34 karma · joined January 20, 2014
It's not always possible depending on the type of organization. A lot of managers are offended that you would work on something they didn't ask you to work on, even though they would never think to ask you to work on the thing, and even though the thing you work on pays for itself over a short period of time.
Other times I just say fuck it. If you want to pay me to do the same investigation over and over, I'll mentally check out while I collect my paycheck and spend my extra time teaching myself skills for my next job.
One comment I have is I don't agree with some of the advice about integration testing. Integration testing shouldn't mock anything (except maybe truly external 3rd party stuff like analytics). If your app is interacting with your own internal services, for example to do user authentication, you should interact with a real instance of that service, not a mocked one. This way you don't run into the situation where the interaction between the angular code and the authentication service is broken, but your tests think everything is fine.
I used to unit test extensively, and I still do, but I now prefer integration tests as my primary form of testing. It's more important to me to know that major features that span across modules/service layers are working. Most of the time the low level testing isn't as necessary if I know that things are working at the higher level.
I think mocking in integration tests is a cop out. First, it's by definition not an integration test (but I hate those conversations or blog posts where people haggle over the definitions of "unit test", "acceptance test", "integration test", "smoke test", etc). Second, you should take the time to set up an actual environment that mirrors production as closely as possible, and run integration tests in that environment. This means running integration tests against services deployed to real servers that hit real databases.
The BTC website is pretty bad too, coming from someone who doesn't already know what all those different frameworks do. There is no sample code or "getting started" guide to illustrate what the project does or why I would want to use it. Especially curious about why I'd want this template over others? What are the competitors/alternatives?
BTW, isn't Firefox is open source? You can replace my exaples with their open source equivalents (LibreOffice, Gimp, etc) and the argument still stands.
A uni should teach algorithms and data structures, networking, threading, etc, not how to use firefox.
Likewise, a uni should teach operating systems, threads, hardware, compilers, memory management, etc, but not Linux.
Even programming languages are generally not taught as ends in an of themselves. Good schools expect you to pick them up as part of the class. For example, the course is about data structures, but you will code in Java, C, and LISP during the semester, and you are expected to pick them up. They may help you a bit in learning them, and give you background on them when you move to a part of the course that switches to a new lang.
If/When I decide to leave there is no counter offer that will get me to stay. For me it's a matter of principle. By the time it gets to the point of counter offers, they've already missed their chance to pay me what I'm worth.
I don't know. Maybe only two out of the ten I've worked with have not cared and have actually listened to the advice of the techies. In one case, the techies forced them to rewrite everything in a new language for no real reason, and blamed a lot of shortcommings of the old system on the language when it was really a problem of poor implementation. In the 2nd case everything worked out great, and that founder's business is really taking off, traffic still nearly doubling every year. His stuff is running off a single WordPress server.