Bringing you the latest news from the Linux World.
Dedicated to keeping Linux users up-to-date, with concise
news for all interests
Sections: Main page Security Kernel Distributions Development Commerce Linux in the news Announcements Back page All in one big page
Other LWN stuff:
Archives/search
Recent features: Here is the permanent site for this page. See also: last week's LWN.
|
Leading items and editorialsThe other free software war. There are two competing free software projects out there, both working on solutions to the same problem. There are strong differences of opinion on who has the best technology; a fair amount of sniping goes back and forth. Corporate interests are increasingly visible, and accusations of selling out are occasionally heard. There is also occasional talk of bringing the two projects together, but it doesn't look likely to ever happen. No, the issue is not desktops - it's real time Linux. And while the term "war" overstates the case (just as it does with desktops), "intense competition" certainly does not. And the stakes are high: Linux is clearly poised to take over a large part of the embedded market, and many embedded applications require real time capabilities. There are numerous contenders in the real time arena, but two projects dominate:
Both projects are based on the same idea: a module is loaded into the Linux kernel which pushes Linux to the side and runs it as the lowest-priority process. Thus the full set of Linux capabilities is there to use whenever no hard real-time task needs the processor. Both require that hard real-time code be written as kernel modules, and both provide FIFO-oriented communication mechanisms to move data between Linux and real-time tasks. So what are the differences? RTLinux takes a very spartan approach to features; reliability, absolutely deterministic response, and portability to many platforms are the project's priorities. For example, no dynamic memory allocation is possible in RTLinux. RTAI aims to provide a broader set of features to the real-time programmer, but only runs on Intel systems. RTLinux makes many changes to the standard kernel source, while RTAI takes a minimalist approach to kernel changes. That small set of differences has become a rather wide divide; the two camps often seem to not like each other very much. The situation is certainly not helped by the fact that both projects are headed up by, shall we say, highly determined men with strong opinions. Also unhelpful, in this regard, is that Victor Yodaiken owns a software patent on the real-time Linux concept. While Mr. Yodaiken has never attempted to exercise this patent against RTAI or any company using it, he also has not plainly said that such action will not happen in the future. As the stakes get higher, expect the rivalry between these two projects to become louder and more public. Consider, for example, the case of MontaVista Software, which announced the availability of a prototype, fully-preemptable Linux kernel with a claim that it was the first hard real-time Linux system available. The announcement drew a gentle correction from Mr. Yodaiken (MontaVista is a supporter of RTLinux), and a much stronger one from Lineo, an RTAI supporter (via Zentropix, which it acquired). The desire to be the dominant real-time Linux solution will certainly drive more of this sort of claims and counter-claims. Of course, the quest to be the real-time Linux solution is likely to be as futile as that to be the desktop solution. The same dynamics apply: the competition has caused both groups to push harder to release a quality product while providing choices to real-time programmers. This kind of diversity is a good thing. All we have to do is to keep it friendly and polite... More on the :CueCat. Last week's LWN discussed the sad story of the :CueCat scanner and the Linux driver that the :CueCat's backer, Digital Convergence, seeks to suppress. That issue lacks any sort of resolution at this point, but there are a few followup items that are worth mentioning:
For those who are having trouble with the :CueCat's sophisticated "XOR with 'C'" encryption scheme, here's a bit of Perl code (from the "Dissecting" page) that does the trick; it's attributed to Larry Wall: #!/usr/bin/perl -n printf "Serial: %s Type: %s Code: %s\n", map { tr/a-zA-Z0-9+-/ -_/; $_ = unpack 'u', chr(32 + length()*3/4) . $_; s/\0+$//; $_ ^= "C" x length; } /\.([^.]+)/g; In the end, the :CueCat is a classic example of a broken business model. One can no longer (if, indeed, one ever could) put out a piece of hardware with the expectation that people will not seek to improve its operation with new software. And, in the free software world, only one person need code up improvements before everybody can enjoy the fruit of that labor. Digital Convergence was very generous to give away all of those barcode scanners; but to expect them to be used only with their software was naive in the extreme. LWN readers have spoken. We had guessed that the popup windows with the software announcements would be unpopular. In fact, not a single reader wrote in favor of them. So they are gone, forevermore. We still don't have the announcements sorted by license, but that will change next week... we promise... Inside this week's Linux Weekly News:
This Week's LWN was brought to you by:
|
September 14, 2000
|