Two entirely unrelated things:
I had switched from using Gmail for my domain to an account with Tuffmail but I may now consider moving back to Gmail. Tuffmail offers a great service with many cool email features, but it is nice being able to fall back to the Gmail web interface if necessary.
Things have been really, really busy both at work and at home lately. The good news is that I think we are seeing the light at the end of the tunnel, although you never can say for sure. :-)
Anyway, I did have a few miscelleanous items I wanted to post about. First, I am continuing to test out various BSD operating systems and am still thoroughly enjoying them, especially FreeBSD and OpenBSD. They have their quirks, to be sure, but so do other operating systems. I am debating whether to try using one of the BSDs as my primary desktop operaring system, at least for testing purposes. A few people have asked me whether I think BSD is better than Linux or vice-versa, but I don't really see it that way. No operating system is perfect, and no operating system is the best choice in all situations. Linux is still my primary operating system at home, and probably always will be. It's just that I have found that I _love_ UNIX-like operating systems so much that it's fun to see other iterations and test them out. FreeBSD and OpenBSD are great in their own right, just as many Linux distributions are. The best thing about all of this is that we have _choice_ and that's really what it's all about.
I have also been enjoying using the TTM Tint Task Manager which is a very small piece of code hosted over at Google. I believe an ArchLinux user created this (at least I found out about it in the Arch forums) but it's a really, really cool little tabbed transparent taskbar. It's very configurable, but the way I have it set up is that a small transparent tab will appear in the bottom center of my desktop as I open an application window. As more windows are opened, more tabs appear, and they disappear as I close windows. When I have no windows open at all, there are no tabs. So it's very "dynamic" as opposed to a static taskbar that is always there. Over the past several years, I have usually used pypanel, a nifty little transparent taskbar written in Python, but ttm might have replaced it for good. I've been using ttm for several months now and have really taken to it.
Lastly, it's great to see that so many Linux Reality listeners are going to the Ohio Linux Fest. I have made my reservations and am really looking forward to attending. Pat from The Linux Link Tech Show posted in the LR forums that he was talking to the OLF organizers about combining forces for the Friday night pre-event get together. I look forward to meeting the TLLTS crew, Dave Yates of the LottaLinuxLinks podcast, and all the great LR listeners who are going. Columbus, here we come! :-)
I have previously blogged about the interesting services provided by rsync.net. Now, it appears they are coordinating some bounties for FreeBSD work, most notably for a port of VMWare 5.5 on FreeBSD 6.x. The VMWare bounty is now over $1800! Hopefully, someone will step up and make this happen. Having VMWare running on FreeBSD would be really great.
Over the past few weeks, I have been playing around with FreeBSD and OpenBSD in various configurations. I currently have a file and print server set up with FreeBSD on a Pentium III and a wireless access point/OpenVPN server set up with OpenBSD on a fanless VIA box. Those two boxes are working great.
I have also been testing out both FreeBSD 6.2 and OpenBSD 4.0 (mostly FreeBSD) on the desktop -- on my Dell Optiplex GX620 and my Thinkpad t42 -- with some mixed results. Granted, I have not spent years with these operating systems and I do not know them as well as I know Linux. Nevertheless, I have run into a few issues on both machines, and I have spent countless hours Googling, searching two different BSD forums, reading the man pages and other written documentation, trying to find fixes for these things without much success. While none of these things are really major deal-breakers, they are a tad annoying.
Now, for the things I like about the BSD's, and why I may just keep one or two of them around on the desktop somewhere here in my house:
I really like both FreeBSD and OpenBSD and I will continue to work with them and learn more about them. I am quite sure that some of the issues I identified above can and will be fixed in future versions of these operating systems. I would definitely encourage any Linux user interested in learning how other Unix-like operating systems work to try any of the BSDs.
I have been pondering some difficult issues over the past few weeks.
There have been many debates over the issue of whether to use or include proprietary kernel modules, drivers, or firmware in historically free and open source operating systems such as Linux distributions or the BSDs. On the one hand, the argument seems to be that the more that can be done to make user transitions easier, by using or including drivers for proprietary graphics cards, wireless chipsets, multimedia codecs, web browser plugins, and the like, the better off we will all be since an increased *nix userbase will lead to more software, more hardware, more everything from independent hardware and software vendors. I myself have said the very same thing on my Linux Reality podcast. Several efforts here and there have come to light more or less trying to advance the cause using this pragmatic approach. One such recent example is that of the Linux kernel developers agreeing to provide free developer time to help companies create Linux drivers. The kernels devs said they would, if necessary, sign NDA's, although the resulting code or driver would still be GPL'd as stated in the FAQ to the original offer. Seems reasonable.
On the other hand, the position espoused by people like Richard Stallman, OpenBSD founder Theo de Raadt, and others would argue (I think) that agreeing to use proprietary "stuff" be it hardware or software locks you out of your freedom. Who is to say what is in that binary blob serving as a video driver or wireless firmware, especially those things that get inserted into the kernel? To counteract these blobs, many groups, including the xorg developers, the OpenBSD developers, and others have been working hard to create free, reverse-engineered drivers for various *nix platforms. I recently installed OpenBSD on a Thinkpad t42 and was pleasantly surprised that the Atheros AR5212-based IBM mini pci card in the laptop was fully supported in OpenBSD using the competely free ath(4) driver. Linux does not have a totally free driver for this card -- one must use the MadWifi driver. The problems with this driver are explained pretty well in the Thinkwiki site:
The "official" driver consists of a BSD/GPL wrapper with an unmodifiable HAL (Hardware Abstraction Layer). This HAL is not binary firmware as with the Intel wireless chips, but a piece of code that needs to run in the Linux kernel. It consists of header files for which no permisison to modify is granted, and pre-compiled object files. The vendor's position is that the Linux community simply needs to accept this sourceless HAL, since in principle the Atheros chip could be tuned to any frequency, and thus produce RF interference with systems operating in those frequencies.
This binary HAL is unacceptable to the Linux kernel developers, and the Atheros driver in this state will never become part of the official kernel. Some OpenBSD developers, facing the same issue, reverse-engineered the binary HAL and have produced a freely-licensed HAL for Atheros chipsets. Hopefully a driver based on this free HAL will be included with the Linux kernel at some point in time, and picked up by the mainstream distributions.
Thus, there is a fundamental difference in how these types of hardware restrictions are handled in various *nix operating systems. Linux distributions use the precompiled, binary code and the OpenBSD developers reverse engineer the chip to make it totally free. Interesting.
Anyway, the recent offer by Linux kernel developers caught the eye of many, including some on the OpenBSD mailing list. Theo de Raadt, specifically, had this to say:
You are trying to make sure that maintainers of code -- ie. any random joe who wants to improve the code in the future -- has LESS ACCESS to docs later on because someone signed an NDA to write it in the first place.
You are making a very big mistake.
I have been fighting against this for nearly a decade.
I don't know what you signed an NDA for in the past, but I can ASSURE YOU THAT OTHER PEOPLE WITHOUT THE DOCS CANNOT FIX THE BUGS IN THOSE DRIVERS SO THEY SUCK MORE THAN DRIVERS WRITTEN WITH FULLY OPEN DOCS.
PERIOD.
You cannot FAQ that problem out of existance.
It is people like you who are closed.
Tough words, indeed. And these are tough issues. I have heard some people complain that the next release of Ubuntu is not going to include proprietary drivers and codecs, that only by including such closed source or proprietary code will Ubuntu compete with Vista and OS X. Others complain that even making access to such drivers and codecs available is problematic and comes at the cost of promoting free alternatives, such as the excellent Ogg Vorbis audio codec. The former group might argue that we need to get more people using *nix, that if we restrict or prevent the use of such closed source stuff, then the *nix desktop will not "work" per the expectations of the legions of Windows users who have all this stuff working all the time. The latter might argue that *nix users need to take a stand, to only use completely free software and hardware drivers, and to pursue the goal of keeping our freedoms alive by forcing companies to provide documentation, by not using closed code to view web sites or operate graphics cards, and by supporting free standards, even if it comes at the cost of short term desktop usability and increased userbase.
Is there any way to have it both ways? Probably not.
Although I have historically been in the camp of the first group -- the group that uses some of this proprietary hardware and software -- really thinking about the arguments made by the second group has given me pause. Seeing what the OpenBSD developers have done to reverse engineer stuff I can only imagine what would happen if more people like them joined the cause.
I decided to get a hardware RAID card for a new server I am putting together on a P3 750mhz, and went with the 3ware 8006-2LP PCI SATA Raid Controller which you can read about on Newegg. This card is really great -- you just stick it in a free slot and boot the box. There is a BIOS on the card so when you first boot the machine, you go into the card's BIOS and select your settings. I went with good old Raid 1 mirroring. After a reboot, you are done. The OS sees the two drives as just a single drive and your Raid is all set up. Plus, 3ware has great Linux support -- the drivers for this card are already part of the Linux kernel.
Very slick.
So, anyway, I decided to put Debian Stable on this server since I am pretty used to running Debian on my servers and there have been no issues. However, I have recently been spending some time with OpenBSD and FreeBSD, trying to learn some of the nuances of those operating systems and how they differ from Linux. I really like how clean these OS's are -- I was shocked at how elegant both /etc directories were in the two BSD's as compared to some Linux distros really clutter up the /etc directory. Since I have not spent much time migrating stuff over to this new server, I am sorely tempted to replace Debian with one of the BSD's and see how it goes. I really like how secure OpenBSD is, and running so many services in a chroot seems like a good idea.
There are so many great options to consider. Isn't that what the free and open source community is all about? Choice? Damn, I love it.
| : foot.html,v 1.2 2008/10/17 13:43:05 chess Exp $ |
|