I think it was roughly twelve years ago that I first got into Linux – via an early build of Slackware. I had Unix heads up by then – AIX and some Solaris, but I hadn’t really gotten into the whole “roll your own” thing. I had gone to a computer show to try and pick up a box I could use so I could teach myself how to upgrade a Novell 3.12 to a Novell 4.11 without melting everything down, and was deeply intuiged by this “Unix without a proprietary box” concept.
To say that first install was a nightmare is an understatement. Slack was pretty far ahead, but the internet was still pretty skimpy – you had to be on IRC or Usenet to really have the skinny on how to deal with kernel faults, or getting X-Windows to work. I think it took me a weekend, to get that first box running – it was an old P133 – and I remember feeling so accomplished.
Now, over a decade later, as I am trying to finalize in-place upgrade paths for Ubuntu snap images running on a VMware cluster – it almost frightens me how far we have come in so little time. Most of the fundamental understanding and sensibilities I have in regards to technology comes not only from having to fix other people’s mistakes and problems, but from making my own. As technology progressively gets slicker and slicker, it gets a lot harder to worm a huge mistake in by yourself, without a full fundamental understanding of the archetecture.
I’ve almost bricked my Pre twice since I bought it. Once, I was able to ressurect with some voodoo I found online. The other time, I was really figuing I was going to have to hit it with a rock, so I could get a new one under equipment replacement. In both cases, the only reason I was facing the problem was because I had already undertaken steps most normal people would have bailed on, well before the first page of documentation (there were seven). The reason both times? I forgot to lock my filesystem on the way out of hacking – SOP for embedded devices, something I used to think about all the time, but, due to my operating systems getting smart for me, not something I think of anymore.
This is less of a crotchety old man musing, and more of a wonder for the future issue – at some point, we are going to have to abandon the archetecture that we are on, if there is any chance of Sci-Fi coming out ahead of Luddite Armageddon. Who will be writing those systems, when everything is so slick it can second-guess you, and be right most of the time? Your average consumer doesn’t care about most of this – they just want the clicky clicky to take them to their email, or porn, or game, or social network. Where will the tinkertoys come in where everything is slick as glass?
Yeah, this is the crap that keps me up at night.