3 ms·
I commented a few months ago that if I was starting a new company today I’d use Rails without hesitation. https://news.ycombinator.com/item?id=17355776 https://
by whalesalad 8y ago
I commented a few months ago that if I was starting a new company today I’d use Rails without hesitation. https://news.ycombinator.com/item?id=17355776 https://news.ycombinator.com/item?id=17355776
Well, I’m starting a company and our current prototype is Rails. I don’t regret my decision at all, although after spending the last four years writing exclusively Python it’s been a little jarring. The muscle memory is still there though and after two weeks I’m almost back to 100%.
I really missed Active Record and all of the conveniences there around migrations. When you’re prototyping and constantly evolving your data model - but still intend to throw real traffic at it - flexibility here with the hardened foundation of PostgreSQL is vital.
The framework and language aren’t red hot anymore but they are far from dead or even stagnant. There’s a tremendous amount of active development happening in Rails.
Meanwhile I’ve been learning Elixir via a course called “Elixir for Programmers” and love it a lot. I didn’t want to risk building things on Phoenix at this stage. It’s full featured and likely more than capable but the smaller ecosystem and it being a first production app for me wasn’t worth the risk. I’m dying to put the actor model and OTP to work in an environment where it would really shine.
Anyway... still loving Rails and Ruby! A lot! With SASS, Haml, rake tasks, lots of data fixture tools etc... and no BS with React and webpack, slow pipelines and maintaining two codebases for an API and a Frontend... you’re left with lots of functionality being built daily. It’s great.
Don’t get me wrong, React is awesome and it’s certainly the new foundation of UI development. But... the ecosystem is still learning to walk imho. Redux isn’t the answer. GraphQL is great on paper but the way people use it is not ideal (coupling data to ui). Long way to go before I’d go 100% JS.
This reminds me to book a ticket to rubyconf.
- paulie_a 8y agoTo be fair the technology stack is almost irrelevant when you are starting a company. write it in perl, python, rails, hell write it in old school cgi and use c++. It all comes down to how many sales did you make today. The clients will rarely care what language was used behind the scenes.
- whalesalad 8y agoThat’s a pretty myopic way to look at it. Sure, use what you know and forget the branding of your technology, I agree. But you need to be cognizant of short term vs long term gains, gasoline you can pour on the fire (libraries and integrations), the team you are going to build, the deployment story, the ops story, etc... I’ve seen companies that made poor technical decisions up front spend years paying for those mistakes. One could argue that those mistakes are what even put them in the position to where they could pay em down... but there is definitely a way to measure twice/cut once and save yourself from a world of hurt while still moving quickly and making sales.
- paulie_a 8y agoMy point was that the technology in a startup is nearly irrelevant if you don't have sales. Step 1-1000 of building a startup is selling.
- PopeDotNinja 8y agoI refuse to write Perl. If enough people feel the same way, Perl is not a good choice.
- wolco 8y agoLong term perhaps but short term it means increased salaries.
- PopeDotNinja 8y agoYou're not wrong.
- vgy7ujm 8y agoOr you are not the right choice to hire...
- ddebernardy 8y agoIt's not an irrelevant choice. Your talent pool will look very different with each of the options you mention.