One of my greatest bits of luck was getting access to these and other Unix workstations&servers as a teen.
If you run old SPARCstation OSes, a few points of interest...
If you get the OpenWindows era with Open Look look&feel, one of the things that were newish around then is the pinnable menus and dialogs. You could sometimes tear off a menu and turn it into a persistent window that's more like a toolbar (which was also emerging then).
Open Look (or Sun) was also introducing small rectangular drag&drop target widget on windows, when drag&drop was new. Which I don't recall seeing later, so maybe it was a mix of less drag&drop than we thought we might have, and more an expectation of the user to learn what things could be dragged and where they could be dropped.
Behind Open Look also a number of different toolkits and window system backends that might be behind a given Open Look application and desktop. My favorite toolkit was XView, and it was very easy to work with in C, and used varargs sorta like GTK later did. Desktop-wise, besides X, there was the XNeWS server, which supported a kind of display PostScript on steroids, and you could write entire applications in PostScript (though their OpenLook widgets looked poor early on, compared to the pixel-perfect XView result).
Maybe the key thing to notice about OpenLook... At the time, the other main X GUI look&feel that workstations were moving to was Motif, which was mostly just rectangles. Someone once posted screenshots of a dialog in Open Look and then another in Motif, with all the text removed, and, in Open Look, you could pretty much tell what all the familiar widgets were, and you immediately knew how to interact with all of them. That was considered valuable at the time. (Though HCI in general has moved more towards brochure-like and magazine-like motivations in the interests of the creator, and away from motivations we had before, for new-user and power-user usability in the interests of the user.)
If you go back a little further with your workstation, to SunOS 4.1.x or 4.0.x, you can forget about X and NeWS, and see something a bit different: SunView. This was a monochrome GUI, maybe sorta like an old Mac, but without all the things the Mac had, and backed by much more sophisticated workstation/server OS. People did powerful things with it, on a workstation with a megapixel display and 8MB of RAM. The GUI toolkit was the predecessor to XView, and similar.
For workstation application programs (like my employer developed), it wasn't unusual to implement your own portable GUI layer, including an entire window systems. So, our stuff and Interleaf's, for example, would create a native window for whatever workstation desktop was in use, and implement our own desktop within that. And they were all different, and often innovative, sometimes in ways that have been lost.
I didn't see how all the different CAD, EDA, etc. packages worked, but sometimes the workstation was pretty much an appliance for running a single program, and on occasion would even be branded with a physical badge for the application software on it.
You can rediscover some of the same flavor still alive by looking at UI innovations by, for example, Blender, where (last I used it) they built UI from scratch, and they seemed to innovate in the interests of the users, including power users. Maybe more accessible to us on HN, some IDEs are places that get creative innovations often in the interests of power users.)
You'll also see innovations less-familiar in some other workstations, like Apollo Domain. (A lot of Sun stuff might be a little too familiar, because they sorta won, albeit in the form of Linux and the BSDs today. SunOS was also arguably the easiest out-of-box platform for building GNU-ish software on the Internet, before Linux was sufficiently ready, so that's another regard in which they both influenced and were influenced.)
Also, if you end up trying a real-metal vintage workstation experience, don't forget to hit the degauss button on your 21" CRT, be startled by the sounds and the ray behavior, and each time wonder in the back of your mind whether there's any health risk to be standing near it each time that happens. :)
> I didn't see how all the different CAD, EDA, etc. packages worked, but sometimes the workstation was pretty much an appliance for running a single program, and on occasion would even be branded with a physical badge for the application software on it.
So true.
One of my first jobs was installing/supporting some big CAD/CAM/CAE software package; mainly for the automotive industry but also a lot for toy factories and some others. The thing is not only was the workstation an appliance for our software, but many times Sun, SGI, HP, IBM, they actually sold their workstations through us. We'd sell the software and offer the client, if they didn't have the appropriate stations, to get them a nice deal. We'd be the ones doing the hardware delivery and set-up, in exchange for the deal, I guess.
I particularly remember one such delivery. It was a fancy design studio where they did like perfume bottles, luxury products' packaging, etc. They were running everything on Macintosh boxes. Around that time, Apple design was all curves and cool colours and all that. The office itself was quite well decorated and also focused on curves and nice colours. Anyway, I got there with a very rectangular, very grey-and-ugly-purple, very heavy HP box. All the designers were pretty appalled that I would set that thing right there in their office just because some "stupid software" needed to run on it.
It was even worse when they finally saw the actual, rather industrial, software running xD
Sun was making an effort, like when they went from sturdy industrial/utilitarian Sun-3 looks, to the dimples and lavender accents of the SPARCstation 1 and 3/80. :)
My first impression, back then: this is a 5-figure engineering workstation running 5-figure engineering software. Why are they trying to make it look like a $15 plastic consumer hair dryer.
I actually like Sun boxes. They look simple but stylish enough.
[Just in case] The box I dropped at that studio was, sadly, an HP. It may have been a Visualize X-Class [0], which was ugly but at least it was new and modern at the moment. But most probably it was just a Kayak XU [1]. Those things were ugly and they weighted a ton and we dumped them everywhere with some 3D card and a RAM upgrade.
To a designer, I would guess those really look like midtower PCs someone was told to spiff up, and they came out looking... much like midtower PCs, with powder blue and other bits. :)
HP also made some other attempts...
Our HP 9000/3xx HP-UX porting workstations had a decent very sturdy lab HP modular/stackable boxes, with a reasonably stylish grill that unified them. It also wasn't a mismatch, now that I think of it, for some other HP gear I had personally, like a 2392A terminal I had.
HP 9000/7xx (HP-UX on the hot new HP-PA architecture) were different, and not as carefully standardized-looking that I recall.
I once saw some very interesting minicomputer(?) design in a museum that might've been HP? It was built into a desk.
For targeting cosmetic-sensitive designers, pivoting engineering workstation vendors could've taken a cue from Next, or some of the more clean white Macs over the years (even the Apple IIgs). Or gone wild and daring like SGI (which was maybe analogous to the fruit iMacs, in that it might look cheap if you didn't know what it was, and it really wanted attention).
Update: Here's what Sun did with the PC midtower, when they made the 386i to run SunOS 4 with SunView, and possibly be a gateway for PC users to migrate. I used one of these for a while as my main workstation, and it was nice. It looked a bit like a PC midtower that had been refined and had high build quality. https://web.archive.org/web/20160402201116/http://sites.inka...
Back in one of my first mom-and-pop ISP jobs, my boss was interested enough in exotic hardware that he decided to buy a Sun Microelectronics motherboard with an Ultrasparc IIi CPU. We were able to put it in a large form factor ATX case, find a compatible video card and Ultra2 SCSI drives and install Solaris 2.7 (or 2.8?) on it. We later deployed it as an LDAP and MySQL server amongst other things. It was a real eye opener for me having come from the world of Windows/Linux on x86.
Later he bought a bunch of DEC Alpha motherboards and CPU's as well, I had fun installing Debian and compiling pre-1.0 versions of Gnome for them. These experiences made me far more comfortable with non-x86 systems and I'm grateful for that exposure more than 20 years later.
> For added fun, a Sun Type 5 keyboard converted to USB
We must have _very_ different definitions of fun. The pile of mush that was the Type 5 is up pretty high on my list of "never again" keyboards. I was a full on Sun fanboy, but even I was the first to admit that their input devices could be pretty awful.
I don't know whether to blame the Type 4, something about our desks, or something in the water... but our entire development team were getting RSI problems, like I haven't heard of before or since.
(Fortunately, there was chatter on the Internet about this problem at the time, and I learned the very simple workstation adjustments that worked for me. Ever since, I've been able to type like crazy, with no discomfort. To this day, I give everyone The Talk about typing RSI injuries, the set of usual culprits I know about, so you can adjust and see what works for you, and the absolute necessity of not working through typing pain. And if someone is noticing typing-related pain, they need to get a referral to a specialist(s), they need to keep on the specialist so they don't fall through the cracks, and they need to do what the specialist says.)
My understanding is that what works for a person varies, but here's some of the approx. dozen changes that I ended up making for myself (sorry have to be quick right now):
* lower keyboard shelf so forearms could be parallel to ground. I've found many desks can be lowered. one place, I found the mount for the detachable commercial-grade desk legs matched that of some lower side table in our lab area; so, in a late-night operation, the table used for lunch catering trays got strangely taller, and I suddenly looked strangely taller next to my desk height.
* wrists straight in line with forearms, as viewed from both side and from top. (I actually found my first-in-ages typing soreness briefly last year, when I used some laptop with a numeric keypad, and i realized that one of my hands was twisted to the side awkwardly to keep the screen centered. fortunately, that was only for a couple days, and i was able to switch back to a sane laptop
* some people like wrist rests, but others say wrist rests encourage bad positions or pressure; i tried them then switched to without, but i suppose they might work for some people, and i don't know the research
* both keyboard and monitor straight ahead of torso, not to one side
* maybe eyes even with top of monitor (this advice might've been intended for 13-17" 4:3 CRTs, and I'm not really good about it)
* arms should hang comfortably at sides. if I can't get a chair without armrests, usually they can be removed with a few bolts.
* feet flat on the floor might've helped
* some people benefit from typing breaks. and also from vision breaks, and stopping to focus on things in distance. i didn't seem to need either for any kind of discomfort (though I am a bit near-sighted, so maybe vision breaks would help)
* stress seems to contribute
* pressing simultaneous-multi-key combinations with the same hand, especially if stretching/twisting/straining the hand, is bad. one big thing I did with Emacs is switch from holding down Alt for Meta, to pressing and releasing Esc (except for Meta key combinations I might be pressing many times, like for undo or for incremental search, in which case I could hold down Al). I also tried to use other modifier keys with two hands -- one hand for the modifier, the other for the non-modifier.
* there was some of talk about the RSI from switching hands back and forth between finding the mouse and finding home row. Also, Sun used this Mouse Systems mouse with an optical pad that has to be placed, rather than maybe the body naturally finding where a good area of travel for the mouse is. And the mouse buttons could've been easier to press.
* The Sun keyboards also had a set of special function keys for things like copy&paste, and I don't know whether moving off home row to reach for those and then go back to find home row is a significant strain.
* (Later on, I switched to Trackpoint keyboards, which let my fingers stay around home row, and requires only very slight movement and pressure. I vaguely recall first trying Trackpoint when in a very stressful situation, and i did get some soreness in my Trackpoint finger, IIRC, but it soon went away. I mentioned this to the Trackpoint inventor, and he seemed unsurprised, but maybe slightly sad that Trackpoint wasn't perfect in all ways. I suppose the newer designs of Trackpoint cap might also help, rather than the hard sandpaper one I was probably using at the time. Today, I put 15-year-old new-old-stock Trackpoint keyboards into 10-year-old Thinkpads as my primary workstation.)
* (I also now have a "lectern" standing desk, $200 on Amazon, that fits a laptop, beverage coaster, and a small device like smartphone or embedded device. I never bother to adjust the height, but it's good to stand for chunks of time, and it can be moved around to wherever the light is good, view out the window, etc. Some partners and customers require videoconf apps that I don't trust on workstation, so those go on an iPad that gets put on a little arm clamped on back of the lectern as needed; and then the arm is mostly hidden when not in use, by clamping upside-down to the back of the lectern.)
* (Today, I don't have to think about it, and I just do whatever is comfortable/natural, which I suspect doesn't look like the good posture that the earlier bullet points in this list might sound like. but I suspect they're good starting points.)
I want to emphasize that these things are preventative measures, and based on folk wisdom. It might be worthwhile to bring a workstation ergonomics expert through office proactively (after Covid), to look at the setups and practices of anyone who wants. And if anyone is already seeing symptoms, it's crucial that they not try to "work through the pain", because that can cause permanent damage. Including permanently no longer being able to type at all, and/or chronic pain. If you're having trouble, it's important to find a doctor who'll take it seriously, and then to take the doctor's directives seriously.
Hey, I'm moving the mouse but the pointer isn't going much of anywhere! What's wrong?!
Oh yeah, the mouse isn't CORRECTLY ALIGNED with the mouse pad! I'd better check if the mouse pad has moved out of its proper position on the desk. And pay a bit more attention to how I'm holding the mouse so that I don't introduce too much rotation about the vertical axis.
That alignment was something I never had to worry about with any type of mouse I used before or after. On the other hand, it worked pretty well when it was lined up right. It really seemed to track motion more smoothly and accurately than mechanical mice.
This thread brings me pleasant memories of drunken post student union bar C hacking sessions in the 90s. Apart from the mice. And openwindows. Fuck both of them!
You're not thinking of the types 4 or 6? They were both horrible sponges.
The Type 5 is also a quiet rubber-dome-with-slider model, but pleasantly tactile with very high quality keycaps. (I'm fond of it, and so inclined to be defensive - in fact I'm using a Type 5c right now. It was my favourite among the 90s Unix vendors.)
yeah, I still have a 5 around for my 32 bit sparc machine... 4s were worse, and I dunno if I ever used a 6. The 5 is well built, great caps, looks pretty, interesting and mostly good layout... I just could never be happy with those domes, and it was a huge letdown.
Plus RS/6000s came w/ a Model M, so clearly _not_ the best 90s unix keyboard. :P
As a teenager, I scored a sweet job installing new SPARC 5 workstations at an electronics company for a few weeks during an otherwise quiet summer. Man, those were some amazing computers. The cases opened with a thumbscrew and were so easy to work with. Elegant design. Super robust. For a kid who had only ever worked with cheap PCs, Sun hardware was an absolute dream.
I also liked having to replace some hardware on one (maybe a disk, was a while back), and wondering what tools I'd need.
I opened up the case (just a thumbscrew or latch) very easily, to find a Sun branded official screwdriver sat inside the case in a tool holder, ready for me to use.
Was a nice sign someone really cared about being able to work easily on those machines.
Had an IPX at that time too, stacked high with peripheral boxes. It acted as my auto-dial gateway to the internet. Of course that was connected to an HP 700 series workstation, a SGI Indigo^2, an IBM PS/2, and a Linux box. The good old days of finding surplus workstations at computer flea markets.
It seems like it's basically a commercial GPS SoC where people have figured out how to run their own code on the CPU that is part of the GPS SoC.
The IC in question here is a Venus828F. It seems like navspark is either a project of SkyTraq (the company that makes the Venus IC), or SkyTraq have a SDK which this packages in a blob format.
I've used most of the SPARC stations during the late 90s. 4, 5 and 20. They were a lot better than the Pentiums in use. Moved up to the UltraSPARCs, my last one was a Blade 2000. Linux was a champ on those architectures. But I don't miss the blasted bus errors not one bit. AMD is where it's at nowadays, ARM will be next within the next few years.
Just remember that out-of-the-box Solaris 2 had an absolute ton of remote root exploits. So, unless you have access to security patches don't allow inbound connections to the VM. Even with a full patch set it's probably not a good idea.
Is there an army of hackers waiting around to exploit a 30 year old operating system? And not even Windows; but Solaris. Notwithstanding, quite sure anyone installing this is doing it to fiddle, not trade crypto.
I remember it caused some commands to work differenty / no longer with same syntax. Cannot remember a single example though, so it must yave been a while...
Ironically, I got into the sysadmin (and later programming) business because a wise coworker of mine told me to stop mucking around in Windows 95 and grab the Sparc LX in the corner of the shop and install RedHat on it.
Because of the domain I was hoping this was going to be a hardware guide. But I really do appreciate this type of in depth emulation tutorial, because the documentation for these is usually geared towards people who A) are very familiar with emulation and B) are very familiar with the system in question.
Read the article but it did not mention anything about the software licences. Have these old versions been released for free in the meantime?
Copyright expires 7O years after the death of the author. Not really suitable for software lifecycles. Nothing from the 1990s has any commercial value any more.
Sadly there are no free releases of these old versions available. Not even sure it's legal to install from authentic media if the hardware is not real SUN hardware...
If you run old SPARCstation OSes, a few points of interest...
If you get the OpenWindows era with Open Look look&feel, one of the things that were newish around then is the pinnable menus and dialogs. You could sometimes tear off a menu and turn it into a persistent window that's more like a toolbar (which was also emerging then).
Open Look (or Sun) was also introducing small rectangular drag&drop target widget on windows, when drag&drop was new. Which I don't recall seeing later, so maybe it was a mix of less drag&drop than we thought we might have, and more an expectation of the user to learn what things could be dragged and where they could be dropped.
Behind Open Look also a number of different toolkits and window system backends that might be behind a given Open Look application and desktop. My favorite toolkit was XView, and it was very easy to work with in C, and used varargs sorta like GTK later did. Desktop-wise, besides X, there was the XNeWS server, which supported a kind of display PostScript on steroids, and you could write entire applications in PostScript (though their OpenLook widgets looked poor early on, compared to the pixel-perfect XView result).
Maybe the key thing to notice about OpenLook... At the time, the other main X GUI look&feel that workstations were moving to was Motif, which was mostly just rectangles. Someone once posted screenshots of a dialog in Open Look and then another in Motif, with all the text removed, and, in Open Look, you could pretty much tell what all the familiar widgets were, and you immediately knew how to interact with all of them. That was considered valuable at the time. (Though HCI in general has moved more towards brochure-like and magazine-like motivations in the interests of the creator, and away from motivations we had before, for new-user and power-user usability in the interests of the user.)
If you go back a little further with your workstation, to SunOS 4.1.x or 4.0.x, you can forget about X and NeWS, and see something a bit different: SunView. This was a monochrome GUI, maybe sorta like an old Mac, but without all the things the Mac had, and backed by much more sophisticated workstation/server OS. People did powerful things with it, on a workstation with a megapixel display and 8MB of RAM. The GUI toolkit was the predecessor to XView, and similar.
For workstation application programs (like my employer developed), it wasn't unusual to implement your own portable GUI layer, including an entire window systems. So, our stuff and Interleaf's, for example, would create a native window for whatever workstation desktop was in use, and implement our own desktop within that. And they were all different, and often innovative, sometimes in ways that have been lost.
I didn't see how all the different CAD, EDA, etc. packages worked, but sometimes the workstation was pretty much an appliance for running a single program, and on occasion would even be branded with a physical badge for the application software on it.
You can rediscover some of the same flavor still alive by looking at UI innovations by, for example, Blender, where (last I used it) they built UI from scratch, and they seemed to innovate in the interests of the users, including power users. Maybe more accessible to us on HN, some IDEs are places that get creative innovations often in the interests of power users.)
You'll also see innovations less-familiar in some other workstations, like Apollo Domain. (A lot of Sun stuff might be a little too familiar, because they sorta won, albeit in the form of Linux and the BSDs today. SunOS was also arguably the easiest out-of-box platform for building GNU-ish software on the Internet, before Linux was sufficiently ready, so that's another regard in which they both influenced and were influenced.)
Also, if you end up trying a real-metal vintage workstation experience, don't forget to hit the degauss button on your 21" CRT, be startled by the sounds and the ray behavior, and each time wonder in the back of your mind whether there's any health risk to be standing near it each time that happens. :)