Challenge: Make UNIX the world's dominant operating system

And if you can make UNIX the world's dominant OS (instead of Windows), how does history change? Y2K bug deferred for 60 years, fewer viruses...hmm...
 
Actually some form of UNIX has been the dominant OS for quite a while. Taking into account the WWW, 99.9999% of webservers up until tthe latter part of the 90's ran on some form of UNIX box (Sun Solaris Box, OpenBSD, etc...). All the DNS servers at the top level still pretty much run on some form of UNIX box. Before they became pretty much appliances, most firewalls ran on some form of UNIX PC. And aside from financial institutions holding onto the old iron of AS/400's, IBM and HP Mainframes, etc.... a lot of business servers ran some form of UNIX up until the early-mid 2000's. Windows has yet to crack the high-end (i.e. e-bay/paypal runs their database servers off of a huge farm of Sun Microsystems E10K's IIRC...they're not Windows servers).

That said, I'm assuming you're talking about the home consumer market. A couple things could start a definitive POD in say 1989 or 1990:


(1) Apple releases it's grip on it's OS software in the 90's so as to allow for clones and running MacOS on Intel boxes.

(2) Microsoft slammed for antitrust violations for their dealings with computer manufacturers and retail outlets for pre-loading systems ONLY with some form of Windows. US Government allows for retailers to choose which OS's to offer for the platform (OS/2, Windows, DOS, OpenBSD, etc...)

The result of this is Microsoft would not have as much of a monopoly on things (or they might, depends on how they play things). Windows would nevertheless be vastly different as the company wouldn't be directing their product to fly in the face of established standards (e.g. for networking they tried to push their NETBEUI protocol rather than the de facto standard of TCP/IP. Cisco as it came up in usage in IT and grew as a company basically pwn'ed them and Microsoft ended up having NETBEUI to TCP/IP translation until I think either Windows NT or 2000). Windows Server products would probably end up being just a front-end for a flavour of BSD or SysV Unix. And desktop OS's will probably be a lot more like the Linux derivative products we see now such as Ubuntu which are more end-user friendly than the Linux of old.

As far as the home users, pretty much the same outcome as now but with no artificial monopoly by Microsoft. Probably a lot less crashes due to better OS level coding and better communications with hardware developers (rather than Microsoft's current practice of essentially dictating which direction they're going).

As far as viruses, worms, etc... those will still be around unfortunately. With home users, they're not going to get down to whatever is going on with the guts of the OS when they launch applications. Viruses generally weren't written for UNIX boxes as the admins would see the process right away via monitoring their process tables. With Windows the visibility isn't their mainly due to the user; most users frankly are relatively technical idiots as far as this is concerned (i.e. they just want to play games on their PC or surf the internet or email their grandkids or type up a term paper). Viruses only become successful if you don't see what's going on which in the case of Windows currently, you really don't unless you run some third-party tool or Microsoft add-on. This was the same with MacOS as well (remember the nVir virus that would pop up on the resource fork of all of the files? You could manually remove it with something like RedEdit). Viruses remember are an exploit on the unsuspecting masses using OS's that hide the guts of their operations from users that don't see and don't want to see their inner workings. They'll be around but just having a different set of code.

As far as the Y2K bug, this was mainly due to programming (2 digit instead of 4 digit years) on 'old iron', mainly NON-WINDOWS, NON-UNIX machines. So I don't see it deferred or having any issues.
 

Sachyriel

Banned
Girlfriends for certain programmers, who spend their time being manly men and chasing after anything their heart's desire desire's instead of their own geeky desires?
 
What flavour of UNIX? Does Mac OS count, being based on (and more recently, IS) UNIX?

I imagine a GUI of some sort would still be necessary for computing to be widespread. CLIs aren't exactly useful for a number of modern applications.

You could have Apple and Microsoft work together on a GUI OS, using UNIX as a codebase.
 
What flavour of UNIX? Does Mac OS count, being based on (and more recently, IS) UNIX?

I imagine a GUI of some sort would still be necessary for computing to be widespread. CLIs aren't exactly useful for a number of modern applications.

You could have Apple and Microsoft work together on a GUI OS, using UNIX as a codebase.
Actually, the Original MacOS was based on the old BSD (Berkeley) kernel. Jobs used what he had seen from Xerox's PARC project to develop the GUI.

As far as UNIX GUI's, those have been around since the mid 1980's when MIT came up with X-Windows. And that has developed over time to what we see now for Linux distro's. The whole Gnome / KDE interface that people see as their GUI on their Ubunto/CentOS/etc... load is basically built on top of the X-Windows GUI.

As far as usefulness for modern applications, define 'modern applications'. For server level applications up until the mid 90's and still unto this day, you want to have your machine do the most work for the least amount of system resources. This is primarily why you'll never see Microsoft's name around most LARGE server farms as the primary OS (most that do have Microsoft OS's running end up running them as client-application OS's under something like VMWare which is of course fashioned around a specialized unix kernel which has much better memory management and is overall more flexible and less buggy). Having to use system resources to support a GUI were deemed more trouble than they were worth and if it's just to have a point-and-click idiotproof interface you have bigger problems (system managers should know enough how to set up and debug things at the CLI level as well as through a GUI). There are also tasks that are better performed more efficiently through CLI's than GUI's, notably batch commands which perform repetitive functions. UNIX implementations today are actually a mixture of the best of CLI and GUI interfaces (although I do kind of thing the newer MacOS is a bit daft in its implementation in still attempting to bury the CLI interface).

As far as Apple and Microsoft working together using UNIX as a codebase, that's a stretch. Given the past actions of both companies, they would have to have basically taken the UNIX kernel that was developed within education institutions and industry as their own legally and made it basically non-open source. Chances are we would have ended up with pretty much the same buggy OS's we have from Microsoft and Apple with no open source alternatives (due to the litigation that would have happened against folks like Linus Turvalds and others due to copyright infringement if Apple and MS were successful in securing the code for the Berkeley Kernel as their property).
 
Actually, the Original MacOS was based on the old BSD (Berkeley) kernel. Jobs used what he had seen from Xerox's PARC project to develop the GUI.

Speaking as someone with low-level programming experience in both UNIX and classic MacOS I find this very hard to believe. Even the Lisa was based around cooperative multitasking, for example, and was not Unix-derived as far as I know.

Your best bet would probably be Xenix. Actually, Microsoft agreeing to cooperate with IBM on OS/2 is what killed Xenix inside Microsoft; perhaps if they decide to compete with IBM earlier on instead of cooperating with them they focus on putting a usable user interface on top of Xenix and things go from there, with Xenix taking the place NT did OTL and gradually expanding to take over all of the mainstream consumer market as Windows 2000/XP did. Of course, you might well end up with the same situation as you have with NT now, at least for UI applications, where all programming is done to a Windows-style API layered over an opaque kernel.

Whatever happens, you need a way for Unix to get a better UI more quickly - Sun making NeWS open-source/an open standard might help, or free and widely available Motif. Athena widgets and tcl/tk weren't going to cut it by the time Windows became popular.

As for Apple/Microsoft taking over/using the BSD source - this is more or less what happened OTL in any case with Sun, IBM and so on (usually an admixture of System V and BSD code, actually), and there was a lawsuit against the early free BSD derivatives based on Unix copyright. Such a lawsuit would fail just as it did OTL, so FreeBSD would still exist, and Linux isn't based in any way on the BSD source code so would still be immune. You can't copyright interfaces, which is why the Wine project for example can exist.

There's no reason to think that things would be worse than the OTL 1990s, that I can see. No, Apple and Microsoft wouldn't be releasing open-source Unix-like OSes, but there's no reason they'd have to to be successful, and they would architecturally still be far more reliable than classic Macs or Windows 95. It's hard to address this TL without a reasonable understanding of the admittedly-convoluted history of OTL Unix before FreeBSD and Linux became popular, but a look at the evolution of say AIX, Digital Unix, HP/UX and so on should give you a good idea of both what sort of similarities and what sort of differences to expect.
 
Last edited:
Top