Updated: August 1, 2018
Several weeks ago, a reader asked me to write a new "why I use" article, based on my original 2010 report, and share a story on my current (2018) setup of operating system and application choices. I thought this could be a cool idea, although essentially, it's pure bragging. But still, it might shed some light on how I perceive and utilize software, and some of you may find this useful.
On a personal level, I like reading my own articles sometimes, as it gives me a glimpse into a written evidence of my own past. It gives me a perspective on my own past motivations and reasoning, and that's always a valuable lesson. Of course, the test of time is always a valuable one. So let's. The software what I use, year 2018 edition.
I have to admit, not much has changed in priorities. I still value predictability and stability above all. Software needs to work for you, with you, not against you. Hence, I still think the tweaking should be kept to a minimum, backups are critical, and I want operating systems to change WITH hardware.
Hardware + software
I have mentioned this on many occasions. The first time in 2011, when I bought a new (gaming) desktop. It was a good opportunity to upgrade my 32-bit Windows XP SP3 arsenal to 64-bit Windows 7, and that was the primary reason for the upgrade. I also wrote about this in my Meltdown performance article. New hardware architectures make sense for new operating systems and kernels that can support them, and vice versa. Upgrading your operating system on the same box is not going to render any miracles.
Consequently, I always aim on buying high-quality hardware for my production setups, with as much memory as possible (in terms of what the motherboard permits). CPU considerations have become less critical, as the exponential growth in pure computational power has decreased, and the move from a single to two to four cores is no longer that important. Overall system requirements (and architecture) have pretty much remained constant in the past six years.
High-quality hardware means you will not end up with silly performance bottlenecks - which you may assume will be resolved by unnecessary software upgrades. This also allows you to use machines for longer periods of time and make natural transitions to the next generation of hardware + software. Hand in hard, long term support. In most cases, even five years, which is the typical Linux LTS, is not good enough. My systems are designed for long, sustained use, and so the choice of software becomes a limiting factor in a way.
I cannot emphasize how important backups are. I've written a whole article on this (above), and I have since expanded the setup with even more disks, additional backup copies, on-site and off-site retention, various other schemes, and more. Some of you may find the notion of having something like 7-15 copies of important data a bit too much, but then, if that gives you the flexibility you need, why not.
What has changed?
The one thing that has evolved in the past eight years - mobile connectivity. I have acquired a range of portable devices, mostly laptops, but also a variety of tablets and smartphones - do check the hardware section for the relevant reviews and such. But still, these systems are only an extension of the principles above. They do offer more flexibility, but they do not fundamentally change how I work in any way.
My primary choice remains Windows, with production systems running Windows 7 (desktops) and Windows 8 (laptops). The reasons are many - long-term support (10 years), predictable behavior and backward compatibility, and a full range of software that I need, including games, office suite, and more.
Much as I've written back in 2010, the range of programs that I love and need cannot be fully translated over into the Linux world, including 3D and interior design software, my text editor and image viewer of choice, and others. Gaming is still mostly a Windows thing.
As before, I find the security nonsense to be completely overrated. In general, my Windows machines are utterly reliable, easily running 80-90 days without reboots, and stable save for a rare occasional BSOD caused by faulty drivers (non-Microsoft). Those actually come down to the Nvidia issues that I had back in 2013, a crash with a virtualization tool in 2011, and an ATAPORT bug that I discovered while testing MBAM in 2017, although these last two can't be really classified as production problems. In fact, since 2005, with a total of 26 uptime years of production Windows desktops, I've only had four cases of BSOD, all of them hardware-related, two each due to graphics card overheating on my older desktops, and two each due to faulty Nvidia drivers.
Windows 10 is a paradox of a sort. It does allow you to use it like any old Windows, but it also comes polluted with the new-wave fast-rapid-activity crap that contributes nothing to sanity or efficiency. I am also not certain about how it is going to change or evolve, as I expect a full range of functionality from the start, and none of this incremental dopamine-dribble we'll figure it as we go nonsense.
It seems Microsoft is committed to the same 10 year (five plus five) cycle for Windows 10, although this could change - as it has a little bit with Windows 8 and the newest generation of CPU, remember - but in general if you (or I) go for new hardware in the next year or two, you can still enjoy the latest and the greatest, including the necessary drivers and whatnot, and have reasonable peace and quite until the next hardware refresh. But then, there might not be Windows 11 or alike, and Windows 10 will receive lifecycle updates. I am still not 100% sure, though.
Ah, Linux. Well, things have changed here quite some. The world of Linux is one big swirly rollercoaster. In the past eight years, we had Ubuntu come and become the staple food, offering five-year LTS releases, a store, and a semblance of real professionalism that does not exist in the Tux arena. Consequently, today, it is the only operating system of the Linux persuasion that I use in a production capacity, on my laptops.
To be more precise, we're talking Ubuntu 14.04 Trusty, which proved to be a phenomenal release on all levels, and the best Ubuntu ever. Since, no version has matched its quality and completeness, and I am not really sure what the next one will be. Recently, the Plasma desktop has shown some real promise, and Kubuntu 17.04 Zesty was truly great, alas its successor are just average, although, post-upgrade and several months worth of bug fixes, Kubuntu Beaver is quite all right. Still, this remains an open question.
But the question is, why Ubuntu, right? The answer is, I use Ubuntu as a portable system, and it's quite convenient in that regard. Laptop wise, you also good commercial VPN support, plus encryption, plus a decent range of software. Specifically, I use Linux a lot for video processing, and in general, if you need to manipulate files, in my experience, Linux is often easier and more convenient than Windows.
So in a way, Ubuntu complements the Windows needs, and can double in some areas, but it does not fully or completely replace it. At the moment, doing that would mean impairing my own capabilities, and that's not a smart thing to do.
Future potential candidates
CentOS remains my Linuxe fatale, and has been so for many years. I do want to use it in the production capacity, but its brilliant conservative stability is a double-edged sword, as it lacks some of the modern features that I expect or need. Sure, you can replenish those in unofficial ways, but remember - I do not tweak, and I never use hacks in my setup, and going with a private repo that can go away in seven months is something I will never do on my production machines. Unless it's supported and guaranteed by the official vendor, it's not going into the arsenal.
Kubuntu 18.04 might be a good candidate for a successor to Trusty. Or maybe 18.04 with Unity. Well, that's something that time will tell. And I will tell too, because I'm going to write an article about it. Ha ha.
A boring, predictable article, right. But I like it. I am happy that what I believed was the right way to do software (and hardware) wise eight years ago proved to be the right method for me. Tested and tried over time including major changes, hard disk failures, and the long-term need for stability and efficiency. I am not sure if this helps you in any way, but if you want sane, quiet, predictable computing, perhaps there's merit in my words.
Looking forward, say the next eight years, I don't see any big revolution. Whatever happens in the cloud or IoT space, or whatever new buzzword comes around, they won't radically change my mode of work. They might enhance it somehow, but the principles remain in place. I have no time or patience for rapid agile-like bullshit changes, whimsical decisions by software vendors, or alike. That's for people who want to waste their lives while others make money off their misery. So there you go, in the most general and practical terms, the software that makes me tick. The end.