Continuing my theme of looking at some of John Gruber's early articles, he wrote a lot in the early years about things such as bringing fresh software (or porting old software) to Mac OS X. It's interesting to look back on it, because I had such a "Mac OS 9" upgrade/switch experience. A few of the articles focus about people who came to Mac OS X by way of NeXT, Linux, and other UNIX platforms. Gruber holds pretty consistently that the UNIX Switchers (as he calls them) were pretty continuously amazed by Mac OS X.
In that way, Mac OS X was a little bit of a battleground for a while. A lot of Mac users (and even some people coming in from NeXT) thought that the platform would be marred by the mere availability of popular UNIX software (such as OpenOffice, and Mozilla) because if you can run your software in X11, and it still somehow counts as "Mac software" then developers wouldn't bother to build "real" Mac software, because the OS provides enough for UNIX people to get their work done, and the X11 environment let them port other important graphical software quickly.
In the real world, what ended up happening is the more important something was likely to be to "regular people" the more likely it was that somebody ported it to Aqua or built a commercial version. That doesn't stop the early days of Mac OS X from having been a little bit of a minefield, and attitudes didn't help. A commonly cited attitude in the early days was that if you want to see OpenOffice (or another piece of software) on Aqua, you could help port itself.
I question exactly whether the people harboring that attitude had considered to whom they were talking, though. Most of these comments came from people who were already developers at a hobbyist or professional level, and even those who were, were probably not going to take their time to contribute to open source UNIX projects, when they have their own Mac-focused software to build.
Another suggestion at the time was that Apple provide resources and manpower contributing to user-facing open source project, which has some (but not much) credence. The main problem with this is that Apple was (at the time) a pretty small company. They almost certainly didn't have the workforce to do that project, and working on OpenOffice has never really been in Apple's interest, because their biggest focuses were showing off what you could do porting an application forward from Mac OS 9 (with AppleWorks 6) and of course, selling copies of AppleWorks (and later: iWork.)
I don't think there's any solid (or even nebulous) numbers for how many people switched from "UNIX" to Mac OS X, but there are two groups in this category. The first is people who would have been casual users of something that ran on x86 systems (probably Linux) because they were looking for something UNIX-like on the cheap. For that group, Mac OS X probably looked refreshing because you could get on an iMac or iBook for $1499 if you were feeling flush with cash, or a Power Mac for that much, if you already have a display to use.
The second is the group came from commercial UNIX vendors such as SGI, Sun, DECpaq, et al. They were used to paying over $10,000 for reasonably nice workstations (anything better than an O2 or Blade 100.) Even basic workstations (such as the aforementioned O2 and Blade 100/150) cost $5000 or so for a "good" configuration, and are considered to have been unsuccessful. The Blade 150's starting price was either $995 or $1395, depending on where you look, but that system was nearly universally reviled. The Silicon Graphics O2 started at about $5000. The Blade 1000, which is considered to be an actually good UNIX workstation, started at about $2000. In comparison, a really great Power Macintosh G4 was going to cost about $3000, and if you have $5000 to dump into it, you can max the RAM, get the highest end graphics card, one or two monitors, and so on.
The other thing that must have been a revelation for the UNIX switchers at the time was that Mac OS X had Terminal.app, as well as the X11 and Aqua environments, so UNIX software was easily ported.
But that brings us back to the point about Mac OS 9 users, which were essentially "switching" to Mac OS X as a new platform. Most Mac OS 9 users were worried about the state of the OS, because from their perspective, it was a mess compared to Mac OS 9. Mac OS X didn't regain a lot of the features and the Finder wasn't as nice as it had been until 10.5 was released in 2008, and a lot of people say that there are still things the Mac OS X Finder isn't doing as well as the Mac OS 9 Finder did.
With that in mind, it's not hard to imagine why Macintosh to Mac OS X switchers who were even aware of the UNIX capabilities of OS X were unhappy about the prospective that Linux and commercial UNIX users might simply accept the fate of mis-matched graphical applications. What's even worse is that in the early days, new software getting ported to OS X, such as Matlab, wasn't always free. The original version of Matlab required an OS X, cost $1900 (which is typical of technical computing software like Matlab) and required an X server, making it inaccessible to people who want to do this type of task, but are using a Mac because they don't want to become proficient at UNIX (and even worse: the intricacies of using OS X for graphical UNIX software) in order to do their research. (I suppose the question there is whether or not you were already using that tool, or if one of the promises of Mac OS X was that you'd gain some of those capabilities.)
X11 was talked about specifically a bit later. The Mac Experience was (and continues to be) deeply important to Mac users. This is true to the point that even products Apple built and sold to Mac users were rejected, because they weren't Mac-like.
Things ultimately turned out, but in the beginning, there was a real fear that Apple was building "the best UNIX platform" and not "the best Mac platform", which is what they claimed to be building. The fear isn't completely unfounded: Many of the people responsible for the architecture of Mac OS X came from NeXT, which was one of the UNIX platforms from which people switched. Mac users have long been very particular about their systems and the software that runs on them, a right I'd say they've earned, because this is a culture that Apple definitely cultivated, and helped along, with the software built by its fully owned Claris subsidiary.
Mac OS X didn't stop having two large groups of users for a long time. I'd argue that it still has this nature, as people struggle with the concept of using it as a commercial UNIX system (in lieu of just using Linux on desktops) to this day. The difference now is that most Mac (and even technical bloggers like John Gruber) don't need to focus on what popular Linux and UNIX people are writing, because Mac OS X is totally flush with really good, really mac-like software.
Perhaps the biggest problems with Gruber's characterization of the UNIX switchers (as characterized by the "Spray On Usability" series, is that he doesn't just suggest that UNIX users were switching to Mac OS X, but that by 2003, the only ones left anywhere other than Mac OS X were either "too cheap" to switch, or political zealots. I'd argue that this is both too heavy handed, and simply incorrect. In 2003, the following platforms were still viable UNIX "desktop" platforms that were priced far out of the typical reach of a Mac user:
- Solaris on SPARC hardware
- HP-UX on Itanium hardware
- HP-UX on HPPA hardware
- AIX on POWER hardware
- Tru64 on Alpha hardware
- IRIX on MIPS hardware
So it's unfair to say that all UNIX users were hanging onto their old systems simply because they didn't like usability. Though, the higher end you get, the more likely it is that people are using the system through a console in order to do The heavy handedness Gruber applied at the time (and really has been since then) really pertains to the fact that linux users were either cheap, proving a software-political point, or "didn't care about good UI design." (Paraphrased.) I think that it's fair to believe that perhaps there's a third possibility: they may care, or even recognize good UI design, but may not think Apple has something that meets their personal needs.
Gruber's notion of UNIX as something that should only exist on the back end is somewhat problematic. There are reasonable uses for the X11 stack, and Mac OS X is one of the best desktop platforms (to this day) for people who spend a large amount of time using SSH to connect remotely to a large number of systems, for whatever purpose.
Bad ports are not ideal, and there are a lot of bad ports (or just poorly designed applications) to this day, but I don't specifically agree that we need to eliminate the use of UNIX (or keep it on the "back end" and out of client applications entirely) in order to create a good user experience.
Perhaps the direction to look at this is three-tier application design. When designing a web site or an enterprise application, it tends to exist in three tiers. The first is the data layer. This can be files on a disk, or an outboard database application such as Oracle SQL, Microsoft SQL, or MySQL. The second layer is the application logic layer, where the program does its work. The third layer is the user interface.
There are a lot of advantages to this, including modularity, maintainability, and the fact that you can build different user interfaces for different devices, operating systems, or use cases. It doesn't always show up in desktop software, but a great example of it is the Transmission BitTorrent client. This design is, as far as I can tell, not very common in desktop software, but would be a clever way to make porting software from one platform to the next much easier, and make the resultant quality of those ports way better. Transmission has, from what I'm told by people who use it, a very Mac-like interface, but having the flexibility to use the web or terminal interfaces (or let the application run without an active Interface) is of great utility.
People have switched back from the Mac to other platforms as well (yours truly included) for whatever reasons. A lot of it's just aesthetic preferences, and a lot of it's covered in The Default Narrative, a piece I cite very often, because it really talks to the problematic way a lot of people talk about Apple's control over its ecosystem and things they add to Mac OS X. (In the case of The Default Narrative, some security features had just been added to Mac OS X, and a story was published online about some content that didn't make it into an App Store app. Everybody simply presumed that Apple had cracked down on this or denied it, when it really turned out that the publisher simply didn't choose to publish it on their app.)
Perhaps what's most interesting about the fact that Gruber spent so much time talking about the Mac as an emerging (and increasingly popular) UNIX platform is that it ended up playing some part in the death of most of the classic commercial RISC UNIX platforms. In 2002 and 2003 when he was writing all of this, you could still buy a Tru64 box from HP-Compaq, an IRIX box from SGI, and so on. Sun was working on making their entry level systems cheaper than ever before, and somehow, the combination of Macs being both good and cheap, (comparatively) and the fact that they were easier to get your hands on, Intel and AMD both increasing CPU performance on x86 and x64 massively, Windows itself getting a lot better for "professional" tasks, among other things still managed to kill everything.
Today, this gets talked about in a few different ways. Many believe that Apple (and others) are complicit in some sort of conspiracy to kill UNIX, but I think that the efforts on Sun's part to build affordable hardware weren't really matched with an effort to make their OS run well on it, nor was the hardware any good, and it wasn't long before Sun completely reneged from building that type of computer, sticking with much higher end systems, even in the x86 and x64 parts of their product line. Even if the OS ran well on it, it would have been nearly impossible for "normal people" to use.
What would be interesting to know is what the people switching to Macs from UNIX systems were buying. Macs were reaching all-time lows in pricing by 2001, with iMacs that year going as low as $799 for configurations that probably ran OS X as well as the Blade 150 ran Solaris, and had cheaper upgrades available. You'd spend a bit more on a Power Mac, but you'd also get a lot more computer for the money.
I have long thought of the highest end Quadras and Power Macintoshes to be workstation aspirational, and given that a Power Macintosh G4 had basically the same hardware as a Sun Blade 150, an OS that was liked better by end users, and is said to have had much better performance, I'd agree that 2001-2003 or so is when Apple really crossed that line. It was around this time (perhaps even a few years earlier) that Apple had started putting a lot of effort into frequently showing the advancement of the platform and what you could do with it. Even Mac OS 8.5 and on earlier systems may not have been a "UNIX computer" (until you bought your copy of Mac OS X, of course) but was perfectly competent at video and other content creation work, in ways that most UNIX workstation vendors weren't really able to show easily.
All of that, and the fact that Mac OS X comes out of Apple's acquisition of NeXT, and Apple never directly referred to any of its computers as a "workstation" (even in the somewhat famous "sends other boxes to /dev/null" ad) until 2006 when the Intel-based Mac Pro launched. Even the Power Macintosh G5, which Apple claimed gave them a huge lead, was simply a "64-bit Personal Computer."
I don't know what the cause for this was. Perhaps Apple didn't actually see themselves as a serious UNIX vendor, or perhaps even if they saw the obvious UNIX vendor behavior, they did not want to muddle their marketing strategy, which still centered heavily on the "quadrants" of personal computing. Another possibility is that they were fully aware of what they were building, and that it was still a far cry from the midrange and high end workstations such as the SGI Octane and Sun Blade 1000 and 2000.
I need to do some more digging about the process by which Apple came to have its hands on the most popular (and perhaps most viable) commercial desktop UNIX platform. I'd like to take a look at it from the other side, but that'll involve some digging.