|
My experience is vaguely similar, but a decade earlier and longer and without much distro hopping. I touched SLS and Slackware first, but settled on Red Hat by the mid 1990s for consistency on my i386 and DEC Alpha hardware. Then I just followed through with Fedora and some CentOS. For the longest time, my workflow has been almost all XTerm and whatever X11 enabled emacs came with the distro. I've reluctantly used other terminal programs pushed by the distros. For work: autotools, make, and gcc before shifting mostly to Python. Plus BSD Mail or Mutt, until enterprise login forced me to Thunderbird. And Netscape and Firefox. I used to have to run Windows in a VM for office tools like Powerpoint and MS Word, but over time have been able to just use openoffice/libreoffice, partly because they got better at opening MS files, and partly because my career shifts and the changing world around me reduced the need for full MS compatibility. I've developed a strong "data orientation" and a feeling for the short half-life of most software. My important artifacts are data files that I carry forward over years/decades, moving from system to system and tool to tool. I have a strong distaste for proprietary file formats and other data silos where the content is tightly bound to particular software. Consequently, I also dislike or distrust software with a premise of having such silos. While I have quite a bit of skill and practice at building complex, distributed systems from my mostly academic CS career, I'm sort of an outsider to many popular end user practices. I dislike things like integrated IDEs, mobile phone apps, and cloud SaaS that all feel like the antithesis of my interests. Ironically, I have more understanding of how to build these things than I do for why anybody wants to embrace them. I don't actually want to eat the dog food, no matter how well I think we made it... |