by LiveTheDream on 10/29/14, 6:18 PM with 1 comments
by dalke on 10/29/14, 7:17 PM
Salvatore wrote "likely", and not "always", hence it's not a statement that "then we can blame EC2 and go on with our day", only an idea of where to prioritize the search.
Why try to read any more into it than that?
Often, yes, "you’ve got to deal with the environment the code lives in", though EC2 offers lots of different environments so that doesn't appear to be the case there. But, for example, I had a horrible startup problem with Python on the Luster file system, which didn't handle all the file stats that Python does to initialize its environment and import libraries.
The solution wasn't to fix Python, it was to switch to zipimport or put the file on local disk instead of the cluster file system. The timing numbers were to figure out where to spend my time optimizing.
Similarly, things aren't always Redis problems. No software can handle all possible environments - no one expects Redis to work on a 4.77 MHz PC compiled with Turbo C 2.0, no one expects Redis to work on a machine using mag tape instead of a hard disk. Those are clearly cases where it's okay for the developer to throw up their hands and exclaim 'not my problem!'
Salvatore isn't saying that. But DEATH Hodges appears to be saying that we must all support the 8088 processor running on cassette tapes in order to call ourselves developers. Which is clearly silly, and thus reveals the lack of meaning to the logic.