I am running under FreeBSD 6.0 right now and just installed minicom, a terminal program. I have it talking to one of my SSC-32’s now (/dev/ttyd0 set at 115200, 8N1). 
I can also talk to the SSC-32 using Python (also available for Windows), which requires the py-serial module found at Source Forge. 
Now I want to see if I can get FreeBSD to work with my LinkSys USB Bluetooth adapter.
8-Dale
Free operating systems…coool
Congrats. Besides some basic kernel stuff, and a few program quirks, the BSDs are exactly the same as Linux, at least IMHO. I use OpenBSD on some of my machines due to the excellent track record on security, though on them I basically turn off all the functions except SSH and/or NAT and logging. Mileage would probably vary if used as a desktop. However, I know KDE can run easily on BSD, as it was developed for all the unices. We have some Solaris boxes in the astronomy department here that run KDE and work exactly the same as our Linux boxes (with a few minor issues like no flash support) due to being non-x86. Anyway… I’d like to know why you chose FreeBSD?
IMO, FreeBSD is the most user friendly of the BSDs for desktop use, and it’s easier to get sound and such working on it. All my server stuff runs under FreeBSD also - DNS for one of my domains, apache2 with PHP5, mySQL, mail, etc.
If I can get FreeBSD to recognize and use my Bluetooth adapter, I can see possibilities for a UNIX/Linux controlled robot without an onboard microcontroller. It’s something I want to experiment with at some point, perhaps with W.A.L.T.E.R. or my upcoming new BRAT.
I prefer Linux for a user box though, but have been having some difficulty getting Gentoo 2006.1 installed.
8-Dale
I can respect that. Not to make this into a distro flame, but I always found Gentoo a real pain to install. Basically, I never had time to deal with all the issues and without a full knowledge of how a lot of the under-the-hood userspace-kernelspace bridging works (init always tripped me up) it was basically hit the default options and hope it works, then when it didn’t, hope I could find the answer on the Gentoo wiki or forum. I did love the idea of the maximum ability to configure it for your needs, and compiling everything to be processor optimized, etc. I have some Core Duo 2s here with Intel compilers, and I can just imagine if I were able to put the 100-150% speed inscrease I get when compiling my research programs with it (as opposed to generic unoptimized i386, what Debian compiles things as. Using gcc with optimization the Intel compiler is only 20-30% faster on my research programs) towards all programs, but mainly KDE, Firefox, OpenOffice, etc.
Also, why not use Debian for that purpose? I’m not fully up to speed on this, but I was fairly sure Linux supported much more hardware than BSD. I know compiling a custom kernel (if needed) is much easier on Linux. One of Theo de Ratt’s main complaints were the BSD ‘drivers’ included for most devices that even supported BSD used some sort of weird Linux compatibility layer and weren’t really native. Also, Debian has never given me any install issues. The only serious Debian issues I had were configurational, and then only one really oddball items (root RAID with dynamic volumes on x86-64 under GRUB 1.07) gave me any long-term trouble. Everything else was either a kernel issue (not supported yet, waited awhile and it was), or a video card issue (i.e. video worked, but OpenGL performance sucked). The later issues be true under any Linux.
One reason Debian might be good for a PC bot is that you don’t have to install X or any packages you don’t want or need. A default install is about 150 MB and can be stripped down much further if needed. However, since large enough compact flash cards or 2.5" hard drives are relatively cheap, even this isn’t needed. You’re actually in a much better place to try a PC bot than I am. Your Walter could very easily be extended into a Whitebox robotics type machine. (whiteboxrobotics.com/2006/PCBOTs/spec.html) Since there is plenty of room for an ITX motherboard.
I wanted to the same type of thing, but with a biped. Hence my size problem, since a Scout isn’t large enough to hold an ITX board properly. However, a Mega-sized Scout probably would be. Also, a smaller Scout lets me learn about microcontrollers. I’m this (…) close to actually assembling Pete’s boards I got months ago. Lol.
Oh yes, Gentoo can definitely be and often is a pain, especially if you don’t follow the handbook instructions to the letter when installing it. I must have missed something during my last install, because I could not emerge anything successfully once I got to the chroot phase. Installing Gentoo is just a bit above doing a complete Linux from Scratch, which I have done four times. You can really super tweak your make options for everything with Gentoo.
I do have the latest Debian Sarge DVD images which I burned to DVD. I don’t fully get the package installation setup yet though. Yes, Linux does support far more hardware than even FreeBSD does, which is why I prefer it for a main desktop user machine. For some things, I prefer to build from sources, when a distro does not provide a current package and I am not sure how source friendly Debian is. Sometimes, I even prefer to build from CVS sources.
Very true. One reason I like Gentoo for experimentation is I can strip it back even further by just not building what I do not want or need. Even though Gentoo is a real pain at installation sometimes, I think the effort spent to get it up and running has very nice rewards in the end.
My Walter is barely 7" diameter at present, but I am going to make him a new 10" diameter body soon, with the help of a friend. He will still be octagonal shaped. I have designed three new deck sets for him in 8", 10", and 12" diameters, and one deck has tabs for leg attachment. I am planning to experiment with one of Pete’s boards as a new brain for Walter at some point, once I get all the parts together and get a few boards. I really don’t want to go to a complete motherboard, at least not yet.
8-Dale
Since I started this topic, I went ahead and moved it to the Linux section so we can continue.
8-Dale
The kernels are very different, including the way they are structured and built. That is the real difference between Linux and the BSDs. Otherwise, they use the same exact software, except for those items that are dependent on the kernel structure and for deep system level support.
I am running KDE 3.5.2 now. It’s the most Windows like environment available for Linux/BSD.
8-Dale
I am not sure I actually answered this question. Right now, FreeBSD is running my servers and is directly connected to the internet. I need to be booted into FreeBSD all the time on this machine. I did have it dual booting into Gentoo Linux also, but did not use it much because of the servers (Apache 2, mySQL, DNS, mail) even though I have had the same servers (including DNS) running on Linux also. BSD just has more of a “server feel” to me, even though I know very well Linux can do it just as well.
I can run the same software as I can run under Linux for the most part, so it doesn’t really matter which of the two I choose to run. I just happen to have everything running better under FreeBSD at present.
It takes a long time to build a large package on this 733 MHz Pentium 3 with 192 Meg of RAM, regardless of which OS I am running.
8-Dale
Not to nitpick, but I’m not sure if you understand how Debian works. If you don’t have an internet connection, you can use the DVD package repository. However, it is not really designed to be used like that.
Things are very different in the Debian world depending on whether you use Stable, Testing, or Unstable.
The whole idea is you select your usage choice and then periodically run apt-get update to talk to the Debian servers. It will then download a list of all the new packages and dependencies. This is digitally signed to prevent any foul play. You then can install packages with apt-get install, or more generally do an apt-get dist-upgrade to patch any security issues or upgrade packages depending on your choice of useage pattern.
If you pick Stable, i.e. Sarge, it doesn’t have too many new package changes. The only reason for changes in Stable is security patches or the discovery of major bugs. The idea is that if you want to run a server, computational box, moderately secure firewall, etc. run stable. No upgrade should ever break anything. It has all the older desktop options too, but honestly, unless you’re really bad at sysadmining or need extra security (where you should be running a hardened Linux distro or OpenBSD anyway…) you don’t want this.
Debian Testing, i.e. Etch, is like most distros most recent release. It doesn’t have the really bleeding edge stuff, but basically has newer versions of all the programs major programs. It also is changing constantly. Its very good about not breaking things on updates, most of the time. If things break, its generally limited to a few minor programs, and these issues are generally resolved very quickly. Because it is changing all the time, if you need or want to keep things as up to date as you can, you might stress your internet connection a tad. I think this past month I’ve downloaded about 1.8 GB per machine (I only have 7 desktops, the other 100 or so are research machines or servers that run Stable, so I didn’t set up a local mirror) worth of updates as there were a number of newer versions I wanted and while I was at it I just dist-upgraded to get all the security and other updates along the way. Out of this there was only one bug, which as that some additional Latex utility stopped working for a day or so. The idea with Testing is that when it reaches a point where most packages are sufficiently different from Stable, there will be a feature freeze and very little new stuff will go into Testing. This will generally be followed by a two to three month stability test, to hammer out any remaining bugs. Then Testing will become the new Stable. You might ask yourself what happens if one picked Stable durring such an event. The answer is that if you picked Stable (as opposed to the release name in /etc/apt/sources.list) your machine will update from one Stable version to the next, again generally without any incident. I had the majority of my machines go through such an update seemlessly from Woody (the previous Stable) to Sarge earlier this year.
Debian Unstable is essential one step away from compiling from CVS yourself. All upstream chanages are tested by the designated Debian mantainer then put into Unstable. Debian’s beta testers run Unstable and basically report back if a package has any issues the maintainer didn’t find. If no bugs are reported for a long enough time and all the package’s dependencies are either in Testing or also deamed stable enough, the package is copied into the Testing repository. Running Unstable is not a very good idea unless you really need the latest and/ or are doing development work for Debian since it really is a pain to admin.
Debian is also good about mixing packages from different piles. For awhile on some machines we had combination of Testing and Unstable packages, since there were some KDevelop features of 3.4 that I really wanted and it had dependencies which for some reason or another couldn’t make their way into Testing.
The real way to install Debian is to download the 180 MB netinstall iso for Testing (i.e. Etch), which contains all the core packages and a good number of most popular extras. The installer will then help you set up which Debian servers you want to ask for updates and you then only download the packages you really want.
A few more things… sorry for the rant 
but you can also add 3rd party package repositories. Debian is a little anal about the packages it will allow into the official distribution. Things deemed not open enough get pulled. LAME, the MP3 encoder is one of these since it is under patent protection in most countries even though the patent holder said they won’t sue and the project is GPLed. The whole Firefox/ Iceweasel things is another example of this, the Mozilla icons aren’t GPLed and Mozilla won’t let Debian distribute Firefox and call it Firefox without the icons. Hence, sometimes if you are willing to put less free software on your machine to get extra functionality, 3rd party depositories can be useful. Honestly, I think I have 2 non-official Debian packages on my desktop machines (the file servers have more 3rd party packages from IBM since the University uses a proprietary IBM backup utility), though once the Firefox change is made I’ll probably remove the official Debian version and add the 3rd party since my users will complain. If its only one or two packages, you can download and manually install which is what I do for the desktops. The packages for the desktops are LAME and Acroread since sometimes the non-Adobe PDF viewer doesn’t work, i.e. your 3D PDFs. For IBM’s stuff on the fileservers, I just link to IBM’s repository. Since the program only works if you have their equipment they let anyone download it. The final thing is that I’m a bit biased (as you can see) being a Debian sysadmin for my lab and regular contributor to the project since I do have a few machines running unstable or unusual configurations which do find bugs other people miss.
Also, if you only have a PIII for playing around with Linux, I don’t think you should be compiling unless you really need the newest version. Debian Testing’s packages shouldn’t take more than an hour or so to go from base system to fully-functional desktop assuming a reasonable internet connection on such a machine.
I am on a 6 MBps/768 KBps DSL connection.
I do not want to get locked into a usage model setup by somebody else. This is exactly the problem I have with most Linux distributions - I have to settle pretty much for what they want to give me and when they want to give it to me. This just doesn’t work for me.
This is exactly why I don’t use distributions like RedHat, Fedora, or Mandrake anymore and it is why I prefer Gentoo or one of the BSDs now days. I would rather take how ever long it takes to build everything from sources and end up with exactly what I want in the system.
The other thing is I insist on having as updated a build as I can get of certain things, such as KDE. I may even choose to build KDE from CVS if there are things I want right now, regardless of how long it takes to build it.
I may take a look at Etch, but I am not yet convinced Debian is the right distribution for me. Every Gentoo installation I have done has been rock solid and I have rarely had a problem upgrading it when I wanted to. I typically do not care for any kind of prepackaged stock distributions.
What this is sounding like is that you can’t easily get some more updated packages from say Unstable, unless you committ to running the full Unstable tree. Is that right?
I will probably give this a try at some point, but right now I don’t want to invest time in learning a new distribution.
Interesting. Gentoo and FreeBSD don’t seem to have this issue, or at least KDE doesn’t. I have icons for all the Mozilla applications in KDE. I don’t know if the icons came with Firefox and the other Mozilla builds, or if they were installed with one of the KDE package builds, but they are there.
I really do want/need the latest and greatest versions of some things, and I don’t mind spending time building them on this Pentium II/733. I don’t think I would spend the time to build an entire system now though, since this box is right on the internet and running important servers such as primary DNS, but otherwise I would not mind a bit.
It’s great that you have found a Linux distribution you like and that works great for you. Just understand that what works well for you may not work well or satisfy me. In addition to having built several Gentoo systems, I have also built several Linux from Scratch systems.
8-Dale
Again, let me clarify. The usage pattern is only for the freshness of the release. If you want rock solid, you use Stable, if you need pretty stable but want newer features, you use Testing, and if you like to live on the edge with the newest stuff you use Unstable. This is no way makes any choice to what packages you have installed. At the very end of a Debian install it will ask what type of packages you want to install and give a a bunch of common options - web server, desktop, etc. The best thing to pick from my experience is manual package selection since otherwise you end up with a lot of stuff you don’t need. You can use apt-cache search to look for specific programs in Debian’s vast repository (100,000+) of packages. Most stuff is named fairy simply. For example to install kde from a base install you need to apt-get install kde3 xorg kdm. This would install those things and all their various dependencies (100s of packages). Debian’s website also has a search that not only lets you search for names, but things inside packages, advances dependency features, etc.
I can certianly empathize with you on this. In Sarge I was always upset that there were things in KDE that were too new to be included in Stable. However, since around KDE 3.3 I haven’t had too much reason to upgrade the entire KDE tree. My main gripe has been KDevelop so I could write Qt (and get the portability they advertise) and native KDE programs.
No, I meant the exact reverse. If you need packages from a less stable tree, you change testing to unstable (or stable to testing) in /etc/apt/sources.list, run an apt-get update and pick the newer packages you want individually with an apt-get install . An apt-get dist-upgrade would update your entire distro (i.e. all installed packages), which you don’t want. However, say an apt-get install kde3 will only install out of tree packages for the thing you specifically asked for (in this case kde) and dependencies. Then you would change the sources.list back to the previous value and run apt-get update. The newer packages will not be overwritten with the exception that if you’re in testing, it is possible that an even newer version of the package will come into the respository.
Hm… First check that the icons are the Mozilla icons. Friefox on Debian right now has a world icon and uses other similar GPLed icons. Also, Debian requires icons, documentation, etc. (not just code) to be GPLed or suitabled open, the BSD, MIT, some creative commons, and a few others also fly. The Mozilla license for the icons was not deemed one of those. Other distros might not care about icons since the code is GPLed. I never got Gentoo anywhere near able to be running Firefox, so I don’t know. SUSE 9 (before the Novell takeover, I haven’t used newer ones) and Red Hat don’t care. On OpenBSD the icons are used (surprisingly) but I think this is mainly an issue of the OpenBSD version of Firefox is so far behind (doesn’t even used gtk2) it hasn’t hit the one which requires the non-GPLed icons. OpenBSD may be secure, but it honestly makes Debian Stable look uptodate by comparison.
Congradulations. I haven’t been able to do either all that successfully, though honestly the last time I tried was awhile ago and my linux skills have gotten better, hehe. There is an option to build Debian from Scratch, and suposidly options for apt to work like emerge (i.e. instead of downloading the precompiled package, downloading only source and compiling) though I haven’t explored those areas.
I’m not saying that you should use Debian. I didn’t know you had so much success with Gentoo, since I never did, and many of my friends here who use Linux have had issues learning Gentoo. Honestly, I think this whole conversation might make me put getting a working Gentoo system on my list of projects. However, I find installing Debian from a wiped machine to something fully useable to be faster and easier than say installing XP.
I did check my icons for Firefox and Thunderbird. They sure do look just like the ones at the top of the page on the Mozilla website to me.
I just did a bare bones Gentoo 2006.1 install (dual booting with FreeBSD). After I get X and KDE installed with Firefox and Thunderbird, I’ll check the icons available for them.
I’ll be here to help if you need any assitance installing Gentoo.
Installing Gentoo really is not that difficult, although it does require you to have more knowledge of your PC hardware at several steps. I still don’t get why they want you to install pcmcia-cs with a 2.6.x kernel if you aren’t going to have any PCMCIA devices - I did not install it. I went outside for my kernel and installed a vanilla 2.6.19.1 kernel instead of one of their gentoo kernels. I’ve been working with Linux since kernel 0.98, so have been around the block with Linux a few times, many distributions, and lots of kernel changes.
I am really looking forward to working with robotics and FreeBSD and Linux. I just have to get the SSC-32 installed with the ABB on W.A.L.T.E.R. now. I’m not going to connect the ABB with the SSC-32 though, since I don’t have a Windows PC to work with at present. I’ll just run the SSC-32 directly from FreeBSD and Linux for now.
One thing you have to be sure of is that you get your clock set correctly at the appropriate stage or you will have problems with emerge later on. That was my problem with my last attempt at installing Gentoo.
I don’t know how long it is going to be for me to get my Windows PC fixed, so I may have to do my robotics in Linux and FreeBSD for awhile. I could use a Bluetooth link to W.A.L.T.E.R. now, and will prefer to work with Linux for robotics due to it having much better hardware support.
8-Dale
