“Several readers took me to task for referring to Linux, BSD, and OS X as Unix. Lighten up, folks — I’m on your side. No one feels more protective of Unix’s heritage than I. Unix has a rich legacy that deserves to be preserved and accurately conveyed to new generations of computer scientists. It rattles many of us to see that the operating systems that best exemplify Unix traditions today aren’t Unix at all.” Read the editorial at InfoWorld by Tom Yager.
BSD is Unix. Its not even Unix-like. Its a direct descendant of System V, albiet with the code that got it into trouble with At&T stripped out. Its POSIX compliant, as all Unix and Unix-like systems are.
Linux is Unix-like, POSIX compliant. Its a clean reimplementation of the ideas found in Unix with none of its code.
Mac OS X is also Unix, derived from BSDLite 4.4. It is POSIX compliant. By the end of the year, as Mac OS X grows in acceptance, Apple will become the largest shipper of Unix systems in the world. Read that twice. Its true.
HP-UX, Solaris and AIX are also Unix. QNX is Unix-like and POSIX compliant. BeOS was partially POSIX complaint and still close enough that many command line applications will compile cleanly on BeOS.
BSD is UNIX but it’s descended from UNIX version 7, not System V. For that matter, System V is descended from version 7 as well.
Xenix was descended from System III and Linux is descended from Minix, Andrew Tannenbaum’s UNIX version 7-like operating system.
Indeed, good summary.
Just a note, BeOS while being partially POSIX compliant, it is NOT Unix, or even Unix-like. In fact, BeOS was written with the book of XINU next to the developers (“Xinu Is Not Unix”). Partial Posix support, was added many years later, around 1997 (Windows 2k/NT/XP/3k also has a POSIX certification IIRC, but this doesn’t make it a Unix .
As for Mac OS X, it was created from three parts: BSDLite 4.4, Mach and FreeBSD 3.x. These days, the main code additions it gets externally are from FreeBSD 4.x and 5.x. FreeBSD is now the main Unix influense on Mac OS X.
The way I see it… and a few of the people I know — no matter how Unix-like an OS is… it’s not Unix unless it’s UNIX. Surely it may have the same principles, “spirit”, and technical standards, but I’m much more appreciative of things like Linux as Linux, and BSD as BSD, and so on and so on. To me Linux is something grand, and the fact that it wasn’t derived from Unix directly makes me even more happy with it, why? Because it feels like I’m a part of something that’s been developed in my own generation. Linux itself has become sort of a movement alongside GNU and Open Source Software in General… and I’m proud to be a part of it and will forever know the communities style and ideals as being that of Linux. I’m proud to love Linux for Linux’s sake, not for Unix’s sake.
Disclamer: I’ve never tried Linux but I’ve read a lot of about it (mostly install horror stories!) I use Windows XP(tm) for kernel-level VB development and its mostly OK but I could use something better…
I think a combined effort needs to be made by AIX, Solaris, Linux and BSD companies to make Unix for the desktop. For this, I propose some changes:
– Mounting disks as directories? WTF? Drive letters are much better — A: is floppy, C: is the computer, D: is cdrom. What could be easier?
– Different layout for file system on disk. Windows has this right with My Computer showing drives, a printers folder, program files, etc. It even hides files directories I shouldn’t look in.
– File extensions for executables. For a n00b, it is much more intuitiev to have program.exe not /usr/xbin/program or whatever it is.
– Get rid of X11! Network transparency has been shown to slow down Linux more than 1000% (sorry I can’t find URL right now). Why should GUI buttons be pushed down a 1-byte-wide pipe when I have AGP etc?
– 100% windows compatibility. Wine is a poorly implemented hack. What needs to be done here is Sun’s expertise applied (they already have emulator experience with Java, remember?)
– Less open source. Programmers need to get paid! I think some critical apps (Apache, Samba, GNOME) have stagnated, because no-one wants to work for free. How about closing these apps and selling them? You pay $100 for redhat distro anyway…
> Just a note, BeOS while being partially POSIX compliant, it is NOT Unix
Also QNX.
QNX is IIRC fully POSIX compliant, but it’s not a Unix…
>Mounting disks as directories? WTF? Drive letters are much better
Sorry, but I disagree. Drive letters are archaic, and they are limiting to 26 letters (partitions). The BeOS way internally was to identify the disks as they appear in the ide chain, while for the user space, it just mounts the drives, as drives in the desktop, and gives them the name of the partition. For example, if you have a partition called “myWindows”, it will create a shortcut in the desktop, with a hard disk icon, as a directory, and that directory would be called mywindows. What could be easier? The terrible A, B, C and D, while the D can also be many things and not just the cdrom?
>Different layout for file system on disk. Windows has this right with My Computer showing drives
Windows has many things right in its (simple) layout, and a few things wrong. For example, Program Files is terrible and becomes bloated with time. Still though, better than the always bloated /usr/bin/ and/or /usr/local/bin/.
>File extensions for executables.
Sorry dude, but here, you are 100% wrong. MIME filetypes is the answer here, not extensions. Read here to understand the concept:
http://www.beosbible.com/excerpts.html (read both the FileTyping and the Query Power articles)
You seem to be stuck in the 1990’s.
>100% windows compatibility.
Yeah, but this is impossible to achieve. Not without the Windows source code next to the developers anyway. Therefore, what you ask here, is at the very best, impossible.
>Less open source.
I am pro OSS, and pro closed source. I like both. I believe each one has its place. Some times, we get more quality or powerful and innovative apps from closed source companies. In other cases, some open source apps are better than their equivelant closed source (e.g. Apache). Both have a place and just choose what’s best for you and your needs.
*LOL*
Not bad!
lets Mac OS X is Unix for the desktop. No letter drives. Drives automaticly mount and you can run Virtual PC.
Mac OS X is still not “there” 100% though. The filetyping problem exists on OSX (along with the jerkiness of scrolling and resizing windows
Read this great article for more:
http://www.arstechnica.com/reviews/01q3/metadata/metadata-1.html
Yes, the article makes valid points. I am the first to admit mac OS X flaws. I wish apple would work out MIME types and a more robust fs than HFS+
In casual conversation, I will use the term Unix to refer to any Unix-like system, including Linux and Minix. When speaking technically, however, I reserve the term Unix from those derived from the original UNIX source, though I’m fairly lenient on that. For me, the BSDs count as UNIX in addition to the various commerical UNIXs. Mac OS X is an interesting hybrid, with root in UNIX (BSD), and Unix-like (Mach) systems.
> I think some critical apps (Apache, Samba, GNOME) have >stagnated, because no-one wants to work for free. How >about closing these apps and selling them? You pay $100 >for redhat distro anyway…
Yeah Gnome development has definitely stagnated. It hasn’t had a new release in what…a day! Please. I agree there is a place for closed source software. But when Apache is kicking IIS’s arse, I’m glad there are open source developers out there. Sometimes pride is enough to fuel development. I think it means a lot to a development team to say, “Hey we’re creating a better webserver than any other multi-billion-dollar company. Let’s keep up the good work!”
i heard they were working on a new fs for the new os x 10.3. Apparently its supposed to be a lot better than the current hfs+. too bad i don’t have a link to an article… 😛
Heh heh. Eugenia, I think you fed a troll this time. I was a little suspicious when I heard the drive letter bit, but what tipped me off was the “kernel-level VB development.” C++ is a controversial language at the kernel level: using VB in the kernel would cause the breakdown of world order.
Actually, I suppose the post could be interpreted satire, in which case it is rather well done, if lacking in subtlety.
Well done!
You even got Eugenia off-guard!
hehehe! 🙂
Eugenia, watch out! 🙂
dang I knew something was wrong why’s my cd rom L:?
Dually Multiple SCSI HD”S and IDE Hd’s
Least SuSE just showed it as CDrom and the HD’s as data 1,2,3,…
LoL – I was about to say the same thing ! I remember some troll a while back used that “kernel-level VB development” line before in COLA ( Linux advocacy newsgroup used to keep trolls and flamers from migrating to the real Linux newsgroups designed for trouble-shooting, and helping people etc…) and everyone there pratically laughed their ass off.
– Less open source. Programmers need to get paid! I think some critical apps (Apache, Samba, GNOME) have stagnated, because no-one wants to work for free. How about closing these apps and selling them? You pay $100 for redhat distro anyway…
Well since closed source programmer get paid, why can’t any of them just write another apps that supposed to be better than the above. I think this is a better way rather than trolling everywhere.
MIME is not better than file extensions. it is possible to edit MIME info and associate jpeg with wrong application (e.g.realplayer). If rem format is not handled by audio application installed on the system then it will be not possible to open it.
On the other hand if one copy doc file on the machine that never had word or other MS product installed but it has WP insted then double clicking on doc will open it in Wordperfect if there is no other app that will claim doc file as self.
It is possible to associate systemwide extensions with application. MIME does not impose what app will open specific file, but that must be pre-defined otherwise if one has installed IE, Netscape, Phoenix (or Firebird) then what app will be launched if doubleclicking on the file that is html type? If htlm was associated with IE and one de-install IE then html will be open with next incharge, probably netscape.
MIME idea seems to be also from 90′(Copyright © 1999 Peachpit Press and the respective authors.)
I personally prefer the native UNIX way, namely magic file headers (to see for yourself on a UNIX-like system, type “file *” in any directory)
yup, but it is not ideal too. For example sxw files can be recognized as zip archive. (But this is not the only reason of liking it)
It would be nice to make more use of them, though; for instance, they recognize other attributes (date, block size, special properties) that MIME and extensions cannot.
LOL!! Nice troll…
For those who bited: “Kernel-level Visual Basic development!” was supposed to be the key word. RTFLOL!!!!!
Aaaahhh, thanks that cheerede me up. Good one though….
if vb kernel development is a bad idea… what about running a kernel in a .net runtime running in a c-64 emulator :oD
with an abacus on an expansion card in an isa slot for a “floating point unit”
I take it then that this abacus expansion card I just bought was a rip off?
@that whole thread about the VB winkernel hacker: Sometimes I feel like I’m at the zoo, watching people feed the trolls from a big cereal box.
MIME is not better than file extensions. it is possible to edit MIME info and associate jpeg with wrong application (e.g.realplayer). If rem format is not handled by audio application installed on the system then it will be not possible to open it.
You can also rename a file and change its extension from .jpg to .ra and you’ll have the same problems.
It is possible to associate systemwide extensions with application. MIME does not impose what app will open specific file, but that must be pre-defined otherwise if one has installed IE, Netscape, Phoenix (or Firebird) then what app will be launched if doubleclicking on the file that is html type?
You misunderstand what MIME types are for. MIME types alone only solve half of a problem; how do you reliably identify the type of any given stream of bytes? All a MIME type does is to codify the file type for future reference.
How the system decides which applications handle which MIME types is another matter, and is upto the system. The “best” way to do it is to have applications register the MIME types that they are able to handle, and then have a set of default applications configured system wide. When a user double-clicks on a file, you look up its MIME type, look in the system database to find which application can handle that MIME type, and you launch the application.
Believe it or not it is not much different to how Windows does it; the only difference is that you use MIME types to encode the file type whereas Windows uses the filename extension.
Yeah, conversations about meta data concepts are tough to fit into this little edit box
After all, its not like companies such as RedHat, HP, IBM and others *ever* hire open source developers. Oh wait, they do….dude, just because you work on open source software doesn’t mean your begging in the streets. A good majority of the major open source developers are paid for what they do.
That would be a very stupid idea to change the Mac file system in OS X, so that we could have problems not just with the 90% of the world that uses Windows on the desktop, but we would have trouble communicating with the other 75% of Mac users. Mac OS X was not designed to run as a large server anyway. Who would buy it when they could use Linux instead?
What we really need is native burning that can be done in OS X in the ISO 9660 level 2 or Joliet format so Windows users below XP and Linux users without hfs tools can view our CD-ROMs.
Rubbish.
It’s certainly possible to be paid to work on Open Source development, and yes, companies like RedHat are a good example of employers that hire developers to do this.
That said, it’s still a drop in the ocean compared to all the other development going on. The vast majority of Open Source development is volunteer-driven. Not that it matters where it comes from, as long as the code is good.
You have the profile of a windows user, so why don’t you stick to windows? It’ll suit every single one of your needs.
You don’t understand what UNIX-like OS’s are all about. Although I must admit some of the linux-guys are starting to forget it too…
Listen carefully, and remember this: I _DO_NOT_WANT_ my OS to look and behave like Windows. If I would want that, I would be using Windows, right?
I find it easier to have my cdrom in /cdrom, and my writer in /writer instead of trying to remember which one was D: and which was E:, and having windows swap them around from time to time, without any reason.
Why should UNIX have a different file layout? The current one is very logical. Why should it change? Because _YOU_ don’t like it? Give me a break.
File extensions? You can have file extensions if you want to, most people use them for common filetypes. Executables don’t have an extension, in general, indeed. But have you ever noticed that they have the +x bit set?
And there’s still a fine utility called file.
X11 is nice. Network transparency is something I use almost every day. There may be some performance issues, but then again, my X11 runs at user level (regular user, not root user) instead of kernel-level on windows. Think about the security implications of it! Besides, when I use a fast WM like eg. blackbox, I can easily beat win xp when it comes to desktop responsiveness.
100% windows compatibility? Why? I don’t need to run any windows apps. If you need to run windows apps… *tada* run windows!
And your final point is even more funny… Open Source had proven itself to create lots of quality software.
The reason why is easy. Picture 2 programmers: one is an opensource guy/gal, the other one works 9-to-5 at Microsoft.
The opensource guy/gal is sitting behind his/her computer creating a program he/she finds interesting, in total freedom. He/she can create it as he/she likes.
Now, the M$ guy is sitting at his computer coding as fast as he can, to be able to leave at 5pm. He doesn’t really care about the quality of his work, because he’ll get paid either way. And even if he has been enthousastic in the past, he still just has to do what the boss says, even if that forces him to create totally broken designs, while knowing it.
And Opensource guys DO get paid. I like the software I use (which happens to be OpenBSD), so I buy the cd’s instead of just downloading it. I pay them because I like what they do, not because I have to.
OTOH, I would never pay M$ anything, because I find their products not worth their price, and I totally disagree with their way of doing business. But then again, I don’t use any M$ products, and never will
In short, there is no “one and only userfriendlyness”.
I find the M$ Windows user interface very non-userfriendly, because it keeps trying to hide from the user what’s going on; try solving problems without decent error messages for instance.
Configuration is all over the place, and for some reason, some configuration options tend to change without any reason. On UNIX, config files are human readable text files. That’s pretty darn userfriendly if you ask me.
Also, windows doesn’t allow you to do complex tasks efficiently. There’s no (easy) way to chain programs together and make hard things easy, like you could do with a simple shell script on UNIX.
Others (most people, indeed) find the windows-way of doing things easier. So their definition of user-friendlyness is clearly very different than mine.
So let’s just keep the diversity we have now, so that everyone can use the system he/she likes best. If you’re happy with windows, there’s no need to be whining that other systems should be more like it, because they really shouldn’t.
I never have sed that one is better than other. I’ ve sed that there is not much defference. In contrast to Eugenia’s statement.
You only proved my intentions Vanders
re:Syntaxis
I don’t believe I said the majority are paid, merely that the majory players (like the core kernel developers) usually are.
~Christopher
As I understood it, your argument was that MIME types were useless because they couldn’t do what file extensions do. Which is not the case.
For example you argued that a user could change the MIME type, breaking the file association, but at the same time your argument implied that a system which uses filename extensions would not have this problem, and that is certainly not true.
My point is that MIME types can do everything filename extension matching can do. They can also do a heck of a lot more that filename extension matching cannot do, but you do not seem to see this.
I still couldn’t understand why there are people that still complaining about getting paid when talking about open sources. There are a lot of cases where OSS programmer got a very good income.
BUT the most important is that, those who expect to get payment for their work, the answer is simple enough…. GO AND WORKS ON MANY OTHER CLOSED SOURCE APPS, UNDER CLOSE SOURCE OS. There are a lot of opportunity available to works for in Windows, Mac, QNX and many other proprietary platform. Why keep on nagging about the nature of OSS.
No,
I provided this arguments to show that you can do the same things with MIME (as with extensions). Everybody knows that it is possible to wrongly associate apps and extensions. I wanted to say that you can do that with MIME to (that concerns other problems too). So basically there is no advantage in using MIME over extensions and there is no advantage of extensions over MIME. There are simply two different approaches. Neither is better. It is possible to do more with MIME as with metadata.
if you have wasted time to argue with a ridiculous proposal like Generation Movement, it is worthless. Time is gold, common Lady and Gentlemen.
Almost of you are *nix fans, so you have a philosophy of computing technology, and you would dedicate your work as the way you choose. Do not mention to those who “never tried Linux” and know nothing but windows.
cheers and regards
dieu
Suffices are slightly more limited than MIME typing in its most simple implementation. A program may go by a file type in its requesters. A good example is when I save a JPEG file in M$ Internet Explorer. In the file requester, only files which seem to be JPEG files are displayed. Files which end in .JPG, that is. Which is downright silly, since the proper extension would be .JPEG. This shouldn’t be the case with a system, even a minimally implemented one, which works with MIME types, since MIME typing would be an OS service, and the programs wouldn’t specify a particular file extension, but only the desired MIME types. How the OS would identify the proper file types is another matter, but at least the program would have no say in it. File extensions become a bit limited (not to mention that it’s an overall inelegant solution) when there are several alternatives, some of which the programmers may not have thought of.
“Now, the M$ guy is sitting at his computer coding as fast as he can, to be able to leave at 5pm. He doesn’t really care about the quality of his work, because he’ll get paid either way.”
If you reproduce bad code, MS fires you rather quick. If you try to check-in code which reduces performance… it’s not possible (only for some high-ups), if you break compilation a second time you’re fired, they do regularly code audits, etc. They actually care for code quality.
“And your final point is even more funny… Open Source had proven itself to create lots of quality software. ”
As has propritery software. Actually I can find more high-quality software with closed source than with open source. But that boils down to your personal definition of high-quality.
I think it’s not valid that open source is better then closed source when it comes to code quality and security. You find good and bad in both paradigms and I can not see why one should be inherently better than the other for quality.
That Unix on the desktop thread is funny
I’ll bite, and feed the troll: I get payed for open source (well, php development. Not a lot, but it’s damn fun.
The irony is that I develop all of my opensource (php) code … using closed-source tools (Windows XP, HomeSite 4.5.1).
> If you reproduce bad code, MS fires you rather quick. If you try to
>check-in code which reduces performance… it’s not possible (only for
>some high-ups), if you break compilation a second time you’re fired,
>they do regularly code audits, etc. They actually care for code quality.
Good code, maybe. Secure code, hardly.
Hey, it compiles! I don’t get fired.
Hey, it exposes personal information of 200 million Passport users! I must be in line for a promotion!
> “And your final point is even more funny… Open Source had proven
> itself to create lots of quality software. ”
> As has propritery software. Actually I can find more high-quality
> software with closed source than with open source. But that boils
> down to your personal definition of high-quality.
Something that doesn’t allow a gaping security hole exposing 200 million user accounts for several months. OSS can identify, patch a hole, and release code in 24 hours. History shows us this.
What’s Microsoft’s excuse?
> What we really need is native burning that can be done in OS X in the
> ISO 9660 level 2 or Joliet format so Windows users below XP and Linux
> users without hfs tools can view our CD-ROMs.
It already does this. Try using the operating system for a change. It also supports ISO 9660 level 1 to be COMPLETELY backward compatible down to 8.3 file naming (lowest common denominator).
Just because it isn’t in the GUI does not mean it isn’t available. It’s like saying technology that isn’t in Gnome isn’t in Linux.
As I tell every MacOS X user: “Look beyond the icons!” ((C) 2003, Juan Carlos de Burbon.)
No one can figure out if it’s dead or alive, or where it is.
Let’s get to the bottom of this question, since I’ve seen it confuse so many people over the years, that it’s become quite tiresome.
The real question is, “What is UNIX?”
People talk about UNIX a lot. But it’s obvious from listening to them, if you pay attention over time, that they’re not all meaning the same thing when they invoke that name. One of the biggest things that’s hurting the free software community right now is that we’re suffering from loads and loads of hype, but nobody knows what any of the words actually mean, and there’s no agreement. It creates the impression that none of us know what we’re talking about, and I’m not entirely sure that impression is wrong.
So, what is UNIX? It’s actually quite understandable that so few people can actually say, since unfortunately, the answer depends on the context of the question. The English language is full of many such words and expressions, which is why it’s so hard for many people around the world to make sense out of English. Computer terms are supposed to be more precise though. The term ‘UNIX’ is unfortunately not one of those precise computer terms. It’s pretty vague.
Here’s a list of the various contexts you may see the term used in, and the appropriate meanings for the contexts:
IF YOU ARE A PATENTS/TRADEMARKS LAWYER — then the term ‘UNIX'(R) (notice the registered trademark symbol after the word ‘UNIX’ in all capital letters) is a trademarked name of the Open Group. It first belonged to AT&T/Bell Labs, then it was sold to all sorts of other parties who hot-potatoed it around for many years, past owners including even Novell. The Open Group are the current owners. They currently publish what they call the Single UNIX Specification, which is basically a description of what they will bless as “real UNIX.” In the past however, being considered “UNIX” by the Open Group has historically had less to do with the specifications of an actual operating system’s software than with just getting the Open Group to like you. Also, there’s a fee involved, paid to the Open Group. Basically, in this context, UNIX is a name, with a registered trademark symbol after it (the (R) in a circle), which comes from the Open Group, and you can use it if they say you can. Sun, SCO, HP, IBM, Apple (MacOS X/Darwin), and DEC have all had this blessing of permission. Linux never has, though whether that makes it less ‘UNIX’ in any real sense is something we’ll get to later….
IF YOU’RE A SOFTWARE HISTORIAN — then ‘UNIX’ is something that hardly anyone uses today, either commercially or privately. That’s because in the strictest, most historical sense, ‘UNIX’ is the operating system that was published by AT&T during the 70’s and 80’s, known to most in its most recent form as AT&T System V Release 4 UNIX, or just SVR4 for short. While many modern ‘UNIX’ operating systems contain gobs and gobs of licensed, fully legal and authorized code taken directly from the codebase of this ancestral AT&T UNIX — including Solaris, Tru64, SCO, and others — none of them are actually AT&T UNIX per se in its most vanilla form. The last time I saw a machine running vanilla unmodified AT&T UNIX was in 1991. In the strictest, most historical sense, only this original UNIX is ‘UNIX’, and everything afterward is a derivitive or variant. That means that in this context, virtually nobody runs UNIX.
IF YOU ARE A TECHNICAL SOFTWARE-ORIENTED PERSON — then to you, probably the term ‘UNIX’ means a family of at least thirty or forty different very similar operating systems made by various groups and companies. They’re all almost completely abstracted from the hardware they run on to the point where everything from the kernel to the apps can be recompiled in C on another platform. They all are based on the philosophies of POSIX as far as things like syntax and filesystem design goes. They all are multiuser, use tty’s to associate terminal sessions with running programs, mostly use the X Window System for a GUI when a GUI is desired (though there have been exceptions to that), and tend to treat everything as a flat text file nomatter what kind of data it really is so that programs that are good at working with text files can manipulate anything from actual text to MPEG videos. Common programs you expect to find on such a system include awk, vi, perl, sh, grep, and the whole bunch. Systems that fit this description include everything from the original AT&T UNIX to BSD, SunOS, Solaris, HP-UX, AIX, IRIX, Tru64, Ultrix, SCO, and God only knows how many dozens of others. By this common definition, even GNU/Linux, the operating system spearheaded by Richard Stallman, whose project literally is an acronym that stands for “GNU’s Not Unix”, *is* very much UNIX.
This is the common definition that gets thrown around the most, because let’s face it, most of us are software-users, not patent lawyers or historians. It if looks like a duck, quacks like a duck, smells like a duck, then it’s a duck. We don’t care about patents and trademarks on names owned by the Open Group, and we don’t care about AT&T’s old original stuck in the 70’s UNIX that nobody runs anymore. As far as we’re concerned, they’re *all* UNIX. Solaris is UNIX. BSD is UNIX. Linux is UNIX. HP-UX is UNIX. IRIX is UNIX. SCO is UNIX. Tru64 is UNIX. THEY’RE ALL UNIX.
But even though this is the definition that makes the most sense for the most people, as soon as you start using it, *somebody* from one of the other groups — someone who’s more concerned with names or history than with function — will spring from the shadows and start talking about “UNIX and BSD” (implying that BSD isn’t UNIX), or “UNIX and Linux” (implying that Linux isn’t UNIX), or even say that Linux is better than UNIX (which badly confuses people who consider that to be saying that Linux is better than itself…say what???).
<continued, because of silly message length limits…>
Even worse, this perception seems to have generated a whole new generation of Linux users, not aware of their platform’s heritage, whose belief that they are running something other than UNIX has led them to speak of “Linux apps,” as though Linux had applications of its own. It is true that there are certain programs that are either so badly written that they will not recompile to run native on Solaris or HP-UX or other UNIX systems without nearly resorting to nearly rewriting the program from scratch (fortunately not too many programs like that), and it’s also true that Linux has certain features like ALSA sound which can only be taken advantage of on Linux, which tend to make an application very Linux-centric if it uses them. By and large though, 90% or more of what are glibly refered to by naive Linux users as “Linux apps” are simply UNIX apps which happen to be compiled on Linux to run on Linux. These days, thanks to the magic of GNU autoconfigure and automake scripts, a C program can hop from platform to platform with an ease that our ancestors back in the 80’s trying to get programs to port between then much more divergent UNIX systems would have thought highly unlikely. Consequently, the perception has been created for many newbie Linux users today that they are on a platform that is not UNIX, but has something vaguely to do with UNIX, though they don’t know what exactly. They’d be very surprised to learn that nearly *all* their programs, including ones which came from the Linux/GNU developer community, are UNIX apps in every sense.
So basically, we’ve got a real terminology problem here, and it makes us look bad. We throw around terms like UNIX, and we can’t agree on what they mean. Add to that the ongoing debate about how Linux is pronounced (which still comes up occasionally), the distribution wars, and various little trappings of smirky hacker culture which I enjoy but many corporate people frown on such as names like ‘Slackware’ or the cute penguin Linux logo, and the sum result is you get a world that other people look at and wonder which asylum we all escaped from. It makes it very hard to tell someone in all seriousness that this is the most technically superior software available today. Usually, the person who is questioning you about this is not exactly the type of person who’s familiar with C compilers.
Let’s just all agree that all operating systems that functionally fulfill the expectations of what’s come to be known as a UNIX system *are* UNIX. Let us leave all the fine hair-splitting to the lawyers and the historians.
-Brent