29 comments

[ 2.9 ms ] story [ 81.1 ms ] thread
(comment deleted)
Title should specify they're only discussing mobile app development.
It should also, either specify what they learned, or go full clickbait clarifying that I won't believe number four.
75,000 CI/CD builds we love (number 31,284 will blow your MIND!)
Since this is an advertisement, why would they tell you what they learned? If they did, then you'd be able to, theoretically, compete with them.
Hi thanks for the feedback! The purpose of the article was to share our unique data and make some sense of CI/CD use cases. Do you have suggestions, what would be a interesting read for you and we will put it in our roudmap. Best
+1 I came here to comment the same. At least on HN.
Hi. Thanks for the feedback. I will specify that!
Request for Post: web dev version of this. maybe from heroku/netlify?
This seems little more than an elaborate bit of PR for "Nevercode"?
Nevercode? Lol. If anyone brought up something called "nevercode" in a meeting I'd laugh them out of the room.
Its the new hot Harvard Business trend where you make a company out of SaaS subscriptions and marketing!
"Elaborate" is putting it kindly. It's just soulless, blunt PR copywriting.

I'm surprised they got to the front-page with this tripe?

Thanks, this is why I spot check the comments before opening the article
This is a prime example of numbers with no value added. Just spitting out figures. No insight. For example relating frameworks to build times.
Hi, thanks for the feedback! Relating frameworks to builds times is an awesome idea! We will start working on it. Do you have any other suggestions to make more sense from our data?
Saying that build times is time you've saved developers is a very odd conclusion - I doubt most developers sit watching the compiler build an app...
Build time is not just compile time. My app is Ruby so deployment time is nearly nil, but we have a regression suite that takes inching past an whole hour, that we run before every merge, and consult in detail before prod releases.

I hope I am an outlier (admitting that regression testing should not take this long, as there are larger projects with much broader scope that get the job done in significantly less time than that), but without CI this strategy would have already failed a long time ago, and it has helped us immeasurably with onboarding new developers and keeping release quality high.

"Build" time is not compile time.

I hope you don't watch any of that either.
Without CI, what choice would I have? Go find another computer when I think my feature branch is ready to merge? (...stop working for an hour?)
Put your running build process in the background and continue to work on something else?
That's exactly what CI does. It saves me from doing it, and it can do so with concurrent workers (say, if I finish the next thing before the hour is up. Should I be running the two concurrent test suites on my local machine? Or holding them in a queue, manually...)

My build spins up a browser and does expensive feature testing in a browser window to prove that our Rails app and JS code play nice together. For a long time, this was working in a way where the browser would stay behind other windows and not get in my way. Then a Chrome upgrade came and now, at the start of every example, the window moves itself to on top of the other windows on the screen, and steals the focus.

I know you're going to tell me to file a bug somewhere, but I suspect this change was made on purpose for a good reason. (Maybe you remember pop-unders?)

CI solves this problem for me, rigorously, and in a way that is unobtrusive and not brittle. If you're not using CI, you probably don't understand the idea that "all developers should be able to run the test suite, but none of them should have to do it." That's the big idea. CI helps our team to increase their velocity, and keep it up.

Hi! Thanks for the meaningful contribution to the conversation. I share your thoughts on CI. I think it's a irreplacable tool for mobile app developers that actually reduces the hassle and repetitve grunt work. That's why this article was published in the first place! Take care!
@yebyen It's seems you know what you are talking about. Would you be interested to test Nevercode out and share you experience?
I'm not sure if their numbers are sound. For example they claim to be able to deduce hosting service market share from their build numbers, but considering GitLab has built-in CI/CD system and GitHub doesn't it would make sense to me that the numbers they see are not directly comparable.

Interesting numbers but no real value without more analysis IMO.

The term CI/CD suddenly appears everywhere. Can «continuous integration and continuous delivery» be shortened to CI&D or CIaD? Or how do you pronounce CI/CD?
I’ve facetiously pronounced it as ‘cicada’ in the past :)

I’ve only ever heard it pronounced ‘see eye see dee’ which is pretty quick to say while still being intelligible.