Flask! It's Pythonic and good. Once you've done a couple of Flask projects, you can spin up a new microservice in a couple of hours. It's much lighter weight than Django, so it forces you to keep fewer moving parts in your head while hacking.<p><a href="http://flask.pocoo.org/" rel="nofollow">http://flask.pocoo.org/</a><p>Lately, I've been using Flask on the back-end, and React on the front-end.
Rails.<p>You get lots of great stuff out of the box. Perfect for me to get up and running quickly for simple/side projects.<p>Writing Ruby feels nice and is always a welcome change of pace from a lot of the heavy JS stuff I am doing on frontend atm.
Django<p>Good solid base for quickly developing something quickly.<p>Iterating on a database schema has never been easier
since 1.7 the migration tools are built in and the ecosystem of plugins has lots of tools that make development of a prototype extremely easy. It just keeps on getting better, just the other day I found an plugin to automatically generate knockout.js powered AJAX forms without me having to write any of the JavaScript.
I guess I'm old school. For simple projects I still use Perl with CGI.pm and a flat file database that stores data in CGI.pm's "name=value" format in a text file.<p>If I need templates I'll add "HTML:Template", for email I'll add "Mail::Sendmail", et cetera, to my perl cgi script.
Google Firebase.<p>As a front-end developer I just can't warrant the time setting up a backend, setting up a database, defining routes, handling authentication and the other nonsense that seemingly takes a solid day to get working.<p>Firebase to me is the best choice for simple projects, I don't use anything else anymore.
Pyramid and Postgres on the backend, Mako templates w/bootstrap on the front.<p>When I'm prototyping, my data model is stored procs that return json structures to the templates. I use SQLAlchemy core to connect to the db and execute the procs.<p>A prototype like this is super fast to get running, it gives you a really clean app structure that you can hand off to pretty much anyone and have them work on features, if you're lucky enough to have a team. If not, it's pretty easy to keep it all in your head at once.<p>As things grow, you can replace components as you need to.<p>You start off rendering JSON anyway, so if you need a more advanced interface up front, you can use React or Angular or Ember to hook into the data model. Whatever you want, really.<p>If things get more complex on the backend first and you need more structure there, you can refactor the SQL (because it was probably really gross the first time you wrote those "dump everything" procs). If you end up with a really complicated data model, you may have to eat the performance hit and switch to Alchemy proper for maintainability. Either way, the choice is yours.<p>Sometimes the simple app is the one you need, and the biggest challenge is scaling quickly. That's easy too because the database handles transactions inside the procs. Run one instance of the app per core on your server and put them all behind nginx and repeat on as many machines as you need to cover your traffic needs behind HAProxy. Now you're maximizing multi-core performance per machine without having to write async/await code.<p>It's probably not for everyone, but I find it gives me an enormous amount of flexibility, scales well in every sense of the word, allows the app to develop organically, and has very few drawbacks.
Clojure and Elastic Beanstalk.<p>You take an arbitrary Ring-compatbile HTTP server, write a handler, and glue together whatever you need to.
And you most likely don't need a routing library, you can just use an array-map for that. When it is time to deploy, you create a jar that contains your whole project and upload it to AWS.<p>Rollback is easy, too, as you don't overwrite your previous version. Funny enough, it means you can treat versions as immutable values of your application over time. I use yyyymmdd for versioning in most projects, that works great with this approach.<p>There are many discussions on virtues of Clojure's dynamic development environment, so I'll skip that.
The language you know the most would be best fit. Maybe Python/Rails/Go could be most optimal fit, if you know very good Java or PHP and don't want to waste time on learning something new - stay with that.
Mojolicious. ( <a href="http://mojolicious.org/" rel="nofollow">http://mojolicious.org/</a> )<p>I am a sysadmin.
I like Perl and its plethora of modules.
With Mojolicious, I can write a simple web app/prototype within the hour.
Just PHP and MySQL.
I coded a simple back end for a web app 10 years ago with very simple templates. Runs fast as hell and I don't need any API documentation or dependencencies or whatever.
ASP.NET deploying to Azure, if you're into the ecosystem it can't get much simpler than that (a web page up an running in minutes).<p>It's now cross-platform and there's an Azure free tier.<p>Of course, other people can get similar productivity within their own most familiar stacks :) (someone mentioned an hour for a similar setup with Rails)
AWS Lambda functions and some kind of database. Lately that has been graphene. AWS SWF, SQS, SNS added in the mix allow me to POC distributed apps quickly. Finally, I've worked with API Gateway recently when needing to support mobile devices.