The following was written by Slashdot Reader knarf
Being a former journalist myself, I was rather disappointed at your recent column about the upcoming RedHat IPO. To be quite frank, there were more faulty facts in this short column than I thought possible. A quick summary:
1: Unix has been, and still is a money-maker for a lot of companies. Look aroud large datacenters, network hubs, computer centers ad financial institutions and you'll find a lot of big, expensive Unix-based systems doing all the heavy work.
2: Unix was not thrown in the `public domain' as you suggest. AT&T fought long and hard just to try to prevent others from making something which resembled Unix. The mere mentioning of the name `Unix' in connection to something not from AT&T was enough to be sued, hence the frequent use of words like `Un*x'. This also led to names like XENIX, AIX, SINIX, ULTRIX, DG/UX, etc. The rights to Unix have moved from company to company for a while, currently the Santa Cruz Operation (SCO) is the `owner' of the System V code. The Berkeley code was `cleaned' and released as BSD 4.3 lite (the `lite' stands for `without AT&T proprietary code'), and serves as the code base for some commercial and a lot of open Unix varieties. It is the Berkeley version which is in the public domain (under the BSD license), not AT&T's `original'. The BSD license has nothing to do with the GPL license, under which Linux is licensed.
3: The differences between all those Unix varieties have been greatly exaggerated, mostly due to marketing and media forces. Unix is not a perfect operating system, but at it's core it is rather clean and consistent. Any user of any Unix will have little trouble moving to another variety. That is not to say they'll like the experience, but THAT is something completely different. For some reason people tend to get religious about their choice of operating system, be it Unix version X of Windows version Y. For software developers, there are several ways the differences between the Unix versions can be solved. the GNU auto* tools are a prime example of this. Writing software for several varieties of Unix is no different from writing software for several varieties of Microsoft Windows, but at least the Unix developers have the benefit of the GNU tools and the often much better documented API's. In case of the free Unix versions, there ARE no `undocumented' API's since you have access to all source code.
4: Linux Torvalds did not create the `command set' for the kernel which got to be called `Linux'. He initiated development on the kernel, and has been managing the development effort ever since. What is generally called `Linux' is actually a distribution containing the Linux kernel combined with the GNU tools (which provide most basic user commands) and a lot of other software. From very early versions onward the X window system has been part of those distributions, making it possible to run GUI command interfaces on top of the kernel. Most early distributions contained the XView toolkit and the ol(v)wm window manager, giving Linux users access to the well-known `Open Look' style of user interface which has been available to Sun Microsystems users for a long time. Open Look eventually got replaced by CDE on Sun systems, while other toolkits (Motif, Qt, GTK) took over the Linux `desktop'. But early Linux users had access to a GUI interface as well.
5: Linux distributions all use the same kernel (give or take a patchlevel number, eg. 2.2.5 versus 2.2.9). They all use the same GNU tools (the `command set'). They all use the same X window environment. They all use the same basic toolkits and language interpreters (Perl, Tcl/Tk, etc.). The differences between distributions are mainly in the locations of these tools (`where are they installed'), and in the extra stuff they deliver (`what more comes with this thing?'. There are some cases where some distributors are quick in picking up on the latest developments, while others wait for stuff to calm down before moving to the `latest and greatest' version of some basic component. An example of this is the choice of C library (leading to incompatibilities between distributions) or the choice of loader format (a.out, ELF). There are all temporary problems though. The move from the a.out format to the ELF format is all but forgotten by most Linux users. The move to the GNU C libary will probably fare likewise.
6: There is nothing to be gained for Linux distributors in making their version of `Linux' incompatible with others. They rely on both open source developers as well as commercial vendors to supply them with software to run on their version of `Linux'. If they behave like you foretell, they'll loose support from the hard-core open source developers, and the commercial vendors will probably follow. Commercial vendors would rather support one or two basic varieties of an operating system, not a whole bunch of them. If a Linux distributor wants those applications to run on his distribution, he'll make sure it does. Now of course there is the question *which* distribution a commercial vendor will support. If they are smart, they'll support one or two (say RedHat and Caldera), but make sure their software runs on the `official' (not finished yet) `Linux Standard Base' or `File System Base' based distributions. That way, everybody wins, and they'll sell more software.
7: Sure, the Linux desktop user interface can be `improved'. But those improvements are probably not what you think they are. The command line will not disappear. It is a much to capable and valuable tool in the hands of even lesser Unix gods to be done away with. If RedHat is hiring people to make their Linux distribution more palpable to the Windows user community, that's fine. It will still be Linux. Until now, RedHat has licensed all their developments under the GPL (GNU Public License), and they are likely to continue doing this. If they don', they'll loose support from the open source developers (look at what happened to the KDE initiative, although that rift has been mended by making the Qt toolkit open source `compliant'). So they will most likely remain smart and keep the GPL flag flying.
8: RedHat is not losing $130.000.000, they lost $130.000 dollars. This is pure disinformation, and might be cause for RedHat to sue you. Please make sure you know what you write about before starting.
9: There is no such thing as `VA Linux'. VA Linux Systems is a hardware vendor in support of Linux. They support several distributions. They may have their preferences (on their website it says `...caldera is quite close to Redhat, but i've found I prefer RH's gui over caldera...'), but they point anyone interested to both RedHat as well as Caldera and SuSE and cdrom.com (Slackware, others) and linuxmall.com (all distributions, $1.89 per CD-ROM, also FreeBSD by the way).
10: I do not remember Microsoft being a proponent for open source software, not them being a beneficiary. On the contrary, Microsoft has from the very start supported a closed, `business-like' approach to software development. They have from time to time thrown some goodies at the developer crowd to get them aquainted with Microsoft tools, but that has nothing to do with `supporting volunteerism'.
11: Netscape did not create a `potential operating system', they created a browser with and API which was seen by Microsoft as a threat to their operating environments. The browser would commoditize Windows, since it would not matter anymore which platform was used to run the interface to whatever applications were used. Microsoft has done it's best to counter this move by embracing and extending Java and the `WWW protocol set', and Sun Microsystems has helped by being too tight-minded with regard to Java extensions. IBM has indeed helped the Java `platform', and is still doing so. They seem to be on a quest to keep Microsoft out of the higher-end application service market. Of course they are not benevolent gods, but they are much less malign than you portray them. Sun Microsystems HAS given access to the Java source code (throug htheir `community licensing scheme'), IBM HAS given and still gives a lot of software to the community (take a look at their AlphaWorks site, www.alphaworks.ibm.com).
12: Your comments about Microsoft flirting with Linux border the ridiculous. What do you think would happen if/when Microsoft embeds Linux in their products? What system would run Linux software better? A Microsoft operating system with embedded Linux, or Linux? You might want to compare this with running Windows software on Unix systems (which is possible by using products like WABI, or toolkits like WINE). Remember, when IBM called OS/2 `a better Windows than Windows'? What did you think when they said that? Did you try it? Now, Microsoft would come along and call their Windows 2001 `a better Linux than Linux'. What would you think when they say that? Would you try it? Even more important, what would you write about it?
13: My final comment, number thirteen. About your final comments. Why do you insist that people have to want to use `Microsoft Linux' to make it a viable proposition? I can very well do without Microsoft, and so can a lot of other people. This has nothing to do with ideology, but everything with stability, choice, and performance. I get more work done in less time using non-Microsoft products, and time (as you should know) is money, especially for the people who pay me to do my work. And why do you think the future and viability of Linux is dependent on the future of RedHat? If RedHat disappears of the face of the earth, there is still SusE (rather popular here in Europe), TurboLinux (rather popular in the Far East), Debian (rather popular amongst developers anywhere and everywhere), Slackware (with a dedicated crowd of followers), Caldera (rather popular in a lot of businesses), etc. Should I go on?
I won't. Linux has the marks of becoming an important player in a lot of markets, with or without RedHat. You may be right (or you may be wrong) when you say an investment in RedHat is not the best way to make money. But you are dead wrong in your reasoning against RedHat in particular or Linux in general. The IT market is not dependent on Microsoft, it can survive without them. So can Linux. Unix is not dead, and it does not seem to be dying either. There is a lot of money to be made in all thing Unix, and Linux is one of those things.
Linus did not start Linux to `kill Microsoft', and most Linux users use Linux because it fits their needs in a way that Microsoft Windows can not. If Microsoft embeds Linux within their own stuff, that's fine with me (as long as they abide to the licensing terms). Since Linux withoug any Microsoft-extensions fits my needs, I do not see what I would gain from such a development, so I will refrain from using `Microsoft Linux'. If they do produce some useful extension (and release it under the GPL) I will use it. If they embrace and extend Linux in the way they embraced and extended other `standards' I will not use their extensions, because I do not want to be tied to a single vendor. Such ties cost money, as experience has taught me. And I;ve got better things to do with my (or rather my employers) money than give it to some company because I am locked in their embrace.
My parting comment? Please do some more research the next time, or ask a knowledgeable person to proof-read your column. It may `only be a column', but a lot of people tend to trust the media a little too much for their own good. As you probably know...
With kind regards,
Frank de Lange
/* Former editor in chief, Unix Info magazine
* now full-time developer... mostly Unix, some Windows
* speaking for himself, not his employer