tab in your browser.
- Type "gitpod.io/#" and paste the URL from step 1. Hit Enter.
Don't have your laptop? Want to quickly review code with a peer? The process is always the same.
Consistency
If your project has special requirements, Gitpod provides custom docker images. Your existing team and every new team member or open source collaborator runs the same environment. No more "it worked when I joined two weeks ago".
Terminal
A terminal is provided, based on the Docker image (see above).
Built-in code reviews
This is a nifty feature I quite like. Just as you can prefix a Github or Gitlab project URL with gitpod.io/# to start the workspace, the same applies to a pull request URL. Look at diffs and leave comments right from within Gitpod. More on that in the documentation.
What about my VS Code extensions?
That's supported in beta. Once configured, extensions are installed automatically when a workspace is opened.
Oh wait... it costs money... monthly?!
The overall feedback from my post reviewers was about Gitpod's monthly subscription fee, so it's worth I address that before I publish.
To start, there's a free (as in, the same cost as your local development IDE) version for open source projects that gives you 50 hours / month.
If you want access to private repos, you start with a 30 day free trial. After that, it's $9 / month for 100 hours of usage. If you need unlimited hours, it's $25 per month - that's less than $1 per day.
To put this into perspective: A 13-inch MacBook Pro costs $1,299. A Pixelbook Go costs $649 (or buy an even more inexpensive Chromebook). You save $650 and can invest it into an online IDE.
- Divided by $25 = 26 months
- Divided by 12 = 2 years and 2 months
Look out for...
It's important to be fully transparent and list two gotchas that come to mind.
- Online only: Yep, you need internet access to work on your code.
- Startup time: Yep, there's a startup time of a few seconds when you open a workspace. I've used Gitpod for a few weeks now and that never bothered me in the slightest. You may feel differently, but for me it's well worth all the benefits I get out of Gitpod.
Summary
I most recently developed https://github.com/mikenikles/markua-docs-addon entirely in Gitpod. The efficiency and convenience of this experience led me to write this post and to share my experience with a wider audience.
There is a lot more to how I work, reduce friction and increase productivity especially when it comes to web development. If you're interested, please head over to https://leanpub.com/cloud-native-web-development and sign up for updates of my upcoming book, Cloud Native Web Development.
👋
Latest comments (4)
Excelent post Mike! I'm seeking for a cloud-based solution too, and GitPod could be one!
I have a doubt about testing, how does it work with Gitpod for both frontend and backend applications?
Thanks for the feedback Guilherme. Frontend applications can be tested like you would on your localhost. Once you start the server, Gitpod shows you how to access the website. You can even make it public to share with others if you want their feedback while you're developing a feature.
Backend applications are the same. In my current project, it's 3 NodeJS apps in a monorepo that I start with a single
npm run devcommand at the project root. They run on different ports and can communicate with each other just like on localhost.Lastly, I also use Cypress.io for end-to-end tests. That runs either headless or through a graphical interface.
Do you have specific worries about testing?
Sounds great, Mike. My concern is more related to it, how to run the environment as I am in the localhost.
I have a NodeJS API with MySQL and a React website.
And what about the database, is it running on a Docker image for your backend?
Unfortunately, there's no Docker support at the moment the way you'd use it loally. It's a big feature I'm waiting for too.
However, your dev environment runs as a Docker container and you can customize the image that's used. So in my case, I install the DB as part of the environment and can access it on localhost. Docs: gitpod.io/docs/config-docker/