3 ms·

You would be surprised what can happen with "agnostic runtimes" Running the same libraries, on the same base OS is invaluable to me - but that can depend on wh

by mugsie 10y ago

You would be surprised what can happen with "agnostic runtimes" Running the same libraries, on the same base OS is invaluable to me - but that can depend on what you are developing. > Also, unless IT agrees to put beefy machines under each one's desk, the server is quite different from what developers can use anyway. With "the cloud", these days that is not necessarily true. a t2.large / m2.large are 2 cores, 8GB RAM. - well within the reach of a basic laptop. > Finally if there is any issue that really needs testing on the server, that is what remote connections (including remote debugging) are for. Never debug in production. No developer should have access to production.

pjmlp 10y ago
> With "the cloud", these days that is not necessarily true. a t2.large / m2.large are 2 cores, 8GB RAM. - well within the reach of a basic laptop. Which is why we just connect to "cloud" and clone our own instance from a production server, if we really need to. > Never debug in production. No developer should have access to production. Maybe in a perfect world. On real life I have been in projects where developers were exchanging dlls and jar files on them.
- mugsie 10y ago
  OK, the problems seem more culture based then. All of these issues tie together. Developers can't access production, so they start using production like environments, which allows for proper Ops work (Infrastructure as code - e.g. config management), which allows for easier setup of production like environments for developers ... and continue cycle. Its hard to move to that, and takes time, but it really helps. I know it can seem overkill, especially for small teams, but the tech debt paydown that happens when you end up doing it in the future can be huge, and massively disruptive.
  - pjmlp 10y ago
    I kind of agree, but to be honest I have been on and off of server side projects since the mid-90's, and can count up to five the amount of projects "done the right way".

by mugsie 10y ago

pjmlp 10y ago

> With "the cloud", these days that is not necessarily true. a t2.large / m2.large are 2 cores, 8GB RAM. - well within the reach of a basic laptop. Which is why we just connect to "cloud" and clone our own instance from a production server, if we really need to. > Never debug in production. No developer should have access to production. Maybe in a perfect world. On real life I have been in projects where developers were exchanging dlls and jar files on them.

mugsie 10y ago
OK, the problems seem more culture based then. All of these issues tie together. Developers can't access production, so they start using production like environments, which allows for proper Ops work (Infrastructure as code - e.g. config management), which allows for easier setup of production like environments for developers ... and continue cycle. Its hard to move to that, and takes time, but it really helps. I know it can seem overkill, especially for small teams, but the tech debt paydown that happens when you end up doing it in the future can be huge, and massively disruptive.
- pjmlp 10y ago
  I kind of agree, but to be honest I have been on and off of server side projects since the mid-90's, and can count up to five the amount of projects "done the right way".