chousuke

34 ms·

271.
by chousuke 6y ago
You could replace orthogonal and immutable perhaps, but idempotency has a rather precise technical definition that doesn't really have any synonyms. It's a useful word too, because idempotency is generally a very desirable propert
272.
by chousuke 6y ago
Happy people are more creative, and most engineering is definitely creative work. If most of your work is engaging or satisfying, it's easier to drudge through the bad bits. It's not like physical work where you can just keep doin
273.
by chousuke 6y ago
Personally, I am excited (or satisfied) by systems that work , up to and including the processes required to keep it "in motion". The technology used is secondary as long as its properties and risks are understood and accounted f
274.
by chousuke 6y ago
The short reason is that if the user can do it, so can any application running as the user. I don't know about you, but I'd prefer my applications not be able to inject and read inputs arbitrarily, though it may be that even stric
275.
by chousuke 6y ago
I have to admit the Cockpit architecture is not entirely clear to me, but at least it seems to allow using SSH as a remote transport, so you don't actually need to install the web stuff on all servers. GUIs are great for discoverabilit
276.
by chousuke 6y ago
Has anyone tried to use this stuff at a scale of greater than a dozen servers? I briefly tried it in my homelab on CentOS 8, and immediately hit enough weirdness that I'm just back to SSH with Ansible now. I suppose having a GUI can be
277.
by chousuke 6y ago
It's ridiculously overpriced for what essentially amounts to storing a tiny binary blob on a server somewhere and making sure it's backed up. I would've been happy to continue paying 12 USD / year for that service, but a
278.
by chousuke 6y ago
Is 200kLOC really a lot? Lots of software nowadays has hundreds of megabytes of dependencies, and people seem to be fine with that. Not that I think having tons of dependencies is really a good thing, but for an application that needs a fil
279.
by chousuke 6y ago
Often the failures are more illuminating than the actual solutions: you will see how different approaches can fail, and become better able to appreciate the solution. It's very easy to write out a finished blog post with recipe-like in
280.
by chousuke 6y ago
Right, and nothing prevents you from writing a compositor that talks to an external process to handle window management. It just happens that no-one does it, because it's insane and difficult to make work properly. The compositor'
281.
by chousuke 6y ago
It also doesn't help that by default (at least in my experience), Powershell's tab completion is slow and annoying to use. I don't know if it's technically inferior, but it feels terrible. It does the thing where you g
282.
by chousuke 6y ago
I live in Finland but am lucky enough to have access to private health care through my employer's insurance. (private health care often works faster and the queues are shorter.) The MDs will in fact help structure your treatment such t
283.
by chousuke 6y ago
it's important to understand that evolutionary processes are not random. People experimenting with proto-bread would tend to prefer the best results, even if some steps are discovered by accident.
284.
by chousuke 6y ago
You talk like using a physical server somehow prevents building a highly available system or from using automation. Why would the hardware be any more of a pet than any random VM in the cloud? You could also just run a hypervisor on the hos
285.
by chousuke 6y ago
In my experience, the "later" may come too late. I've seen code copy-pasted and slightly modified over a dozen times, sometimes without even eliminating dead code! Copying a function or even a whole file is fine if you act
286.
by chousuke 6y ago
The point is the amount of "touching". Applying security patches to RHEL is still a change, but it's significantly less risky than upgrading a faster-changing system where you might not even get security patches at all for th
287.
by chousuke 6y ago
It's not really about running servers for 10 years. It's about having a platform to build a product on that you can support for 10 years. RHEL software gets old over time, but it's still maintained and compatible with what
288.
by chousuke 6y ago
As far as I can tell, it's going to be similar to Debian Testing, ie. a rolling-release distribution that will have security support. RHEL point releases will then be snapshots of that rolling release with their own security fixes. I c
289.
by chousuke 6y ago
Not everything chases the latest and greatest though. If you run containers anyway, wouldn't you want the host running those containers to remain stable and predictable? Would you gain value from running your databases on rolling-relea
290.
by chousuke 6y ago
As far as I understand it, that distinction doesn't really make sense with Wayland. The display server / compositor is also the "window manager". Client applications have their own buffer they draw on, and they negotiate
291.
by chousuke 6y ago
I feel like sometimes software could benefit from a better conceptual differentiation between failures and errors. Failures happen, and robust software should deal with them in a predictable manner. Not all software needs to be able to reco
292.
by chousuke 6y ago
It seems to be dynamic. CPU pinning means the pinned process always runs on the specified cores. If you instead group processes into trust domains, workloads can still be scheduled on arbitrary cores, but the scheduler can use the groupi
293.
by chousuke 6y ago
This looks like a mechanism that allows more efficient use of CPU resources. You would still be able to schedule workloads on multiple cores as usual, but if they're all in the same "trusted" set, you can skip side-channel
294.
by chousuke 6y ago
I did this with my S5. It's not the easiest operation, but the nice thing about the device is that a factory reset is an actual factory reset, so you can always start over with factory firmware if you mess up. I use a Wireguard VPN t
295.
by chousuke 6y ago
I think using Ansible is not by itself a problem. There are many things I don't like about Ansible and I don't think it's particularly robust, but it is an improvement over manually written scripts, most of the time, becaus
296.
by chousuke 6y ago
For me, modal editing is great because it's comfortable . I almost never need chords save for those using control, which resides in its rightful place below the tab key. Lots of typing, zero RSI. Modality also enables the vi "min
297.
by chousuke 6y ago
I haven't played ADOM since 1.1.1 (almost 2 decades ago) but I always found it more approachable compared to the alternatives I tried. It starts you off in an overworld and the initial progression is less linear than the traditional du
298.
by chousuke 6y ago
There are other options, though. For upgrades, a virtual machine snapshot is a perfect backup, and with really large databases, you would likely have something using WAL archiving, which provides a continuous backup until the point that you
299.
by chousuke 6y ago
I find it unfortunate that augeas is so magical. Or rather, the design of its interface is very surprising What's up with requiring "--noautoload"? What does "-r ." do? and the transform invocation starts with the s
300.
by chousuke 6y ago
Sure you can do transactions across services. Do you really want to deal with the problem of distributed consensus though? If software especially built to deal with it often gets it wrong (as I think Jepsen tests have frequently demonstra

More ›