I’ve reached a very nice resting point in the ongoing effort to develop a very useful, powerful, stable, and cool computer setup.
This started a while back when I built a computer. In particular, this computer. There are several advantages to building a computer. You can save money or get more bang for your buck even if you don’t pay less. On the saving money side, maybe you have components on hand that you don’t have to buy. I did, mainly mass storage. The case I had, thinking I’d save money there, ended up not working out. You get more bang for the buck because the parts you buy will be better than the ones in the equivilant off the line but cheaper computer, and you’ll have more control over what happens in future upgrades.
For example, say you build a compute from parts for $1,000 and buy an equivalent Dell computer for $800. You save $200 with the Dell.
Then your motherboard burns out. For the computer you built, you replace the motherboard. For the Dell, you may or may not be able to replace the motherboard. You won’t be able to buy the exact one that burned out. The other componants have been designed to work with the Dell OEM motherboard, maybe not with the one you replace it with. There is a good chance that you will even have trouble at the level of getting into and out of the highly optimized Dell case. You can fix it, but it won’t be easy or cheap.
Meanwhile, Dell has designed the motherboard to be as efficacious as possible, meeting the specs at minimal cost. Chances are, the motherboard you bought, that cost you $45 more than Dell paid to put its motherboard in their machine, is sturdier and better performing, unless you are comparing your build to the Dell professional level computer. And the Dell professional level computer comparable to your build does not cost $800. It cost $1750.
So, if you build your own computer, you will not save piles of money, but you might save some, and you probably built a better computer for that price than you’d get commercially, or equivalent. And, there is a very comforting level of comfortable comfort knowing that if something burns out, you just put the new one in, because you totally know how to do that and what part to get. I know in advance more about how to fix a major problem with my hardware than any professional tech anywhere.
The other big advantage is that I get the computer I want. Also, I got to chose the upgrade path. For example, I did not get the most powerful processor available (though mine is quite nice). But, the motherboard I got can easily handle a major upgrade in processor. So, it will be trivial for me to buy a much more powerful processor, if I happen to have a few hundred bucks with nothing else to do with in, and install it in less than 20 minutes. I know I can do this I know exactly what processors I can put in, and I know the exact procedure.
I also can build a second computer really fast. After building mine, I built one for Huxley. Different case, but otherwise almost exactly the same. His monitors are not as good as mine and he has way less RAM. But, it took me a couple of hours to build his. I can do a third one in an hour, easily. I don’t need a third one, but just in case, I’m ready!
After building this computer, I messed around a bit further with the hardware and ended up with my ideal mass storage system.
I had thought earlier that the ideal system was a very fast SSD with a second drive with data. However, since I use cloud synchronization, it actually becomes slightly risky to have the storage media on a separate drive than the OS, because a rapidly deployed desktop will sometimes get ahead of the hardware, and the cloud becomes confused. THis is easily fixed, but only in a reboot performance hit and with some tricks that then have to be maintained. So, I opted for a single giant SSD, using my smaller older SSD as a giant swap file (which is not needed because I have so much freakin’ RAM), and a third drive, a traditional hard drive, on which I store some files I don’t want taking up cloud space (like downloads of an entire already backed up web site or three, that sort of thing) and to hold an on site backup.
I’m running and decided I totally like and will stick with KDE Plasma. I’m not using it so much for he bells and whistles, but rather, because I can perfectly and nicely emulate a Gnome 2.0 environment on a desktop system that is currently very well maintained and not buggy. If you were thinking KDE takes up a lot of resources compared to other systems, be informed: Not these days. They made it very efficient.
Plus, resources? I have lots of resources.
Finally, I just deployed emacs with my somewhat messy but very well personalized .emacs file.
I open emacs with a launcher that opens “blank.txt” which lives in my home folder. From there I rename, or it simply holds whatever I’m working on. Someday I will improve that end of the process, but it is actually so simple, it is hard to imagine much going wrong.
6 thoughts on “A really good computer setup”
I told you that you’d like KDE!
Glad it’s working out for you.
Steve: Yup, it is working great. The real test my be putting it on my older laptop!
“If you were thinking KDE takes up a lot of resources compared to other systems, be informed: Not these days. They made it very efficient.”
Either very efficient, or also as likely that they just didn’t make it any worse, and Moore’s Law did the rest. Desktop environments, like browsers, keep pushing the edge for what is ‘fast’, as the libraries get deeper and deeper in abstraction layers and code bundles full of stuff you don’t need just to get 1 method you do. Loading (and ignoring) all that code uses up time and memory. KDE’s focus on stability rather than new flashy features has given them the Moore’s Law advantage: they’re not adding much code but the processors are almost twice as fast as 3 years ago.
The node.js / npm architecture is horrible for this sort of thing, where packages load the entirety of, say, lodash just to get one efficient array sorter – the weight of the package isn’t worth the speed of the one method you need.
To some extent, it may be KDE standing still while the other desktops become more drenched in molasses. But it is also true that recent benchmarking shows KDE Plasma of today using half the resources its own earlier incarnation used a year or so ago. KDE made a strong effort to decrease its footprint.
I use KDE but I’ve had trouble with it flaking out on me. When it does that I switch to xfce for a while. My motherboard was issued in April 2017 so it’s still kind of new. That may be some of the problem. When I built my computer in September 2017 the Linux kernel had only been working on my motherboard since the July release. It’s been slowly getting better. Anyway, I always build my computers. One of the main reasons is I don’t want to pay the Microsoft Tax. The other is to have control over the components. You are right of course about manufactured computers. I recently replaced a Dell computer that failed for some unknown reason with a similar HP for my Mother in Law. She simply couldn’t afford one I could build for her and she needed a Microsoft partition anyway. The specifications are just enough to run KDE and it does so without any trouble. I too use a SSD for my main drive and a regular hard drive for a backup. It’s only 250GB. My reasoning is that they’re good for about 5 years. In about 4 years when is near time to replace it, SSDs will be much, much cheaper, bigger and longer lasting. Time is on my side.
Exactly. The SSD I’m using as my way larger than needed swap was almost $200 when I bought it a few years ago (and for about a year it went mostly unused so it is low mileage), and the 1 terabyte one I just put in was under $200.