Hacker News new | past | comments | ask | show | jobs | submit login
The Death of Xenix (1997) (linuxjournal.com)
41 points by tshtf on Feb 17, 2014 | hide | past | favorite | 23 comments



Choice quote: The Linux community must agree upon a single software installation and management scheme, just as it standardized file system layouts.


By and large, I think the community already has. dpkg is the clear winner, with the most widely used distributions (Mint, Ubuntu and Debian) all using it.

RPM isn't dead by any means, but I think its continued existence is merely an accident of history at this point. As more and more people and organizations migrate away from Fedora/Red Hat/CentOS and openSUSE/SUSE Linux, I think we'll see its relevance and importance drop off significantly.


Do you have any sources for "As more and more people and organizations migrate away from Fedora/Red Hat/CentOS and openSUSE/SUSE Linux"?

I would like to see one standard to win, but in my environment I do not see big companies going away from RedHat or SUSE. So I am curious to see if there is some trend building up.


Xenix was my first UNIX experience in 1993, after almost 10 years of coding experience in home computers.

That command line experience compared with the Amiga GUI of the time was a bit of meh experience.

The only thing left was my curiosity for OS design and delve into UNIX, while at the same time my teacher gave me the sad news that there wasn't a Turbo Pascal compatible dialect for it and I really needed to use C.


Whenever these old articles get posted, I find myself reading them almost entirely for the sense of borrowed nostalgia. I was playing games on my C64 at the time, maybe playing around with BASIC, not this, so it's not real nostalgia. Still, I love reading about computer history that I never experienced and thinking "oh gee what a simpler time that was..."


I have this crazy idea that Microsoft could pull itself out of the "meh" gutter by becoming a Unix OEM again.

As a longtime developer, I'm certainly not representative of the general computing population, but I have to think that part of Microsoft's long and painful decline is due to a new generation of computer geeks simply not wanting to use Windows. If Valve can break through the Linux gaming barrier with Steam OS, this could spell the eventual end for Windows as an OS that anyone voluntarily uses.

In my (admittedly naive) view, I see legions of young computer hackers who want to write apps, games, and server things. I see them asking the "elders of the internet" what system they should use for programming, and I see them all getting the same answer: "get a Linux box, or at least a Mac if you like it nice and shiny". So they go off and get a Linux box or a Mac. The point is: what they get is a Unix system, and that's what they will come to be familiar with.

At some point their parents or friends will ask them what kind of computer they should get. If they don't say "why do you need a computer? Just get a tablet." they'll probably steer them towards a Mac... partly because of the "it just works" reputation and partly because, well, it's a Unix system, and they KNOW this. And so goes the cycle of OS popularity. A large portion of "what tech should I buy" seems to trickle down from the tech geeks in every family or peer group, and as time goes on there will probably be more and more of those.

So where does Windows fit into this? Visual Studio is nice but otherwise everything is just... different. It used to be that Windows was "normal" and UNIXy things were a bit different, because everyone "normal" used Windows. Now, that situation is becoming reversed.

Personally, I quite like what Microsoft has been doing lately, but no matter how shiny I think their new stuff is , I really don't want a non-UNIX OS on my desktop. I want a normal command line with bash or zsh. I want a homebrew or apt-get package management system, and I want a system where no matter what software package or library I want, I know I can always git clone or untar it, then run ./configure && make && make install.

You can kind of get there on windows with SFU and Cygwin, but it still doesn't feel quite right. I think Microsoft could gain a strong second following with the tech geek crowd by committing to a complete POSIX layer, proper fork(), a unified file system, a proper terminal (even though powershell is cool), and an "it just works" philosophy when it comes to GCC and clang and all the open source software and libraries available online. If they can do this before their "windows is for gamers" window closes, I think they can get back a lot of their lost tech geek crowd and then enjoy the reputation and recommendation trickle down that inevitably follows.

I have absolutely no evidence to back up this advice; I just know that if Microsoft said "Windows 10 will be a certified UNIX", I would strongly consider replacing my Mac with one of those nice Lenovo Yoga laptops.


I see legions of young computer hackers who want to write apps, games, and server things.

These legions increasingly accept corporate approval for iOS and Android app stores as natural phenomena. Success looks to them like SEO and IAP and Advertising revenue. The altars of Stallman and Linus dim. The lsighted logos of their laptops, phones, slates, proclaim that Jobs is their brand of hero.

Linux hackers are running Linux now.


Suppose Microsoft does this with Windows X -- releases a version of Windows with an top-notch UNIX subsystem and excellent tooling.

All the geeks tell their families to buy Windows PCs, because it "has UNIX underpinnings!" Microsoft splits the UNIX workstation market right down the middle with Mac, and Microsoft's share of the PC market increases from 92% to 96%.

How does this help Microsoft financially?

Desktop Windows accounts for only 23% of Microsoft's revenues. So Microsoft's overall revenue increases by 1%. But its costs also increase, because they will have to maintain this UNIX environment. Its costs may even increase by more than 1%, because the dev work won't be trivial. Not great for profits.

Does it help them sell Windows Server licenses? No. Developers code on the Macbooks today and deploy to Linux -- not to Mac OS X Server.

Does it help them sell Azure? I can't see how. Same logic as Windows Server. And Microsoft tries to make a profit. Amazon does not. Thus, Amazon can beat them at pricing.

Does it help them sell Office? There's already a version of Office for the Mac.

Does it help them sell smartphones and tablets? There's still one heck of a chicken-and-egg problem. Developers are more likely to run the Android tools on Windows X. In fact, the Android tools will now run even better on Windows X, because of the improved UNIX support.

It does get Microsoft some "street cred," but how do you turn street cred into profits? Maybe if Microsoft had done this in 1999, then it could've prevented OS X from taking hold as the preferred UNIX workstation of the developer crowd. But in 2014, with Linux and MacOS and Android and iOS already well-established?

Frontal assaults are a hard way to win a war. Better to go around the flank. Apple had single-digit share of the PC market when it scored a runaway hit with the iPod.

P.S. Services for UNIX has been deprecated. The last release was in 2010.


Maybe if Microsoft had done this in 1999, then it could've prevented OS X from taking hold as the preferred UNIX workstation of the developer crowd.

What is funny is that I believe MS acquired Interix in 1999.


And was POSIX.1 compliant before that.

http://en.wikipedia.org/wiki/Microsoft_POSIX_subsystem


At least, it would be nice if MS could help put an end to the infamous SCO lawsuit. Of course, even SCO OpenServer/UnixWare is outdated nowadays and probably would require a lot of work to run well on modern hardware.


instead of unixifying windows, wouldn't it be easier for microsoft to run a windowsified linux distro in parallel to their pure windows line?


Ah, one of the early 386 protected mode OSes, developed back when MS/IBM was focusing on the 286 for OS/2. To be honest, I think 386s was expensive back then.


> Ah, one of the early 386 protected mode OSes, developed back when MS/IBM was focusing on the 286 for OS/2. To be honest, I think 386s was expensive back then.

I don't think this is accurate. Xenix pre-existed the 386 and was used on other processors before and after it supported the 386.

Since I thought maybe I was going crazy, I checked Wikipedia and read, for example, "Version 2.0 of Xenix was released in 1985 and was based on UNIX System V. An update numbered 2.1.1 added support for the Intel 80286 processor"


Couldn't Xenix run even on an 8086? I don't think it required an MMU at all.


I know. I am talking in the sense that its 386 port predated other 386 OSes.


There were successive Xenix versions for 8086, 80286 and 80386.


I remember Xenix as an available OS on the $9K-ish TRS-80 Model 16 (Z80 for I/O + 68000 for processing).


Absolutely. A TRS-80 running Microsoft Xenix was my first Unix-like, back in high school. 6-12 terminals full of students typing in COBOL from their coding forms. It replaced an older Burroughs with a card punch. I thought I was learning some obscure Tandy environment. (I was a Commodore kid, looking down on the "Trash 80"). Little did I know I'd be using a version of vi and sh for the next 30 years.


I remember Xenix as the OS that wouldn't recognize certain brands of floppy drives or many VGA cards, because it didn't support going through BIOS. This brings back memories.

Insert floppy disk 1, boots...

Asks for floppy disk 2. Insert it.... Floppy disk drive not found. (granted this was a laptop that Xenix wasn't designed to support)

Or insert floppy disk 1, boot....

Screen goes black......

Basically with Xenix (at least the versions I played with) your choices of hardware were so limited you needed to use systems specifically designed for the OS.


Of course, now all modern OSes no longer go through BIOS for standard VGA registers or floppy drive access.


But Xenix didn't even support standard VGA registers or any standards regarding floppy drives. It required internal knowledge of specific products to work at all.


Most of us used it via serial port.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: