OK open-sores software developers… if you’re going to make business software products whose whole pitch is centered around the convenience of magically downloading shit from the internet, please do the following:
Include in the manual how to point your piece of shit software at the corporate proxy server. Perhaps you work at the only idiot company on earth that has no proxy, but please consider the rest of the world.
Put said instructions at the top of the manual or at least under an obvious heading like “configuring a proxy server” or “setting up your internet connection”
Make the instructions descriptive. Do not simply say “put the proxy settings in the place where the proxy settings are put.”
Include a test function to verify that the proxy settings are working. After all, connecting to the internet is the raison d’etre of your shitty software, right? Why not put a “test” button to make sure it’s configured, instead of some random message that looks like a local failure?
Why the fuck do developers create software with network functionality that is virtually nondebuggable? Why do they create a function that downloads a file, and when the app fails to connect to the internet to get the file, they present the error message “File not found?” FUCKING JERKS! You could very well catch a network exception and present a meaningful error message if you weren’t goddamned idiots.
This is a special shout out to the developers of Maven, Artifactory, Archiva, and 20 other lame-ass Apache projects who aren’t even worth mentioning. Go suck a rock, your open-source wizardry is lame.
I think it’s because open-source developers are generally rather lazy and only do the “fun” bits, or at least whatever interests them. Which, sadly, doesn’t include manuals or debugging.
Rather, I’d say all people are lazy and consequently so are programmers. The trouble is that open-source programmers don’t have the right incentives(ie a boss) to make them do the debugging and documentation.
My perspective on it is more that most software developers think only “in the box”, i.e. in the machine where they expect the software to run.
Where I used to work, we routinely had problems with communications software where both ends were owned by the same development group. Anytime we had a problem where “A” could not communicate with “B”, their first answer was “It must be a network problem”. Never mind the fact that this was an aggregated quad-OC13 (translation - a huge fucking pipe) at only 20% capacity with a total downtime of something like 40 seconds over the past 5 years. No, “it must be a network problem.”
And yet, even under the assumption that the network is totally jacked-up and unstable, not a single debugging tool to quickly and easily verify the problem?
There aren’t enough :rolleyes: in the whole world.
Agree about the ‘fun bits’ part. That’s why most Open Source software isn’t worth the trouble to try and use. If you want the non-fun stuff done right you have to pay someone to do it. Damn few people are going to spend their free time writing documentation or debugging just for ‘fun’.