aRTee

Anna is right, you need UDF/packet writing, or MountRainier/Easywrite; the latter should be supported by the drive in its firmware. UDF is to be avoided, cdrw discs are more problematic than floppies in the sense that they possibly accumulate more dirt, fingerprints and scratches. UDF has error-correction, the mechanism is better than for cd-rom (or cd-r(w) written in a normal way), that's why a 650MB cd-rw will only give you 550MB or so when using drag and drop packet writing (I used easy cd creator when I still used windows). But this extra redundency is not really enough if you don't take real good care of your discs; one fingerprint on a cd-rom (cdr) will create read errors, that's okay, they thought of that. On a cd-rw it will create write errors the next time you use it; the data may not be written properly. In comes Mount Rainier/easywrite. It has a much better error management mechanism, which also means you get some 30MB less (520MB? out of 650). I saw some messages quite some time ago on usenet, Jens Axboe was working on it, and it should work fine. It should also be included in kernel 2.6, and some 2.4 kernels may have it backported. In any case: for now, just write a cd-rw as you would a cd-r, don't use packet writing, not on windows, not on linux. It's not so bad, you can do fast erases and just rewrite the whole disc. And be very careful with the disc! Never touch the surface with your fingers etcetc....

small hint, in case you can't find or don't have the nvidia installer anymore: on the command line / in text mode you can use the browser lynx: lynx www.nvidia.com then download the installer.

On my sys it was no problem when I installed the second 512MB mem stick. True, I only get to use 890MB (or something like that) of the 1024, but since my gkrellm (or on command line: free) tells me that I don't use even 800MB I just know it's not worth it recompiling or using the enterprise kernel. Did you see if your bios settings are ok? For instance, the amount of mem allocated for the graphics buffer; my guess is that should also be 256MB...?

tyme: LUCK! :)

Have a look here: http://linuxtoday.com/developer/2003071400826NWKNDV Note Linus' request: So give it a try! And maybe Mandrake will offer a 2.6 test kernel in 9.2, I for sure will vote for it on the club rpm request list.

About OpenOffice: 1- you can turn off the splash screen, forgot how, somewhere in the config menu's 2- when your system crashed and you reinstalled, did you keep your home partition? There may be some crappy / broken config settings remaining. Check for files/dir named something like .openoffice (make sure you turn on 'show hidden files' in your filebrowser, or on the terminal use 'ls -a' ) in your homedir, rename it to something else. Then retry. Alternatively, create a new user and log in as that one, then try OOo and see if it runs properly again. As an aside note: welcome aboard, and hope we can help you fix your remaining problems in the next couple of days. As someone else hinted: in some cases (kopete?) you may have to resort to another program. If OOo was running properly to your liking before, I wouldn't switch though, just get it running like it should.

You can also check my specs (see sig) to see what my config files look like. Normally setting up k3b should not be a problem, but I've had problems with it too at some point; similar to yours, I could not see my cdrom devices anymore in k3b; had to go through the thing a couple of times to get things right. BTW before k3b I used gcombust, urpmi gcombust to install it; definitely for iso's it's not hard to use; for all other purposes the interface is quite confusing with all its options. You have to go to the 'burn' tab and then just select the iso and 'burn iso' option... And do the device setup, at the beginning. Since it tries to do less than k3b, chances are it works right...

Well, guess it all depends on where your priorities lie; personally, linux support comes first. Without that I would have to reinstall windows and that's out of the question... BTW for 150 euro's there are nice 3x camera's that claim no driver install in windows, and if true, they must / should work with the usb-storage thingy under linux..

As if that's not a word! Hah! Well, it is when used as an adjective: the userfriendlyness of linux largely surpassed that of windows. Anway, don't hold your breath, it will be another couple of weeks. I will put it up for discussion in this forum first, as usual.

Hmmm, maybe I should put an 'avoid this manufacturers hardware' page on my website... ;) If all linux sites do so, and all users advice anyone that 'if some day, you want to get away from windows, and really, you want to keep your options open' so don't buy that unsupported crap! maybe it would change their minds. Remember, they have a bottomline, and it's all about money. Touch them there and they will have to change. Anyway, note to self, don't buy the Jenoptik JD 2100 z3s (aka 'Traveler', aka 'Skanhex'), and tell anyone else to avoid that stuff... :D

If you make a backup, why not just wipe the whole disk? Aside, I'm not sure if I get your question completely,... you want to use different partitions for the things you are using them for now? Remember, if your disk has some primary partitions, then some logical in an extended partition, and then your linux partitions, you will not be able to put the first ones (primary AND logical in extended) together... That's why maybe just wiping the drive clean may be much better/easier.

Yeah, it's really cool, PCLO and desktoplinux also posted the link. I've had over 2000 hits from IP addresses that never contacted before. Pretty good for 2 days... :) Next will be an article on userfriendlyness and even 'making a case for the command line interface'... but I have so many other things to do, it may take a month or so, maybe by the end of August...?

30 bucks extra for the extra fan on the PSU? Are they mad? Not worth it. 5 bucks for a separate fan is more than enough... About the win98 crashes, true, if properly set-up and not (de-)installing stuff all the time, and not doing too many crazy things, one can live with it, more or less. The pci gf2mx should perform very very close to the agp version, unless your pci bus is overly crowded. Shared IRQ's may also be a problem. Can you tell me what cards you have in your machine, and what onboard stuff? For instance, my mobo shares the IRQ on the AGP slot with PCI slot 4 (I think), so I left that empty; sometimes you can assign a different IRQ as well. In any case, your pci graphics may be sharing too much with some onboard stuff, like the audio you mentioned.. maybe the 'slot-dance' can improve..?

Small addition: Diskdrake is the wizard, and doesn't allow you to do your own wizardry. Meaning: I have a heavily edited /etc/fstab, and when diskdrake (luckily) prompted to ask whether to write it, my alarmbells started ringing. So I made a copy. Sure enough, it had demolished and left out quite a few of the good parts, the (possibly illegal) entry I had once made so my dvd-rom can be mounted as scsi, but in safe mode will be mounted as standard ide; the extra entries made by and for k3b (that would p!ss me off, not being able to burn cd's anymore, until I would find out what caused the problem). On the whole I like diskdrake, since I am careful and mistrust it. And don't expect too much. And I always check the drive numbers etc. But in my experience they never got swapped...

Just a comment on the side: 1- I completely agree with bvc, a simple thing like a usb mouse should work 2- for those interested, check my config/specs (see sig), I now have 3 mice, 1 normal wired wheel+ball mouse on ps/2, 1 optical wireless (with extra buttons) logitech mx700 on usb, and LIRC mouse (remote control via tv-cards remote eye). I really want to keep the ballmouse since sometimes I forget to put the wireless one on its cradle to recharge. It still lasts for days, but then, I sometimes forget for days to put it back... ;) BTW yes I have usb2/hispeed, and I even think the mouse may be plugged into that. I also have a usb2 scanner. On my system, usb1 and 2 both work like a charm. So I'm not sending any bugreports to Mdk, nothing to report on hardware support -- actually, I'm in linux nirvana since all works! Well, except the lmsensors readout for the cpu-fan speed...

"payed" did I write that????? Aaarghh! Took over your other comment too (ensure..) Well, guess that's what you get when working long nights to finish an article.. Thanks a lot! If you find the time, please tell me any other error I missed. Personally, I've been editing this article too long, so I don't see the mistakes anymore...

For your info, my athlon 2400+ (2GHz, 13x154MHz) runs at 52degC, if I load it it goes up to 67 or so. I have some experience (from my past/first job) on high temperature electronics and laughed at loads of the first reports about the Athlon (thunderbird) power consumption. It went up to 70W, which is much, compared to P1 or PII or even PIII. Then people started to shout: hey this can't be good etc. Who cares, if AMD says their cpu's can run this hot, and guarantee it, then it's fine with me. We had electronics that didn't consume so much, but still a few watts, and we used it at temperatures of 150, 180, 200 or even 220 degrees. Yes, I am talking Celcius!!!!! Even our chip providers would only guarantee 125degC or 85degC, they had no idea we could push their electronics so far. My point: if the manufacturer says it's ok, and guarantees it, why would you care? Except about the power consumption, you do get the bill for that. Oh and you don't want the inside of your case to heat up too much; the better cases have the power supply suck in the air from below, where the cpu is, not from the back. That way, less heat stays in the box, so you hd's etc stay cooler. Also, you can put an extra fan, if you want it quiet put it on 7 volts, there's normally a place for a fan right next to the cpu; make sure it blows to the outside. BTW: The P4 uses as much power as the Athlon, so no big difference there. Yes, the P4 is larger (less power per area) and has a heat spreader. That last thing actually means there's an extra barrier for the heat to go through before getting out through the heatsink/fan.... Another thing: only since the t-bred did AMD put a temp sensor (partially) on chip, before it was the mobo that would read the temp from under the cpu. Of course then people suddenly had 10degC higher readings... Anyway, don't worry as long as your onchip sensor tells you some temp below 75degC. On graphics: The Geforce MX series were ok until the standard gf2mx. All mx's after that are lesser parts and to be avoided. The GF4mx is only a GF3 in architecture, with some parts missing; same for the gf3mx.... What exactly are your problems with your gf2mx? Just that windows crashes? That's normal :D Anything else? Small correction: agp does not go directly to the cpu, it is a separate channel to the chipset (southbridge, in case there are 2 chips in the set, some make combined chipsets, so actually only 1 chip); the pci bus is shared over all pci devices, including any onboard things, like audio, lan, cardreaders, usb, etcetc. Normally, the cpu is connected to the northbridge, which links to the memory and the southbridge. The southbridge links to the agp, pci, etcetc stuff. Forgot where ATA/IDE goes, but I think it's on the southbridge too. Basically, on the upcoming Athlon-64, the memory controller is integrated onto the cpu, so there's no northbridge.. On heatsinks: My main thing in heatsinks is that they have to cool well, but at the same time be silent. Bad thing: in shops they just have no clue... but anyway, read hardware forums, they will be able to tell you which hsf to avoid. Also, there are plenty of tests on the web on hardware sites.

Hi Darkelve, thanks for the links, I had a quick look them, and it seems I have quite a bit to do to make my site more presentable. Maybe it is because I normally just adjust the size of the window I read my site/any site in to be comfortable; I don't like sites that don't make lines long enough (for my taste); that being said, the point is not that I like my site but that it's usable. So I will take the time and see how I can improve the site.... That being said, I was actually more asking if you agree or not on the contents... but since you had no comments to that... :D

bvc, the basic problem is marketing. To give you a slight bit of my background, since most may only know me for being a linux advocate (with my site and all)... I'm an electrical engineer, microelectronics designer, analogue/mixed signal; during my studies I knew quite a few fellow students who were crazy about computers; I studied from 91 'till 97, computers went from 386sx to the first pentium IIs in that time, and win3.11 to win95 (98 ), or on a linux scale: linux 0.1 up to ehmmm CorelLinux more or less. Being surrounded by freaks I picked up some stuff, in class I picked up other stuff, like basic computer architecture. RISC vs CISC etcetc. I didn't like computers since they were too slow to do anything fun, like creating your own cd's, mixing audio, video capture and editing etcetc. After finishing my studies, I found that computers had moved up, memory was more affordable, harddisk sizes became larger, so it became interesting. I followed the market, and with my background, was capable of understanding more than many other people with an interest in computers. Note that at university they had never explained the Intel x86 architecture. Reason: economically it was extremely successful, technically it was considered a disaster. The 386 was just a boosted 286, which was just a 8086 with loads of addons, the dx finally had the fpu integrated into the cpu, it was a mess... When Intel had problems going to 66MHz (486), we were doing practical work on 99MHz RISC cpu's from HP. They were much faster, they could emulate a 386 and it would have the same performance as a 386 at 156MHz!!! So later, I understood: the architecture determines what a processor is worth at a certain clockspeed, combined with the memory bandwidth. Now, zoom back to the problem today: competing with different architectures. Since the 386 AMD has been making Intel compatible cpu's, same architecture, same motherboard/platform, same 'environment'. Hence, with a higher clockspeed comes a faster processor. So clockspeed, expressed in MHz, became the bottomline to compare cpu's. Then ( ?IBM and ? ) Cyrix had a flirt with 'ratings', they had a different architecture, their 150MHz 686cpu would run integer calculation as fast as a 180MHz Pentium (if I recall correctly)... So how to sell this beast? Selling it for a 150MHz cpu would not do them much good, it was a better design, performing better. Clock for clock, it would handle more instructions. So they stuck the rating 180 onto their 150MHz cpu's. But it backfired. Floating point calculations had become more important, I don't remember why,.. and their '180MHz-equivalence' rated cpu's were not all that... Since the Intel Pentium, other companies, including AMD, were not allowed to copy the cpu design. So by definition, they had different architectures. From there on, things started to diverge. AMD had the K5 and K6 designs, which were partially from other companies, the first k6 was from Next if I remember correctly. All in all, they had more or less the same performance as the Pentium (II) per clockcycle, except in floating point (fpu) performance, where AMD's cpus were quite a bit slower. As it happens, fpu calculations are very important for games, so AMD couldn't sell as well. Things changed with the AMD Athlon. That was an ace in the hole. It was faster than the PIII, came earlier than Intel had expected. It was faster clock for clock but also could ramp up faster, so AMD beat intel to the first 1GHz cpu. MHz was still everything, since as before, the performance per clockcycle, called IPC (instructions per clock), went up with each new architecture. Then Intel came with the P4. They were late, had had to alter their roadmaps to deal with the Athlon. Public statements of Intel in 96/97 about the AMD cpu were that it would probably start at 350MHz in 98/99 and be at 600MHz in 2000. And they based their roadmap on that. And on the idea that the K7 (athlon) IPC would not be higher than that of the PIII. Wrong on both counts, AMD managed to have higher IPC and reach 1GHz in 2000. So they had to mop up. The p4 (willamette) was a too young born baby, it seemed designed to get high framerates in Doom, but that was all. They had implemented very cool features, like the P4 tracecache, but not enough memory. The p4 really needed/needs a very high memory bandwidth, that at that time only Rambus RDRam could give. But it was too expensive. In any case, a P4 1.4GHz was not really faster than a PIII 1GHz. For the first time, a new generation actually had a lower IPC. And there is the problem. Marketing. How could AMD sell its Athlon1.4GHz (Thunderbird) when Intel was offering its P4 at 1.8GHz at the same time, for the price it was worth, performance wise?? People don't know/understand IPC, processor architecture, cache and RAM bandwidth. But they can read: 1.8GHz and compare it with 1.4GHz. Athlon/K7 has exclusive L2, 16-way set associative cache (on chip memory). The PIII has inclusive L2 8-way set associative cache, and the celeron 4-way, half the L2 cache of the PIII. The P4 has trace cache, and has 20 pipeline stages. All of this makes the difference. So comparing an Athlon at clockspeed X with a P4 at clockspeed X is comparing apples and oranges. But 'clockspeed X' is easy to write.... So for the next iterations of the athon, the Palomino, Thoroughbred/T-bred, and now the Barton, AMD decided to use the old Cyrix trick: The Rating. The thing that made Cyrix the one forgotten cpu maker. Very risky. Marketing. Stick 1800+ on it and price it lower than a P4 1.8GHz, but make sure it has the same or better performance, and you're set. Originally, they had the standpoint that the rating numbers were comparing the new K7 iterations with the older Thunderbird (that went up to 1400MHz), since there were some slight architecture changes that indeed improved the IPC. So a 1800+ running at a real 1533 MHz (I think) is as fast as a (hypothetical) Thunderbird at 1800MHz. Which explains why it is faster than a P4 1.8GHz, because a T-bird 1.8GHz would have outrun that P4 too. Now they are getting closer and closer to shooting themselves in the foot again, ... the rating is now used more to compare to the P4. But the P4 has been altered, it has more cache, of which it greatly benefits (it is still memory hungry) and SSE2, and possibly some other architecture improvements. All in all, the rating is just a number. For all practical purposes, just check your wallet, and see what AMD or Intel platform that money will buy. Then check on benchmarks which of those cpu's performs best, and focus on the benchmark that looks most like your computer activities. If you don't play any 3d games, don't pay attention to those benchmarks. etc.. One other thing: you may want to check with which platform/cpu you can most easily upgrade. Remember: memory upgrades are easiest, next are cpu upgrades (sometimes needing a new bios to properly recognise the cpu). No software hassle, just reboot and it works. In the recent past, AMD has been better for cpu upgrades than Intel. Today I'm not sure, since AMD has the 64-bit Athlon-64 coming on (clawhammer)... On memory bandwidth: again, comparing apples and pears. P4 has 400MHz qdr ( = 100MHz x4 ), then 533 (133x4), now 666 and 800 Mhz. Athlon has moved from 200 to 400MHz (100x2, 133x2, 166x2 and now 200x2). But on Athlon it is 64 bit, and on P4 it is (if I remember correctly -- I'm now more focused on the upcoming AMD cpu's) 2 channels of 16 bit. Do the math. The P4 architecture has always had more bandwidth, but also needs it more than the Athlon. Basically, for a 2.4GHz P4 vs a 2400+ athlon, they are more or less the same speed, and where I live the athlon based system is cheaper. Problem for the consumer: these numbers, they are just marketing. MHz and GHz... So what to make of it...? Same for AGP. I don't think you really need or will notice a difference between AGP8x and AGP4x. 128MB of RAM on the video card tells me that it will rarely send any info through the AGP bus anyway....

In case anyone cares, it's up now. Enjoy!

Well, didn't know that rename only takes care of one %20 per filename, but I bet it's still faster to type arrow-up, enter a couple of times than actually do things with a script, if you have to figure out how to do scripting... ;)

On my graphics card, in the manual, it explicitly said the hardware GPU was fully supported. I have seen a samsung laser printer (cheap!) with large letters on the box: Windows and Linux compatible (RH7.2 or so was mentioned). My motherboard had audio drivers for linux included, and a friend just bought a similar board where also the onboard LAN driver was included, with the instructions (readme) in the same dir on how to compile and use them... Yes, it is happening.

The same happened to me, see here: http://karely.emmanuel.free.fr Have a look with IE too,... out of protest I made the extra pages to get people to use something other than IE. Note that IE for Mac works fine! So yes, they have crippled .png transparency on purpose, since Unisys is a good friend of MS and owns the gif patent, which currently is the only transparency capable image format. But gif is only 256 colors... :( So that's what you get if you allow a company to be a monopolist... Sorry, I don't know what to do about it... I just caved and converted to gif..

Easiest way: just start the installation of 9.1, choose install (don't try to upgrade it), then select to install it on the old 9.0 partition. It will then ask you if it can format that partition, make sure it does, and make sure it doesn't format any other partition that you want to keep. BTW, make a /home partition too... trust me, you want that. The mdk9.1 system partition doesn't have to be larger than 4GB or so (unless you want to do very special things), so you can delete the 9.0 partition, then create a new / (4GB) and a /home (6GB) partition in the same space.

Read the manpage for 'rename' man rename or in konqueror, put it in the location field: man:/rename basically it works like this: rename "string you want to replace" "string to replace it with" [filename to do it on] In you case, you can do rename "%20" "_" *.mp3 which will change all the %20 in all filenames that end in .mp3 in your current dir into an underscore. To do it on any file, just do: rename "%20" "_" * and rename "%20" "_" */* will do the same on all filenames one directory level deeper. Just use arrow up and add a /* each time, to do it 5 levels deep in 5 seconds... ;)