I started a startup with my own hands. Of course, I want to do everything to the maximum cool. I can not find in the network sensible information about comparing server languages ​​(platforms) such as php, ruby, pyton , etc.

Firstly? I would look at performance tests, but any other analysis and comparisons would be interesting.

Also interested in the article compared to the database.

Specify the direction, give references.

  • 2
    - C10k (Problem 10,000 connections) in different languages ​​/ platforms - Specter 6:52 pm
  • one
    @Stas Kazanin, you understand that you have formulated the question very widely. Any language, any DBMS has its own niche (in which to use them more correctly (taking into account the experience of the performers)). What task you have and what resources you have you can only guess. Want specific answers, ask specific questions with sufficient input. - avp
  • 2
    This is where serious performance issues are taken (as a whole). - avp
  • 2
    @AlexWindHope, just because node.js is lagging there? waiting for your example of true tests =) - Specter
  • 2
    @Spectre, did you figure out these tests yourself? I looked only at the habre, I did not even get into the git, because a picture is drawn from the Habrovsk text, what is being measured, what and how is unknown , and it is also not clear by whom. IMHO is nonsense. How many machines are running clients? How is the 1ms client startup interval provided? What are the characteristics of LAN? What kind of server? And a bunch of similar questions. Discussion is consistent. - avp

1 answer 1

"Cool as much as possible" is a properly composed TK (including SLA), as well as the corresponding TK architecture. It is useless to compare the database or languages ​​without knowing the problem. It can be either Erlang or Java, or Python, or Oberon, or C, depending on security requirements and verifiability of the code, hardware requirements, performance, maintainability, availability and cost of manpower, etc. There are no "silver bullets" - everything has not only advantages, but also disadvantages. You must first determine the requirements and criteria, and only then compare the options for them. Without this question about a spherical horse in a vacuum. Same performance - what is critical for you - reading, writing, sampling? How much data? Without understanding the situation, even the answer to this question is not straightforward.