36 ms·
Ask HN: I need help to choose the right technology and start building my product
I am starting a new business and I am facing the following recurrent question : how to choose the right technology to build my product?
I don't want to dive into the technologies I already master just to save me some time, I would be too afraid to end up writing all over again from the beginning because the technology I chose is not appropriate.
Here are the following stuff I'll have to handle:
- big amount of data (some tables might have ~1M entries)
- users (there are different kinds of users)
- reporting (for instance, I'll have to know in real time what is done by the users and also run reports on actions already done)
- build an API to allow mobile apps to consume features of the product
- all the fancy stuff to make nice mobile apps (push notifications etc.)
I am not waiting for a straight answer saying that I should use "that" or "this" technology but more for hints to help me make a decision! Thanks a lot!
- spicyxtreme 15y agoThe languages that are thrown about alot nowadays are Ruby, Node, Python and probably a couple more. I guess you could start with a comparison of all these languages and see which fits your app best. Something like this http://fabianosoriani.wordpress.com/2011/09/11/when-to-ruby-on-rails-when-to-node-js/ http://fabianosoriani.wordpress.com/2011/09/11/when-to-ruby-... any of these programming languages can achieve the criteria you put forth
- steventruong 15y agoHonest question but why wouldn't you use languages you already know, especially if you anticipate to do all the things you listed from scratch.
- thana123 15y agoJust because I don't have a predilection language so I would be ok to start with something new and more suitable to what I wanna do
- thaumaturgy 15y agoJust go with what you know and are comfortable with, and make sure you make it possible to change various parts of your architecture later on. It's easy to over-think and over-engineer when starting out, and it's usually not that hard to switch to a different DBMS later on or even use a modular architecture that allows you to build different parts of your application in different languages as you go along. The real danger for you, right now, is that you won't make it to 1.0, not that you'll have to write it all over again from scratch.