One of the things that gets me is analysis paralysis. When I first started programming (albeit not professionally) most people didn't have the internet, me included. So I had to just write code and figure things out. For a very long time it's how I did things. Now, with blogs, SO, github, etc. I fall into the trap of, "well, I need to do X, so let's see how others have done it." instead of just trusting myself and doing it. I don't know when it started getting worse, but I hate it and I hate myself for allowing it to get bad.
Fighting this is a constant battle. One thing that helps is a deadline. Another is a pomodoro timer. I really wish I could unplug my internet cable for a while each day, but people would freak if I were not easily accessible.
>I really wish I could unplug my internet cable for a while each day, but people would freak if I were not easily accessible.
I had the luxury of living without internet for about two months last year. It was very influential. I would save up a list of things I needed to use the internet for, and do it on public wifi on the weekends. The rest of the week I was offline. Naturally this lead to more productive and proactive browsing (like saving documentation for offline viewing (or simply learning to navigate offline docs), locally storing wiki articles/blog posts/potentially relevant youtube videos, and using RSS feeds for efficient browsing on limited time).
Now that I have internet again, I've re-implemented some of those methods, like limiting my use of web browsers, especially full blown graphical ones like Firefox. For example, instead of surfing youtube, I use a script that takes my search and downloads the first x videos returned (or just download whole playlist with youtube-dl). I try to use $browser for less than an hour per day, and have found that postponing launching Firefox until later in the day (but not too late) helps - so those little scripts help a lot by not being potential "cans-o-worms". At some point, I go through my list of internet-todos I've gathered during the day, realize most of them were just stupid time sinks anyway, and then search or address the ones that survived. I've also kept the "proactive browsing" habit, and storing a lot of stuff locally (even whole websites).
Of course, for more urgent tasks, like things that could a quick solution /now/, I don't hesitate to launch $browser, but I try not to deviate from completing said urgent task and closing the browser as soon as it's done - no tangential searching allowed.
I guess one way to sum up the workflow I'm trying to achieve is a "power users" approach to the NoSurf[0] philosophy.
Another thing which helps is doing a project in a language or framework which you're not yet as familiar with - hard to get in analysis paralysis when you don't even know how to start the analysis. The thing is if you're choosing a language or technology which is extremely appropriate to the task, for example learning a grammar builder IDE to tackle a parsing problem, you'll hopefully still produce a "good" (or especially, "good enough") solution despite your beginner status. And then maybe when you come back to the domain you had analysis paralysis in you might have a better perspective.
Fighting this is a constant battle. One thing that helps is a deadline. Another is a pomodoro timer. I really wish I could unplug my internet cable for a while each day, but people would freak if I were not easily accessible.