I mean switching to Linux but using a homebrew port instead of the native package manager? Really?
I really don't get why that seem's so off.
You will be running back to osx in no time after a horrifying Linux experience. So take my advice: save yourself the headache and stay with Apple.
The parent's goal is not to embrace change, is to get back to having the same things they liked and worked for them on the new platform.
If they could be satisfied with any random thing, they would have just as well stayed and followed the old platforms new direction.
I think that's understood. It's just that trying to make Linux behave like OS X might end up being more painful than just embracing change. That doesn't mean they shouldn't try, I guess, but it's something to keep in mind.
If you switch from MacOS to, say, Linux, you'll inevitably lose some things you value, but there are also inevitably things in the new OS you would miss if you switched in the other direction. You owe it to yourself to go find a few.
Not every non mac laptop comes with shitty battery and components ;)
Oh yeah, and you don't have to be afraid an OS update will make your laptop unbootable. Has happened to me under Ubuntu twice.
- Apple Kerberos does not support MS-KKDCP and multiple factor auth. Services that require Authentication Indicator are simply not available from macOS. Their Ticket Viewer is even more simplistic than in 10.6 timeframe.
- Their IPSec implementation is weird and haven't managed to make it work with Strongswan yet.
> Oh yeah, and you don't have to be afraid an OS update will make your laptop unbootable. Has happened to me under Ubuntu twice.
Just wait, sometimes it happens there too. There's a reason for that recovery partition.
Perspectives :)
That's not true. Apple's yearly OS upgrades can and do brick machines. At least with Ubuntu, I'd stand a good chance of diagnosing and fixing the issue.
Also, we are just talking about Ubuntu here and not all Linux distributions. For example, NixOS is a Linux with an entirely different method of managing upgrades.
Built into the OS :
Automator (aka Unix pipes for GUI apps)
Preview (doubles as a comprehensive image + PDF editor. Scans your signature from a piece of paper held in front of the webcam)
OS-wide data detectors : hover over a date in any text field, set up a calendar event instantly.
Time Machine : dead simple backups
Notes (now with native Evernote import)
Photos (with easy to use bread & butter editing + ML-based search/cataloguing, not to mention integrated backup+iOS sync)
iMovie
Garageband (+ comprehensive audio/MIDI routing)
Comprehensive cross-device integration with Messages / Facetime / Photos / Continuity
Before anyone replies with a list of supposedly equivalent desktop Linux apps : Apple has never been about the What (ticking boxes in a checklist), more about the How (the overall experience)
I'll admit that quality has suffered under the new 1 year OS release cycle, but in my experience it's still a notch or two above Windows or Linux's best desktop setups.
For years I used WordNet as my dictionary because it was the only electronic dictionary I knew of that was less than 90 years old and didn't put the definition on a web page cluttered with distractions (and because paper dictionaries are too slow and tedious compared to electronic ones).
But over the years it slowly dawned on me that it is not a good dictionary. Then I realized my Mac came with a good electronic dictionary.
(Another former Linux user.)
For example, despite the shallow learning curve, in our house, time machine was both unreliable and patronising (the time travel UI). For me, I'm using ZFS snapshots for my backup, technically it's more than a notch better than time machine.
Another example would be Notes, Apple have demonstrated over many years that they cannot write reliable bidirectional sync. Unison just works.
That's the key issue. If you're starting out, do not like to customize, or have not yet reached your (locally optimal) setup, then I guess Mac is superior.
However, having perfected my setup over the last 12 years, I find the default Mac tools quite limited; e.g.
Time Machine is inferior (except in UI) to Borg, to which I've switched recently from Bup, to which I've switched before from a set of rsync scripts (which are time-machine equivalent, and I had them running long before time machine existed).
I don't make music (so don't care about GarageBand), and I also care about privacy so all the cross-device integration that apple offers goes out the door.
Thunderbird is, for my use case, superior to Apple Mail; Firefox is superior to Safari.
Finder puts junk all over my file system, and the old hacks to stop it (aseptic, and a couple of others) are no longer supported under Sierra.
I like Unity (the v on 14.4 - I think v7) with a few tweaks much better than the Mac doc, though that's a matter of taste.
And I like the transparency of the system. My main laptop ubuntu never breaks, but when I do weird experiments on other hardware, I can actually tell what's going on - if my Mac had trouble booting, it's likely the recovery partition will come useful - but there's basically nothing else I can do.
Example: in college I had to work with some C code written by a professor who was apparently a Pascal diehard. He included a header in everything he wrote which #define'd Pascal keywords to their C equivalents:
#define BEGIN {
#define END }
...etc. I like Pascal just fine (Turbo Pascal 3.0 forever), but this just seemed belligerently pointless.
Eventually the person gets more acclimated to the second system on its own terms, and when the need arises to use a third system, they just go straight to learning the idioms of that system first, they don't try to make the third system conform to a previous system. I've got to think that the Pascal guy stopped using that header after a while, and didn't bother making a new one when circumstances made him move to a third language.
Anyway, so IMO that's the general pattern that people might think is off, because we've all had to deal with it in one form or another, but I don't think that's necessarily what's happening here.
Because you're fighting a losing battle.
There's absolutely no way to make Linux work the same way as macOS, simply because they do things differently by design. If you try, you're going to waste countless hours, have a horrible time, and at the most only make Linux slightly resemble macOS in terms of the experience.
I actually find that in terms of UX, Linux is a lot more similar to Windows than it is to macOS.
For me, what I missed of macOS when I tried out Linux was the easy shortcuts, easy access to special characters (accented letters, etc. without the compose key), the focus on drag-and-drop to do things, the quality of the look & feel of most apps, and most importantly the consistency that macOS has. There is of course no way to make all Linux developers stick to the exact same guidelines, but I had a feeling on Linux that every time I installed an app, I had to relearn a lot of things. Shortcuts and other conventions are always the same on macOS, on Linux they change alongside many interface elements, down to the windows controls.
In general though, I had a better time when I embraced how Linux does things, rather than try to recreate my macOS workflow.
Ultimately, I have no interest in wasting hours and hours configuring and fixing things up, just so that I have freedom with the software on my computer. I don't like many decisions that Apple has taken lately, and don't especially like how expensive and outdated their hardware is, but my hourly rate is pretty good, and with the money I make from gained (or not lost) productivity because things just work I can buy not one, but probably 10 MacBook Pros, and save myself a lot of frustration. I can then have freedom not from using open source software, but because I'm done working and I can go grab a beer with my friends or girlfriend instead of staying home configuring X.org.
As a developer, switching to Linux can give you a very nice custom enviroment with CLI applications like mutt or zathura and a DE like xmonad or i3. But you can't have a GUI as polished as MacOS, although things are getting closer.
One of the big innovations of Linux was package management. So using homebrew is indeed a bit absurd.
Actually, I would argue next-generation package managers like nix or guix are extremely interesting and have tons of advantages. Maximizing benefits of a MacOS to Linux switch maybe makes it advisable to use one of those.
Why not?
> you can't have a GUI as polished as MacOS
Why not?
All it would take would be for a company to publish a Linux UI library and a standards doc, and to curate an app store. I don't see any reason that a Linux machine could not be made to look and feel like a Mac. There's no magic in OS X, just a lot of hard work.
I'm not saying this would be easy, only that I don't see any reason why it wouldn't be possible. And given how badly Apple is dropping the ball, I also don't see any reason why it couldn't be lucrative. When Google started, everyone thought that the search market was mature and that Yahoo and Microsoft would lead the computer industry forever. I don't see any reason Apple could not be unseated just as they were.
You lose 90% of the value of having UX that 'just works' if the driver support is buggy or isn't there. Developing and testing drivers is expensive.
Apple solves this problem by designing and selling the hardware as well, and controlling the whole stack. They then charge extra margins for the privilege of using a Mac.
MS solves this problem by shoving the driver work onto each hardware vendor.
So while you can make and publish a very polished Linux distro / library, what's the business plan for getting paid for it, when the UX only gets you a fraction of the way towards having a seamless 'out of the box' experience for users?
But as a counter-example, Intel® HD Graphics 530 can easily handle Half-Life 2, with power to spare.
Because for the last 20+ years, FOSS haven't managed to be consistent, employ professional level designers, not rewrite everything from scratch every 5-6 years, and maintain parity with commercial OS offerings at the Desktop GUI level.
Why? Because of lack of resources, programmers hoping to what's interesting and not to grunt work that must be done, people dropping out of projects and no formal structure to force a continuous experience so new members get to change whole direction, etc.
And plain lack of resources. GTK+, the core basis of Gnome and tons of apps, one of the more important libs, only had one person fully maintaining it (at least a few years ago -- the person complained publicly in blog posts).
Contrast with the Windows or Cocoa gui libs that have teams of tens of programmers.
>I'm not saying this would be easy, only that I don't see any reason why it wouldn't be possible.
Anything is possible, it's all 1s and 0s. Whatever it's there or at least probable is more important.
Of course, one can beat MacOS at its polished GUI game. But it requires coordination. A single distro can do it. I expected Ubuntu to follow this path.
With so much fragmentation it's not easy though. X11 vs Wayland, GNOME vs KDE, a myriad of desktop systems, etc.
However, fragmentation comes from the bazaar model that makes Linux so resilient. I would never trade this for a bit of polish.
Also, as a programmer, I prefer simple CLI applications that compose well than GUI silos. That's why the current situation does not bother me much.
That has historically been antithetic to free software culture and its bazaar style of development.
I'm not sure I agree with that. Frankly, I don't find macOS's GUI to be particularly polished, or even pleasant. Of course, people that like macOS usually do, but, well, that's somewhat self-selecting.
I honestly think thinks like Cinnamon or KDE are nicer that the macOS GUI.
On Linux we have a myriad of alternatives.
Personally, I came to the same conclusion recently. But I went the Hackintosh route instead of switching to a Linux/dual-booted Windows machine. The latest rMBP’s disappointing hardware and battery life, not to mention the cost premium, made the decision fairly easy for me. That, and I wanted to go back to using a separate desktop and laptop after repeatedly hitting performance bottlenecks with Adobe Lightroom on my previous rMBP. My photography pretty much nixed a Linux option unless I wanted to dual-boot into Windows more often than I’d prefer. There are alternatives like Darktable, but even though Lightroom is a resource pig that needs a diet and a lot of optimization work still, it’s my preferred catalog software that I’m most comfortable in. And when you’re trying to sort through 3-5k+ photos shot over a very busy weekend, comfort is the deciding factor.
As a developer, I have no problem with the popular Linux distros. I’ve worked with Linux servers for years, and I love the approach to package management package management. I’m perfectly comfortable with CLI apps. But when I’m not doing dev work, I don’t want to be saddled by the need to constantly boot into Windows for other tasks.
The Hackintosh route gave me the best of both worlds. Better hardware that I can upgrade on my own terms, my preferred GUI environment, and no need to dual-boot unless I want to game. I had a bit of extra work to deal with in the beginning to get everything working properly (with my Skylake build, proper audio support and the wireless networking card in particular), but it didn’t take much of an effort thanks to a very robust community ([tonymacx86.com]). The only other issue is checking on discussions for new software updates to make sure there aren’t any problems before I go and install them. But that’s not too big a deal, all things considered.
Tl;dr - I guess my point is that we all have our preferred environments and will go to a hell of a lot of trouble to maintain them. Or, if that’s not an option, then to at least get things close to them. Even when that can mean more trouble later on. Using a homebrew port strikes me as a less-than-advantageous approach, but I can imagine why the original author did so.
darkwake=8 worked for me on a Skylake setup (Z170X, Core i5, nvidia graphics card with web drivers). Sleep and wake work flawlessly.
From one phtographer to another, and not to be a dick, but if you're shooting that many photos in one weekend you're doing something wrong.. the only scenario where I can imagine reasonably coming up with that many photos is motion photography with burst. Easiest rule I can give you is to pretend you have an old-style camera with a limited amount of rolls. Imagine how many rolls would be reasonable for the session you're having if it was 1980? If you stick to that, your photo quality will dramatically shoot up and the effort you have to expend sifting will go down a lot.
In my personal work, that's not an issue as I shoot almost exclusively with manual focus glass (Loxia lenses, adapted glass, etc.). With focus magnification and peaking, it's very rare to ever have to reject a photo for blatant focus issues even when I'm using a lens with a razor-thin DOF at f/1.2.
Isn't it OBVIOUS? Because they very much like their existing platform and how it works, but just dislike some things enough (e.g. the hardware options offered lately) that they want to move away.
> However it turns out Linuxbrew manages it’s own version of all the packages it tracks (makes sense I guess). This end up duplicating many of the packages available or already installed by the system package manager. This works on macOS since there is no system package repository but did not sit well with me on a Linux system.
But I am still wondering why linuxbrew exists to begin with. I mean, I understand the point of having snap or flatpack in addition to native distro packages, but linuxbrew?
There are also these universal package formats that some directly from the developers.
The Linux 'solution' to this problem is to compile from sources. linuxbrew is a way of automating that, if you will.