So far, none of the bugs I have reported seemed to have been really touched – we dont even have to talk about resolving. One bug is really severe. I must say that this is the worst bug response reaction I have experienced on any distribution so far. Is this because there are so many bug reports? I dont know. But sure enough i will have to switch the distro if my notebook keeps crashing multiple times a day. This is because I will get a data loss at some point. I would have installed OpenBSD if this would also provide flash playing ability. I hate that situation. i think flash is evil – OTOH flash works really well on Linux and also on Linux it is a nice replacement for other proprietary formats. am sure technically it would just work with Linux emulation. I got it partly working with Opera+Flash once. But then again, I dont like Opera because its ugly and proprietary.
But I think I will install OpenBSD on the notebook anyway, soon. I dont see any benefit in running Ubuntu for me personally, any more. I generally like Ubuntu because it has made a free operating system more popular and it gives a common user base from which one can build on. But something like a freeze is a bug that needs to be resolved and if this bug is still untouched and not even decided if its serious or not – than actually there is no support and i can not expect that anybody will EVER touch this bug. As I am working daily with this notebook I can not risk on loosing the data. This troubles started with the last upgrade. That brought a finally stable working wireless but then this freezes.
Another annoying bug is that the mouse displays weirdly and does not react any more. I then solve it with pressing like: ALT+F8 and ALT+F7 and the display gets redrawn and the effect is gone. I now have seen that I can get this every so and so seconds if I use EOG to view images (not so much with gthumb, though). This bug is also untouched still.
Ubuntu is not the only distro that doesnt resolve issues. But for me, personally these bugs are not acceptable any more. I dont demand a resolution NOW or SOON, but I demand a reaction on serious issues.
I get the impression that on Linux you see really a lot of changes in packages – like Ubuntu and Fedora (Update: The information that Foresight also already uses FF3 was wrong! my fault! Sorry for misinformation!) using FF3 beta with the effect that many web designers cant use Linux any more if they depend on extensions that arent ready for FF3. Its like partly they dont really care about the user. its some kind of overpopular. So popular so that you dont care about single issues. OpenBSD supporters often dont react positive to criticism and requests but I think they would never replace FF with a new version that degrades the functionality out of the respect for users who depend on some functionality. I would consider this to be more user friendly than the actions on the Linux side. And thats what actually brought me to OpenBSD: If I use my computer I like it to provide a given functionality also on the next day. I also like new functionality and new software – but I the prefer to rather compile experimental stuff myself or get into ports/package building.Mostly I should be satisfied with what a distribution provides me and which was good some weeks ago – or is good enough for all other users. I now think the “latest and greatest” is for developers.Linux distributions that give that to users actually do something what Microsoft or Apple does to their users: make them public beta testers. The problem is that there is never and endpoint where you can say: Now Linux is stable – development goes on and on and there is no time to breath. And at some point one can get tired of all the new stuff that is not working. I still cant say that all that is bad. Linux is a testbed for new GNOME stuff for instance. This means OpenBSD can wait an learn from what they do. I am not sure if this helps getting some stuff faster into OpenBSD. I think some technologies depend so much on Linux that OpenBSD cant really follow and that if some stuff would be build with the existence of OpenBSD in mind a lot of stuff would be much faster also be available in OpenBSD (like NetworkManager).
Right now I cant really help building ports for OpenBSD because I dont want to use my working OpenBSD productive desktop as the basis (which I rather maintain with snapshots). I have asked some people via IRC and mail to give me advice on how to set up a build system additionally but havent got any responses so far. So any advice here would also be welcome. I know the tutorials but thats not the question. The question rather is: How can one make a build system that doesnt conflict with a stable OpenBSD installation? is it enought to have some kind of chroot environment to build packages and then install these testwise on the stable system? or would I need a dual boot installation? How do you manage this? What options do I have?
Today I found this description about OpenBSD philosophy:
Unlike some of the other Free OSs, the leaders of the OpenBSD project have no desire that everyone in the world runs OpenBSD. Perhaps more than usual, most of the OpenBSD users believe in the “Right OS for the Job”, not “Our OS is The Answer, now what is your question?”. (here)
I cant say that this was new to me, nor that it surprised me. But somehow I found this view refreshing as an opposition to many distributions that maybe too much think about marketing. Maybe marketing is a silly concept at all? I can imagine some of you might say you think like that for a long time. But maybe thats maybe because you never seriously thought about it as a possible partly positive concept?
One other thing came to my mind today: I have invested large parts of my free time into packaging some months ago, but now I dont do it at all. This was due to my switch to OpenBSD on the desktop. This was mostly a smart move. I am only missing a few things now. But then maybe some of you also know this effect: When you are into something you think its important and you think you need to do it – until you switch your habbits – and then you dont do it any more – and if you look back you maybe cant imagine why you really did invest THAT much time into that. I think this is true for many hobbies. We are all kind of blind. We all think we only have the time we use and we also think what we do is essential, while mostly I would estimate that most of what we do is (sorry) useless crap, trumpery. This is also true for the business world. And thats also why the rubbish piles get larger and larger – people consume more stuff and throw it away. Maybe what we need would be a different culture – one which is already there – like: be efficient in what you do but have fun and cultivate your lifestyle – without using all this bloated products … just live and act. Maybe I am just tired of all that nonsense you hear every day (or have to buy). To me it seems we tend to live a copy of life and not live they way we could. Its all just a cheap copy. But maybe its just the view from me of my life – but I see this also in other peoples lifes, from the outside.
I now have virtualized all my servers with a vmware-server. I have chosen to use the rPath LAMP appliance for the web stuff and an OpenBSD Mailserver from Johan Allard. You might say why I buy a virtual appliance and not do it all by myself. The reason is simple.
- my main provider Firstdedicated decided to increase prices drastically (nearly 100%). Because of that I has an extraordinary right to end the lease which I did.
- But that ment I had to set up all servers very quickly.
I had only did some minor testing before so I was lucky to get a helping hand to set up vmware-server on my new provider Hetzner. The new server should run about 4-5 virtual servers.
As I looked more deeply into what the LAMP appliance does I was not so happy with some decisions. Like they configure it with no root password. This leads to a phpmyadmin that tells me that I shouldnt work with a root mysql access without no password. And sure thats true and also I have changed that. Then I think the LAMP appliance primarily builds on rPath – and is not built up from the ground. So there are a lot of packages that one does not need. Like Postgresql. I like Potgresql and indeed think its better than Mysql. But I think a LAMP appliance which means “Linux, Apache, Mysql, PHP” should just contain that. The on a simple “conary updateall” which should give me the latest packages I got a brand new kernel but without a vmware trove which is required. The question is why should the appliance update a very critical component like a kernel and then fail to update. Personally I think the kernel should be the last think to update because it requires a reboot and should only be updated either if it gives you very better performance, features or fixes serious bugs. On an appliance not being able to reboot is fatal. So I would rather expect an appliance to be most conservative when it comes down to the kernel. Personally I really dont care about any feature additions kernel wise. Where I do care most are updates to the latest PHP, because that tends to be vulnerable and all other components that are critical. But is working ok for now, but I think I will switch to an OpenBSD Apache appliance soon.
And after my server switch I also updated my OpenBSD desktop to version 4.3. This went quite well. I am still missing Epiphany 2.20.3 but there is a patch I need to test, soon. There are only a few other things I am missing on OpenBSD. When it comes to web this sure is flash support. I couldnt get that working ok with Opera and gnash, the free version has a very high CPU load still. Another thing is support for USB sticks. This doesnt seem to work dynamically/automatically as I would expect. But those sticks are very common, so they need to work if you want OpenBSD as a viable desktop alternative.Shouldnt be too hard. I always had a very high load average on this machine. I am now testing the “cwm” window manager as a GNOME replacement. Why cwm? Because OpenBSD works on this one as a new default as far as I understood. It was inspired by evilwm and so its similar to wmii. I think this is the right direction. Let me show you the quick reference:
C-M-Enter Spawn a new terminal.
C-M-Delete Lock the screen.
M-Enter Hide current window.
M-Down Lower current window.
M-Up Raise current window.
M-/ Search for windows.
C-/ Search for applications.
C-M-n Label current window.
M-Tab Cycle through currently visible windows.
M-S-Tab Reverse cycle through currently visible windows.
C-M-x Delete current window.
C-M-Escape Enter group edit mode.
C-M-[n] Select group n, where n is 1-9.
C-M-0 Select all groups.
M-Right Switch to next group.
M-Left Switch to previous group.
C-M-f Toggle full-screen size of window.
C-M-= Toggle vertical maximization of window.
M-? Spawn “Exec program” dialog.
M-. Spawn “Ssh to” dialog. This parses $HOME/.ssh/known_hosts to provide host auto-completion. ssh(1) will be executed via the configured
C-M-q Spawn “Exec WindowManager” dialog; allows you to
switch from cwm to another window manager without
restarting the X server.
for more see the cwm manual on OpenBSD. This allows very nice interaction with the GUI. cwm works on a file basis and has included some ideas from 9wm (plan 9 like window manager).The development of the official cwm has halted sometimes at 2005 it seems and is now developed as part of OpenBSD. This I think is one of the core principles of OpenBSD – a operating system as a whole – or lets say many BSD strive to do that. But like FreeBSD – they encourage to use Apache as a port . But this means that only some maintainers work on it – and its not considered part of the OS like in OpenBSD. So you cant really blame the OS if it fails. I think the big goal OpenBSD has and would go for if it had all ressources it could have is to have a full operating system that gives the people all the basics they need – so for the desktop – log in graphically, read the web, write mails, write letters, import photos from the camera. Right now OpenBSD mostly gives these features either per default via the command line or through maintained packages. So you get a thunderbird and epiphany and OpenOffice.org or gthumb. But these are all maintained by different parties, they are NOT part of the OS. The OpenBSD team just cant do it all, but I am sure they would love to. Just like the plan9 guys. I think those goals could be accomplished. I dont see such efforts in Linux. Linux has accepted the fact that there are countless projects which each fight against each other. Linux really just is the kernel – and they try to accomplish some minor common sense which you can see in the LSB (Linux Standards Base) and also partly in the FreeDesktop project (which also is for other platforms). The problem is that all those effors go towards a new Linux desktop which does not have a common vision and is basically a “Windows done right” or some wannabe Mac. But that cant lead to something convincing.
I have just removed the flinux.wordpress.com blog of me. Why? Because its primary purpose was to write about “Foresight Linux” in the past. There were many articles who talked positively about it, then as I turned my back on Foresight I changed the name and description and also wrote some negative comments about Foresight. Also the articles in their were very much similar to this blog. So I have decided to remove all old content. I am neither fixed on any sort of Linux or on Linux at all. I use what fits best and my comments are not ment for one specific OS, windowmanager or OS. I have really turned my back on many technologies by now. There is no much sense to talk negative about projects that are just aint worth it. I have seen a lot and basically it all boils down to “Don’t believe the hype”