How come these small things are always lacking in linux?
It is'nt. Try printtool. Red Hat's print setup GUI. I use Debian, but printtool is apt get'able. I set my Xerox DocuPrint P8ex up, using any driver that claims to do PCL (I used a Samsung printer driver) and from then on it just works perfectly. There are probably even easier to use tools, but why should I care? Printtool works for me, so why should I bother with something else?
With Linux (or *BSD for that matter), I can print to PDF from *any* app that can print to a file. Simply print to file and then use ghostscripts ps2pdf to convert the file to PDF.
Linux is not lacking, most users are. If you are someone who is not willing to RTFM after failing somewhere in a free OS, then go back to Mega$hit.
You need ram depending on what you do. For most laptop uses having more than 256 megs of ram is useless.
I would agree that 256M would be way more than enough for something like an OpenBSD firewall, with it's super small footprint. However...
On my Debian machine, installed at 2.1 with an on-going upgrade to Debian-testing, my/lib is 28MB and/usr (apps, X, fonts and other libraries) is 1.3GB minus/usr/src/linux. Having as much of this, of which I use most often, cached (Thanks to loads of RAM) will improve performance thanks to less paging and less fetching of data that has already be fetched before.
Of course i use linux and not bloated windowsXP.
Of course I am impressed beyond belief. I see you are a Mandrake junkie. Congratulations on your successful install of Linux.
As far as procesor benchmarks, I checked them myself, the 1ghz processor totally smokes the G4 at specific things, I mean like x4 the speed of the G4 at certain things. Some things the G4 is closer to the same level, but these things are very rare things, like using a certain feature in photoshop.
As far as ram, Dell supports over 1gig of ram too. As far as battery life, you win, Tibook beats dell.
I win? This is not a competition HanzoSan. How old are you? You obviously have little knowledge of CS, seemingly because you are young and closed minded. Trust me, if you want to compete here, although there are plenty of morons, there are also many people who do know CS very well, you will loose with this attitude until you open your mind. The smartest man, is the man who always questions himself and is not afraid of being wrong and adjusting his opinion accordingly. Here, ego will get you nowhere.
Having more ram wont give speed gains,
My desktop 7200rpm UDMA drives perform at about 25 sustained MegaBytes per second, my RAM "destroys" them (I have two, striped (RAID-0)) since my RAM sustains about 800MB/s. With the cost of RAM lately (with the exception of the recent artificial price spike), anyone would be CRAZY not to max out thier RAM. (That 25MB/s is not the stripe speed, but the raw single drive speed.)
Do you know what caching is? If so, what do you think of a 32:1 performance ratio? Would you rather be re-accessing apps and libraries from RAM@800MB/s or file-system/swap on disk@25MB/s?
To rub salt into wounds, my RAM might be considered slow by todays standards (PC100 SDRAM) and my disks way faster than any 2.5" notebook drive, so the ratio between disk:RAM would be much more pronounced. We could be talking about a 100 *times* speed difference or more. But even worse, I am comparing transfer rates here and not factoring in access times. Start using these disks with random accesses, with head movements that delay access to data in milli seconds (0.001s), versus nano seconds for RAM (0.000000001s), and you can see how cached data can effectively be thousands of times quicker (I know, 1nS is one million times quicker than 1mS, however, the entire time is not spent randomly accessing nothing).
having a faster harddrive will.
Of course it is a good idea to have a fast hard drive. But pouring money into your drives, and then skimping on RAM, is an ignorant move to say the least. BTW, the performance gains of loads of RAM, if often a performance area that benchmarks miss. Contrary to popular belief, benchmarks are acurate tools for measuring performance, but not overall performance. They are tools that measure specific areas of performance, that provide results that should be pondered and acted upon by someone who knows how to read and design for those numbers. If you do lots of 3D or even 2D CAD work, you will want fast floating point, 2D bitmap work will want fast CPU integer, loads of fast RAM and maybe good fpu also, video editing might need these and some real fast hdds or hdd array, a server might want loads of RAM, real fast networking hardware and a fast efficient networking OS.
Case in point for lots of RAM... Say my girlfriend is making a poster with Photoshop or The GIMP. It is going to be physically large'ish (typical poster), so high res is in order. Lets say she will be working with a bitmap of 4096x8192 32bit (8bits per RGB + 8bit alpha channel). The raw size of that, decompressed in memory, *without* layers, will be about 134MB. Now lets say shes using 5 layers for her various effects, that comes out to about 670MB being used *just* for the bitmap storage. Add to that OS and app memory usage, and consider whether you want to be doing this with RAM or disk. ; )
With something like this, whether you have a Xeon 2GHz or a Pentium 100, without enough RAM Photoshop is going to slow down at some stage to the speed of your hard disk. In fact, a Pentium 100 with 1GB RAM might "destroy" a Xeon 2GHz with 256MB RAM in this situation. So how important is your MHz now?
Get SCSI raid on a laptop and 1ghz chip will distance itself even more.
SCSI RAID on a laptop? Did you get some fine Jamaican hashish for Christmas or what! You think GHz is da bomb and SCSI RAID is an option for notebooks? Don't get me wrong, I would like a 2GHz G5 TiBook with 2GB RAM, a solid state 2.5" UDMA drive and OS X if I could have it, but I'd be quite happy with the current TiBook maxed out with RAM.
For now it still wins on encoding, decoding dvix and movies. It wins on gaming. It has a better screen, UXGA and if the guy above got the Latitude model which sucks its only because he didnt got the "network enhanced" model instead of the "desktop replacement" model. He had the choice.
I assure you, if you've got 256MB RAM and I've got 1GB RAM, I am going to enjoy my notebooks speed a whole lot more and a whole lot longer than you are. But I could'nt care less what you or anyone else is happy with, because I'm happy with how I set up my computers and thats what matters. This is not a dick measuring contest.
Looks like dell wins
For you, if you think you're the "winner", then good for you.
Lastly, The 1 gig of ram thing, how as if you need 1 gig of ram ofr a laptop,
There are very compelling reasons to max out your RAM on a notebook computer:
1. 2.5" notebook drives tend to be slower than 3.5" desktop drives, so the RAM speed vs drive speed is much wider on a notebook than a desktop, meaning that any dependance on drive speed (say for swapping or re-getting something that could otherwise be cached), makes the notebook slower than the desktop. The speed gains of adding RAM are higher for notebooks than for PC's.
2. Those little 2.5" inch drives are expensive and have higher failure rates, they don't generally last as long as a desktop drive. With more RAM, there is less head movement due to caching, which can lead to longer drive life and...
3. less head movement = better battery session life.
I was sold on the i8000 until I saw the G4 TiBook. I am glad the TiBook can support up to 1GB RAM and when I get mine, that will be the first thing I upgrade it to.
and the powerbook doesnt come with 1 gig of ram either.
At least it is capable of supporting 1GB.
I will take more RAM over Mhz any day. I cringe when I see people complaining (at various work sites I attend) that their P3 750MHz Dell notebook is slow (and they demand an upgrade), when it only has 64MB RAM, so they get the latest machine which is only 30% quicker as far as CPU goes, yet has 256MB RAM, and they think the enormous speed gain was due to the quicker CPU. Blah. Of course, being executives, they're not interested in what I said about the cheap RAM upgrade, they want whats on the pretty web site.
the 2.5k dell totally destroys the TiBook in every area, better processor, better monitor, more ram, (Tibook comes with 512 megs of ram)
Destroys? The P3 1GHz is close to the G4 600MHz in the benchmarks I've seen. "Destroys" seems to be a school kid way of saying, "my PC is 15% quicker than yours!". I like the TiBook screen for what it is (wide screen), but I also like the i8000 screen and the TiBook is capable of supporting more RAM. "Destroy" is something a 2GHz Xeon does to a 4MHz 8080.
Prove you arent biased
You were'nt replying to me, though I can tell you, after 12 years with x86 (some of that repairing notebooks for NEC and DEC), am I much more impressed by what Apple is offering. They offer extreme stability and usability thanks to high quality hardware, limiting what hardware they support and their efforts of extending the super workhorse OS, BSD.
It's the package, the hardwareOS+app meld that works so well. After all these years putting up with x86, my next machine will be PPC. I will probably never buy another x86 again (besides SBC's I use for firewalls, etc, although I might look at PPC SBC's for them also).
For high security, you can encrypt part of your hard disk through the use of a disk image. You can then email this disk image can to other people who know the password. Simply open the Disk Copy utility, make a new image and set the encryption. The image will show up as a volume on your desktop. When your Keychain is locked, or when you send that disk image file to another person, the image is secure. When your Keychain is unlocked, you can copy, move and delete files as you would on any normal hard disk.
I know how to use the loop device to mount fs images, but I would really love to be able to mount an ISO9660 image as writable in Linux or FreeBSD, to perfect an image before burn. And being able to encrypt the image would be neato too, though I know that can be done.
So is there a way to mount an ISO image writable in Linux?
PS. Notice the screenshot at that link shows the recommended crypto to be AES-128! The more I read about Mac OSX the more I want it.
Perhaps a solution Apple could use, would be a new type of focus-follows-mouse, that does one of the following:
1. Only change focus to the window below the pointer if the mouse pointer becomes stationary for a set period of time over a new window (say 250mS). This way, continuous movement over other windows will not annoyingly change focus.
2. Or a slight variation to the above, only change focus to the window below the pointer if the pointer becomes stationary or mouse/keyboard is clicked.
3. What about a "warp portal"! ; ) An area at the top left of every window that will warp the mouse to the menu bar, with ESC or a warp-back area at the top-left of the screen warping the mouse back to the focused window?
OK, I better get some sleep now I guess. Although #1 sounds pretty slick to me.
If this is already an option, please excuse me, my next machine purchase will be PPC/OSX but until I can muster up the money, I have to be happy with hanging out at the local Apple Centre.
The problem is that it's true. I've tried twice to install Linux (PPC first, then intel), and failed both times. Linux sure isn't doing much for me on the desktop.
I've been into Linux for about 4 years (I was curious seeing the kernel on Sydney BBS in 1993 but did'nt bother until Red Hat 5.0), MSx86 for about 12 years and various platforms from programmable calc/comps and various Apples to the Microbee and C64 (machine+basic) since about 1983.
So installing Linux or a BSD for me works almost always to some degree, even if I do commit hours to the task at times (grappling with hardware issues).
Having a bit of a love affair with Debian, Free&Open BSD (and enjoying QNX, Solaris and UnixWare), I'd love to dispute your point, but I can't.
Linux/BSD/Unix+tools are awesome for those who first know hardware very well and are willing to learn the ins and outs of the kernel and tools, but I've had some sort of troubles with almost every distro (barring SuSE, of which I've not used enough to complain of any problems), that I've been able to overcome with some effort that a newbie would have great difficulty with.
The various Linux distro's have had install issues that I can spot quickly and resolve, but what about the newbie? I realise that these issues mostly are due to the distro vendors, but tell that to the newbie who is viewing these experiences as his delve into "Line-Icks".
If I see erratic mouse movement in X, I think gpm, kill it and remove it from the rc's. X startup flashing with mostly black, XF86 config problem, Ctrl-Alt-Backspace, vi XF86Config, fixed. Many distros with really flakey partitioning utils I avoid prefering fdisk or cfdisk from a Ctrl-Alt-F*. Few distros have set up with X to a working GUI, leaving me to use xf86config, which mostly also fails to do so, leaving me to use vi to get it completed.
Hell, I tried Debian 2.2 (which I have now settled on (Debian-testing for home and Debian-current for work stuff)), and found the korn shell.deb to be faulty (in the official.iso). Notifying the Debian user community in a mailing list seemed to fall on deaf ears (weeks later it was "found" to be a faulty package), and 2.2 failed to install without install-script stopping errors on my AOpen AX6B. 2.2r2 even worse, kernel freeze at MCA drivers (yes I used the safe floppies to get it going)! 2.2r4 same as first, though I can get things sorted now. Progeny 1.0 has some flakey partitioning and can easily get it's knickers in a knot, requiring a reboot for another crack at it. Corel would not install unless I removed my Adaptec Ultra SCSI card. Caldera OpenLinux 2.2 progress bar getting to 700% and counting, never actually finishing the install. Red Hat partitioning also sucks and default install footprint and security is questionable.
I think the biggest hurdle for Linux acceptance is firstly the installer, and I must say, Mandrake and SuSE do it very well. Mandrake even allowed me to construct and install to software RAID (0 for/,/usr,/home and/tmp)!
Don't get me wrong, *I* LOVE Linux and associated tools. I enjoy working at the command line and most work I do gets done there. I rip audio CD's with cdparanoia, data with dd, have many CD's permanently mounted (Unix, Perl, and Networking CD Bookshelves, SysAdmin6, etc) and served with http/ftp/samba using the loop device, make anything I like into pdf by printing as ps to file and converting with ghostscript, burn with cdrecord, do most downloading with wget or rsync, work from home with ssh+scp and even fax from the command prompt.
I can do most of this from a GUI frontend with some setting up, plus use office type apps (StarOffice6Beta is a killer!), 2D/3D and PCB CAD, Corel Draw, Sound editing, movie watching, etc etc, but Linux is not about to take over the desktop soon. I would'nt rule it out though at some stage.
But if OSX is *the* OS of the future, I will be happy. Since, Linux in any incarnation is still far from user-proof, whereas OSX is here, now. It's pretty, it works, it's easy, it's hyped on the mass media, it has MS IE and Office (important to a lot of desktop users), it's supported fully by Apple and it is the future of Apple at least. But, once I get my G4 PowerBook and OSX, I'll still use Linux, FreeBSD and OpenBSD, because I am an enthusiest. Less equiped newbies to Linux will continue to run back scared to Windows or find MacOS, until some Linux distro makes install easy and widely compatible, usage easy and worthwhile, and provide high quality documentation to ease the newbie into becoming a hardcore Unix guru at the pace they want, if they want.
Needing to be a hardware and Unix guru to be able install and use Linux is not going to be the road to desktop dominance. Please don't reply with, "I installed Blah-Linux without any troubles... therefore you're wrong...", unless you've installed it on every piece of hardware out there with the same ease, what works for one system can fail for many others.
I agree with most of what you say. However, there are technical reasons why Motorola does'nt just run the G4's into the 2GHz regions.
I can't recall the exact reason, but I think it has to do with fundamental design of the pipelines. Intel being designed to use and exploit as much MHz as possible, whereas the Motorola chips working smarter.
It's pretty much braun vs brains.
If the G4 could go right now, to 2GHz, you think they would not do that? You think that they would not want to be seen as 2 or 3 times faster that Intel?
If they could, they would. I hope the G5 kills IA64 and I can't wait for the G5 TiBook!
the catalyst which brings the mass market to Linux.
I've been thinking along those lines for a while too. I used to hate virus writers, now I kinda think they'll ultimately be doing everyone a huge favour, since the softest and hardest to fix (fundamentally) targets are from Microsoft.
I'd like to see a Wintel virus that seeks out as many flashrom devices as it can recognise and write over them and all block devices it can with random data. Wipe out the motherboard, MODEM, HDD, Video card, CDROM/CDR/etc, printer, etc and perhaps leave behind a warning that this was all possible because of the long history of extremely poor Microsoft security (printed out on the default printer could be nice, or left in the users own mail account a few thousand times as a nice reminder once they get a new system up:).
That ought to wake some people up! Especially when at least one, possibly many critical parts of their systems are rendered useless until at least the firmwares are restored.
Perhaps the virus could ask the question, "Do you trust Microsoft with your IT requirements?", with a YES/NO prompt, YES leading to the above disaster and NO merely zeroing out C:.
I would have thought that any good cipher would be uncrackable when used with a OTP.
But getting back to my point with simple XOR'ing, XOR a OTP with raw audio. Now listen to the resulting file... you can still hear the original audio through the noise! Because on average, 50% of bits will have toggled, leaving enough recognisable data behind. Of course, do this with a text file and you're as secure as they come.
Simple XOR, even with a OTP, is not perfect for every application. It's not even practical for most.
"The simple-XOR algorithm is really an embarrassment; it's nothing more than a Vigenere polyalphabetic cipher. It's here only because of it's prevalence in commercial software packages, at least those in MS-DOS and Macintosh Worlds [1502.1387]. Unfortunetely, if a software security program proclaims that it has a "proprietary" encryption algorithm-significantly faster than DES-the odd are that it is some variant of this (XOR)".
If anyone thinks simple XOR is a magic bullet, they might want to try this...
XOR encrypt a raw audio file (no encoding, just a raw stream) with/dev/urandom and then listen to the encrypted file as an audio file...
You will hear the original audio file in all its glory along with perhaps 50% noise added!
XOR can be OK with perhaps a text file and a OTP, however any zeroes in the OTP will result in plain text at that byte in the output file. I'd preffer a OTP along with some other data mangling scheme.
The displays would have to be higher resolution than 65,536 pixels to warrant 24bit over 16bit. So greater than 320x200 pixels. Is this the case? I would highly doubt it.
My point being, that there is no point in being able to display more than 65,536 colours (simultaneously) if there are not that many pixels on screen. If there are only 64,000 pixels on screen, then obviously there are never going to be more than 64,000 colours displayed at any one time, which is usually what max colour specs are reffering to.
: )
Sorry to be so picky.
Probably a hold over from the fact that calculator watches were never cool
Hmmm, I'm imagining a watch version of the HP48GX... now that would be nerdy^2! The 5 pixel high time font might be a bit hard to read in the compact watch version though.
Moderators need to mod Charles UP and the AC down.
That's why they use ethernet for communications and just use the PCI bus for the power supply.
There is nothing informative about this! I was supporting products like these from Cubix back in '94. Back then, those products were even using the ISA and EISA buses to carry ethernet between other Cubix cards on those buses.
The PCI bus is just an outdated fancy parallel port.
ROFL. ISA can carry data at 8MByte/S (8bit 8MHz = 64Mbit/S) to 32MByte/S (16bit 16MHz = 256Mbit/S) which provided a far better solution in these setups than a 10base-* NIC that was going to be plugged into ISA or EISA *anyway*!
And PCI can carry data at 133MBytes/S (32bit 33.333MHz = 1Gbit/S)
These cards are usually slaves to the host motherboard, not the other way around. This way they're easier to make and the assumption can be that whatever they're plugged into will be the master of the PCI bus, so no need to fiddle with master/slave config. For usage with a dumb PCI backplane, PCI master cards (one per backplane please) can also be purchased. Though I have'nt looked at this companies offerings.
A Linux machine set up as a web server, accelerated with khttpd and one of these cards running FreeBSD serving the db would be an awesome setup. Nice and real fast. Especially with a server mobo with multiple PCI buses (buses != slots) to seperate the 100/1000Mb NIC interfaces and the db card.
Speaking of *BSD... If Linux continues to have a development VM system in the stable kernel, causing it to lock up hard, the people who came to Linux to get away from that kind of crap in the first place, (who are now more Unix aware) might move to something truely stable like FreeBSD.
Of course, according to Linus, there is nothing interesting in FreeBSD. ; ) Maybe he ought to steal Free's VM, or maybe Red Hat will do that for him.
Software RAID in 2.4 is crap and now the VM... more and more I want to move completely the FreeBSD and hopefully soon get some OSX happening too.
"I don't think this was related to anything I was doing as I wasn't actually doing a compile run at the time - probably just a random occurance, but worth mentioning."
What is this guy smoking!? A Ferrari might take the least amount of time to get somewhere, *IF* it ever gets there that is!
I've hard-rebooted my Linux machine a few times ever, and those few times were mostly since moving to 2.4.10+2.4.13.
Under high load, "the 2.2 kernel gave the best responsiveness".
Under low load, "2.2 again was the most responsive" AND WHILE IT "used the most swap ".
"Also, after one 2.4.13 test run, the system locked up entirely, requiring a hard reboot."
Another "random occurance" to be sure.
"2.4 VM systems are undeniably better". Say what? Better because it is quicker or better because it brings Linux in-line with Windows 95 stability?
Is it just me, or does this look like a guy trying to advertise himself and his company as smart guys who know Linux and want work, through some long winded test that claims that the VM in 2.4 is much better that 2.2 because it is a quicker, ignoring the glaring fact that 2.4 locked up hard for him more than once in his tests?
Actually, as you mention, it's when I close Acroreader that the trashing begins.
From what you say, what I've heard from others and from that article, it seems both VM systems are kinda sucky in the 2.4 kernels.
If FreeBSD supported RAID-0 with Vinum for / and/usr, I'd move completely over to 4.4 Release. If the Linux kernel does'nt get the VM and software RAID sorted out before the next major release, then I might upgrade my PII-300/BX using Linux software RAID-0 to maybe an AMD with hardware RAID-0 and move to FreeBSD that way. Software RAID-0 in Linux is also pretty sucky with 2.4.x, I get a reduction in speed over the single drives! : (
I have never experienced such "system freeze" swapping with Linux, since I started about 4 years ago. Felt like my Win95 days with 8Mb RAM were back.
With 256Mb RAM, I only need to open up a ~1Mb pdf with Acroreader to experience some extreme disk thrashing that takes an age to give back my system. Something I've never seen happen with Linux.
"...but now that I've been using Linux on those same systems for about 4 years..."
You haven't though. Linux uses differnent drivers for that hardware (as I'm sure you know).
I have'nt what? Been using Linux? Are you trying to say that Microsoft OSes crashing regularly versus Linux never crashing, is an indication that Microsoft crashes are due to poor MS 3rd party drivers?
Ever heard of MS certified drivers? Loads of crashes occur with them too. Also, if MS were more open with their specs and source code, we could expect more stable drivers and thus a more stable user experience.
The Linux drivers on my system are mostly from the Linux community, the WARP code for my various Matrox cards besides.
It's the driver in most cases causing the crash.
If my NIC crashes, ifconfig eth0 down;rmmod tulip;insmod tulip;ifconfig eth0 up, NIC back up and system still going.
Graphics card driver crashes, Ctrl-Alt-Backspace, then start X again. If that does'nt work, I can shh (telnet disabled) from another machine and kill X. System keeps going.
Same for sound... etc.
Attribute MS instability to what you like, but don't dare try to claim that MS is somehow not responsible for the instability of their products, driver issue or not. When fresh systems come from Dell or Gateway, with MS blessed everything, crashing here and there is who's fault?
Regardless of where the blame lies, I will continue to use the systems that have never crashed on me, with their mostly hacker written drivers. As opposed to shit OS code further hampered by shit driver code. MS warez are a dogs breakfast.
Yeah, SBM is great. I've been using it since late 1999 early 2000.
It has some great features, one in particular I like is the boot scheduling. You can configure it to automatically boot a particular partition based on the time of the day and get it to hide partitions also, depending on what partition you want booted.
What on Earth would you want with Boot Scheduling?... I supported a medium size.edu environment where students (who thought they were leet haxors) would like to stuff things up on the work stations. Combined with getting the BIOS to automatically switch the PC's ON late at night (when nobody is there), I could get SBM to automatically boot a hidden partition, which in turn runs Ghost to restore the SOE.
It was pretty cool. During the day, students could switch the PC ON, PC would boot straight to NT (NOT see any boot manager) and use it. But at 11pm that night, all the PC's would suddenly come ON, Ghost would come up automatically and restore the student workstation image to all the PC's, then about 3 minutes later (after the Ghosting has finished) the PC's would all power OFF.
My boss and his was pretty amazed at this. It allowed us to remove warez, mp3's, bad print jobs stuck in spools, viruses (mostly macro:), leet haxor settings and generally restore the machines to a state that is exactly the same as every other students, the teachers and all the other classes. Doing this manually on around 300 student PC's was a nightmare.
Another neato use for SBM's boot scheduling is getting all the PC's to automatically start some networked number crunching after business hours. Render farm, search for alien life or a cure for cancer anyone?
I'm pretty sure it messes up a LAN to have duplicate MAC addresses on it.
It does. Duped MAC addresses serviced by the same router are going to cause problems for those devices.
Also some places base their security on MAC addrs, probably not a good idea,
I took care of a site with DEC hubs that allowed you to use a particular MAC address to be a form of 'authentication' for a particular hub port. If someone else plugged their PC into that port, the hub would disable the port, assuming they did'nt change the MAC address on their card.
Oh, I have heard of vendors accidently duplicating MAC addresses (years ago).
I don't know about that 100 figure, but I have heard of some dodgy no-name cards from Taiwan that have MACs that are not unique.
Me thinks that they were no mistake. ; ) They just want to churn out $20 10/100 cards en mass, but don't have the vendor range of MAC's to do it. Or worse, any vendor range of MAC addresses at all.
No two of their cards have a duped MAC though. Their theory is, what are the chances of a customer purchasing a card from us with the same MAC of another device on their network. The MAC is 48bit so there are a fair few to choose from. ; )
Slashdot Mirror
How come these small things are always lacking in linux?
It is'nt. Try printtool. Red Hat's print setup GUI. I use Debian, but printtool is apt get'able. I set my Xerox DocuPrint P8ex up, using any driver that claims to do PCL (I used a Samsung printer driver) and from then on it just works perfectly. There are probably even easier to use tools, but why should I care? Printtool works for me, so why should I bother with something else?
With Linux (or *BSD for that matter), I can print to PDF from *any* app that can print to a file. Simply print to file and then use ghostscripts ps2pdf to convert the file to PDF.
Linux is not lacking, most users are. If you are someone who is not willing to RTFM after failing somewhere in a free OS, then go back to Mega$hit.
You need ram depending on what you do. For most laptop uses having more than 256 megs of ram is useless.
/lib is 28MB and /usr (apps, X, fonts and other libraries) is 1.3GB minus /usr/src/linux. Having as much of this, of which I use most often, cached (Thanks to loads of RAM) will improve performance thanks to less paging and less fetching of data that has already be fetched before.
I would agree that 256M would be way more than enough for something like an OpenBSD firewall, with it's super small footprint. However...
On my Debian machine, installed at 2.1 with an on-going upgrade to Debian-testing, my
Of course i use linux and not bloated windowsXP.
Of course I am impressed beyond belief. I see you are a Mandrake junkie. Congratulations on your successful install of Linux.
As far as procesor benchmarks, I checked them myself, the 1ghz processor totally smokes the G4 at specific things, I mean like x4 the speed of the G4 at certain things. Some things the G4 is closer to the same level, but these things are very rare things, like using a certain feature in photoshop.
As far as ram, Dell supports over 1gig of ram too.
As far as battery life, you win, Tibook beats dell.
I win? This is not a competition HanzoSan. How old are you? You obviously have little knowledge of CS, seemingly because you are young and closed minded. Trust me, if you want to compete here, although there are plenty of morons, there are also many people who do know CS very well, you will loose with this attitude until you open your mind. The smartest man, is the man who always questions himself and is not afraid of being wrong and adjusting his opinion accordingly. Here, ego will get you nowhere.
Having more ram wont give speed gains,
My desktop 7200rpm UDMA drives perform at about 25 sustained MegaBytes per second, my RAM "destroys" them (I have two, striped (RAID-0)) since my RAM sustains about 800MB/s. With the cost of RAM lately (with the exception of the recent artificial price spike), anyone would be CRAZY not to max out thier RAM. (That 25MB/s is not the stripe speed, but the raw single drive speed.)
Do you know what caching is? If so, what do you think of a 32:1 performance ratio? Would you rather be re-accessing apps and libraries from RAM@800MB/s or file-system/swap on disk@25MB/s?
To rub salt into wounds, my RAM might be considered slow by todays standards (PC100 SDRAM) and my disks way faster than any 2.5" notebook drive, so the ratio between disk:RAM would be much more pronounced. We could be talking about a 100 *times* speed difference or more. But even worse, I am comparing transfer rates here and not factoring in access times. Start using these disks with random accesses, with head movements that delay access to data in milli seconds (0.001s), versus nano seconds for RAM (0.000000001s), and you can see how cached data can effectively be thousands of times quicker (I know, 1nS is one million times quicker than 1mS, however, the entire time is not spent randomly accessing nothing).
having a faster harddrive will.
Of course it is a good idea to have a fast hard drive. But pouring money into your drives, and then skimping on RAM, is an ignorant move to say the least. BTW, the performance gains of loads of RAM, if often a performance area that benchmarks miss. Contrary to popular belief, benchmarks are acurate tools for measuring performance, but not overall performance. They are tools that measure specific areas of performance, that provide results that should be pondered and acted upon by someone who knows how to read and design for those numbers. If you do lots of 3D or even 2D CAD work, you will want fast floating point, 2D bitmap work will want fast CPU integer, loads of fast RAM and maybe good fpu also, video editing might need these and some real fast hdds or hdd array, a server might want loads of RAM, real fast networking hardware and a fast efficient networking OS.
Case in point for lots of RAM... Say my girlfriend is making a poster with Photoshop or The GIMP. It is going to be physically large'ish (typical poster), so high res is in order. Lets say she will be working with a bitmap of 4096x8192 32bit (8bits per RGB + 8bit alpha channel). The raw size of that, decompressed in memory, *without* layers, will be about 134MB. Now lets say shes using 5 layers for her various effects, that comes out to about 670MB being used *just* for the bitmap storage. Add to that OS and app memory usage, and consider whether you want to be doing this with RAM or disk. ; )
With something like this, whether you have a Xeon 2GHz or a Pentium 100, without enough RAM Photoshop is going to slow down at some stage to the speed of your hard disk. In fact, a Pentium 100 with 1GB RAM might "destroy" a Xeon 2GHz with 256MB RAM in this situation. So how important is your MHz now?
Get SCSI raid on a laptop and 1ghz chip will distance itself even more.
SCSI RAID on a laptop? Did you get some fine Jamaican hashish for Christmas or what! You think GHz is da bomb and SCSI RAID is an option for notebooks? Don't get me wrong, I would like a 2GHz G5 TiBook with 2GB RAM, a solid state 2.5" UDMA drive and OS X if I could have it, but I'd be quite happy with the current TiBook maxed out with RAM.
For now it still wins on encoding, decoding dvix and movies. It wins on gaming. It has a better screen, UXGA and if the guy above got the Latitude model which sucks its only because he didnt got the "network enhanced" model instead of the "desktop replacement" model. He had the choice.
I assure you, if you've got 256MB RAM and I've got 1GB RAM, I am going to enjoy my notebooks speed a whole lot more and a whole lot longer than you are. But I could'nt care less what you or anyone else is happy with, because I'm happy with how I set up my computers and thats what matters. This is not a dick measuring contest.
Looks like dell wins
For you, if you think you're the "winner", then good for you.
Just don't try to give me a lesson in CS.
Lastly, The 1 gig of ram thing, how as if you need 1 gig of ram ofr a laptop,
There are very compelling reasons to max out your RAM on a notebook computer:
1. 2.5" notebook drives tend to be slower than 3.5" desktop drives, so the RAM speed vs drive speed is much wider on a notebook than a desktop, meaning that any dependance on drive speed (say for swapping or re-getting something that could otherwise be cached), makes the notebook slower than the desktop. The speed gains of adding RAM are higher for notebooks than for PC's.
2. Those little 2.5" inch drives are expensive and have higher failure rates, they don't generally last as long as a desktop drive. With more RAM, there is less head movement due to caching, which can lead to longer drive life and...
3. less head movement = better battery session life.
I was sold on the i8000 until I saw the G4 TiBook. I am glad the TiBook can support up to 1GB RAM and when I get mine, that will be the first thing I upgrade it to.
and the powerbook doesnt come with 1 gig of ram either.
At least it is capable of supporting 1GB.
I will take more RAM over Mhz any day. I cringe when I see people complaining (at various work sites I attend) that their P3 750MHz Dell notebook is slow (and they demand an upgrade), when it only has 64MB RAM, so they get the latest machine which is only 30% quicker as far as CPU goes, yet has 256MB RAM, and they think the enormous speed gain was due to the quicker CPU. Blah. Of course, being executives, they're not interested in what I said about the cheap RAM upgrade, they want whats on the pretty web site.
the 2.5k dell totally destroys the TiBook in every area, better processor, better monitor, more ram, (Tibook comes with 512 megs of ram)
Destroys? The P3 1GHz is close to the G4 600MHz in the benchmarks I've seen. "Destroys" seems to be a school kid way of saying, "my PC is 15% quicker than yours!". I like the TiBook screen for what it is (wide screen), but I also like the i8000 screen and the TiBook is capable of supporting more RAM. "Destroy" is something a 2GHz Xeon does to a 4MHz 8080.
Prove you arent biased
You were'nt replying to me, though I can tell you, after 12 years with x86 (some of that repairing notebooks for NEC and DEC), am I much more impressed by what Apple is offering. They offer extreme stability and usability thanks to high quality hardware, limiting what hardware they support and their efforts of extending the super workhorse OS, BSD.
It's the package, the hardwareOS+app meld that works so well. After all these years putting up with x86, my next machine will be PPC. I will probably never buy another x86 again (besides SBC's I use for firewalls, etc, although I might look at PPC SBC's for them also).
From http://www.apple.com/macosx/technologies/security. html
Disk Image Encryption
For high security, you can encrypt part of your hard disk through the use of a disk image. You can then email this disk image can to other people who know the password. Simply open the Disk Copy utility, make a new image and set the encryption. The image will show up as a volume on your desktop. When your Keychain is locked, or when you send that disk image file to another person, the image is secure. When your Keychain is unlocked, you can copy, move and delete files as you would on any normal hard disk.
I know how to use the loop device to mount fs images, but I would really love to be able to mount an ISO9660 image as writable in Linux or FreeBSD, to perfect an image before burn. And being able to encrypt the image would be neato too, though I know that can be done.
So is there a way to mount an ISO image writable in Linux?
PS. Notice the screenshot at that link shows the recommended crypto to be AES-128! The more I read about Mac OSX the more I want it.
Perhaps a solution Apple could use, would be a new type of focus-follows-mouse, that does one of the following:
1. Only change focus to the window below the pointer if the mouse pointer becomes stationary for a set period of time over a new window (say 250mS). This way, continuous movement over other windows will not annoyingly change focus.
2. Or a slight variation to the above, only change focus to the window below the pointer if the pointer becomes stationary or mouse/keyboard is clicked.
3. What about a "warp portal"! ; ) An area at the top left of every window that will warp the mouse to the menu bar, with ESC or a warp-back area at the top-left of the screen warping the mouse back to the focused window?
OK, I better get some sleep now I guess. Although #1 sounds pretty slick to me.
If this is already an option, please excuse me, my next machine purchase will be PPC/OSX but until I can muster up the money, I have to be happy with hanging out at the local Apple Centre.
The problem is that it's true. I've tried twice to install Linux (PPC first, then intel), and failed both times. Linux sure isn't doing much for me on the desktop.
.deb to be faulty (in the official .iso). Notifying the Debian user community in a mailing list seemed to fall on deaf ears (weeks later it was "found" to be a faulty package), and 2.2 failed to install without install-script stopping errors on my AOpen AX6B. 2.2r2 even worse, kernel freeze at MCA drivers (yes I used the safe floppies to get it going)! 2.2r4 same as first, though I can get things sorted now. Progeny 1.0 has some flakey partitioning and can easily get it's knickers in a knot, requiring a reboot for another crack at it. Corel would not install unless I removed my Adaptec Ultra SCSI card. Caldera OpenLinux 2.2 progress bar getting to 700% and counting, never actually finishing the install. Red Hat partitioning also sucks and default install footprint and security is questionable.
/, /usr, /home and /tmp)!
I've been into Linux for about 4 years (I was curious seeing the kernel on Sydney BBS in 1993 but did'nt bother until Red Hat 5.0), MSx86 for about 12 years and various platforms from programmable calc/comps and various Apples to the Microbee and C64 (machine+basic) since about 1983.
So installing Linux or a BSD for me works almost always to some degree, even if I do commit hours to the task at times (grappling with hardware issues).
Having a bit of a love affair with Debian, Free&Open BSD (and enjoying QNX, Solaris and UnixWare), I'd love to dispute your point, but I can't.
Linux/BSD/Unix+tools are awesome for those who first know hardware very well and are willing to learn the ins and outs of the kernel and tools, but I've had some sort of troubles with almost every distro (barring SuSE, of which I've not used enough to complain of any problems), that I've been able to overcome with some effort that a newbie would have great difficulty with.
The various Linux distro's have had install issues that I can spot quickly and resolve, but what about the newbie? I realise that these issues mostly are due to the distro vendors, but tell that to the newbie who is viewing these experiences as his delve into "Line-Icks".
If I see erratic mouse movement in X, I think gpm, kill it and remove it from the rc's. X startup flashing with mostly black, XF86 config problem, Ctrl-Alt-Backspace, vi XF86Config, fixed. Many distros with really flakey partitioning utils I avoid prefering fdisk or cfdisk from a Ctrl-Alt-F*. Few distros have set up with X to a working GUI, leaving me to use xf86config, which mostly also fails to do so, leaving me to use vi to get it completed.
Hell, I tried Debian 2.2 (which I have now settled on (Debian-testing for home and Debian-current for work stuff)), and found the korn shell
I think the biggest hurdle for Linux acceptance is firstly the installer, and I must say, Mandrake and SuSE do it very well. Mandrake even allowed me to construct and install to software RAID (0 for
Don't get me wrong, *I* LOVE Linux and associated tools. I enjoy working at the command line and most work I do gets done there. I rip audio CD's with cdparanoia, data with dd, have many CD's permanently mounted (Unix, Perl, and Networking CD Bookshelves, SysAdmin6, etc) and served with http/ftp/samba using the loop device, make anything I like into pdf by printing as ps to file and converting with ghostscript, burn with cdrecord, do most downloading with wget or rsync, work from home with ssh+scp and even fax from the command prompt.
I can do most of this from a GUI frontend with some setting up, plus use office type apps (StarOffice6Beta is a killer!), 2D/3D and PCB CAD, Corel Draw, Sound editing, movie watching, etc etc, but Linux is not about to take over the desktop soon. I would'nt rule it out though at some stage.
But if OSX is *the* OS of the future, I will be happy. Since, Linux in any incarnation is still far from user-proof, whereas OSX is here, now. It's pretty, it works, it's easy, it's hyped on the mass media, it has MS IE and Office (important to a lot of desktop users), it's supported fully by Apple and it is the future of Apple at least. But, once I get my G4 PowerBook and OSX, I'll still use Linux, FreeBSD and OpenBSD, because I am an enthusiest. Less equiped newbies to Linux will continue to run back scared to Windows or find MacOS, until some Linux distro makes install easy and widely compatible, usage easy and worthwhile, and provide high quality documentation to ease the newbie into becoming a hardcore Unix guru at the pace they want, if they want.
Needing to be a hardware and Unix guru to be able install and use Linux is not going to be the road to desktop dominance. Please don't reply with, "I installed Blah-Linux without any troubles... therefore you're wrong...", unless you've installed it on every piece of hardware out there with the same ease, what works for one system can fail for many others.
I agree with most of what you say. However, there are technical reasons why Motorola does'nt just run the G4's into the 2GHz regions.
I can't recall the exact reason, but I think it has to do with fundamental design of the pipelines. Intel being designed to use and exploit as much MHz as possible, whereas the Motorola chips working smarter.
It's pretty much braun vs brains.
If the G4 could go right now, to 2GHz, you think they would not do that? You think that they would not want to be seen as 2 or 3 times faster that Intel?
If they could, they would. I hope the G5 kills IA64 and I can't wait for the G5 TiBook!
the catalyst which brings the mass market to Linux.
:).
I've been thinking along those lines for a while too. I used to hate virus writers, now I kinda think they'll ultimately be doing everyone a huge favour, since the softest and hardest to fix (fundamentally) targets are from Microsoft.
I'd like to see a Wintel virus that seeks out as many flashrom devices as it can recognise and write over them and all block devices it can with random data. Wipe out the motherboard, MODEM, HDD, Video card, CDROM/CDR/etc, printer, etc and perhaps leave behind a warning that this was all possible because of the long history of extremely poor Microsoft security (printed out on the default printer could be nice, or left in the users own mail account a few thousand times as a nice reminder once they get a new system up
That ought to wake some people up! Especially when at least one, possibly many critical parts of their systems are rendered useless until at least the firmwares are restored.
Perhaps the virus could ask the question, "Do you trust Microsoft with your IT requirements?", with a YES/NO prompt, YES leading to the above disaster and NO merely zeroing out C:.
Ahh, a man can dream can't he?
Sony Doggy Style (tm)?
Whoops, I should get more sleep I think.
First, I would imagine that a ADC would be built into the guitar so that the digitisation takes place right at the pickup.
Yeah, makes the old EMG (EMG?) pickups with the pre-amp in-the-pickup seem old!
Second, I seriously doubt that the DCMA applies in this case.
I think he was being funny. At least I thought so.
I imagine this could be extended a lot with digital effects built into the guitar! They'll probably sell add on cartridges to do so or something.
I have been wanting a watch that displays Unix Enoch time in decimal seconds.
Set to GMT, it would be a real internet time.
I would have thought that any good cipher would be uncrackable when used with a OTP.
But getting back to my point with simple XOR'ing, XOR a OTP with raw audio. Now listen to the resulting file... you can still hear the original audio through the noise! Because on average, 50% of bits will have toggled, leaving enough recognisable data behind. Of course, do this with a text file and you're as secure as they come.
Simple XOR, even with a OTP, is not perfect for every application. It's not even practical for most.
"The simple-XOR algorithm is really an embarrassment; it's nothing more than a Vigenere polyalphabetic cipher. It's here only because of it's prevalence in commercial software packages, at least those in MS-DOS and Macintosh Worlds [1502.1387]. Unfortunetely, if a software security program proclaims that it has a "proprietary" encryption algorithm-significantly faster than DES-the odd are that it is some variant of this (XOR)".
If anyone thinks simple XOR is a magic bullet, they might want to try this...
/dev/urandom and then listen to the encrypted file as an audio file...
XOR encrypt a raw audio file (no encoding, just a raw stream) with
You will hear the original audio file in all its glory along with perhaps 50% noise added!
XOR can be OK with perhaps a text file and a OTP, however any zeroes in the OTP will result in plain text at that byte in the output file. I'd preffer a OTP along with some other data mangling scheme.
Some even 24 bit color I think
The displays would have to be higher resolution than 65,536 pixels to warrant 24bit over 16bit. So greater than 320x200 pixels. Is this the case? I would highly doubt it.
My point being, that there is no point in being able to display more than 65,536 colours (simultaneously) if there are not that many pixels on screen. If there are only 64,000 pixels on screen, then obviously there are never going to be more than 64,000 colours displayed at any one time, which is usually what max colour specs are reffering to.
: )
Sorry to be so picky.
Probably a hold over from the fact that calculator watches were never cool
Hmmm, I'm imagining a watch version of the HP48GX... now that would be nerdy^2! The 5 pixel high time font might be a bit hard to read in the compact watch version though.
I use the HP48gx emulator x48 as my system calculator. It is excellent, but has been removed from Debian testing! (WTF!?!? WHY!)
i p Revision R ROM dump of the HP 48GX calculator.
http://www.hpcalc.org/hp48/pc/emulators/ HP48 emulators!
http://www.hpcalc.org/hp48/pc/emulators/gxrom-k.z
Enjoy!
Moderators need to mod Charles UP and the AC down.
That's why they use ethernet for communications and just use the PCI bus for the power supply.
There is nothing informative about this! I was supporting products like these from Cubix back in '94. Back then, those products were even using the ISA and EISA buses to carry ethernet between other Cubix cards on those buses.
The PCI bus is just an outdated fancy parallel port.
ROFL. ISA can carry data at 8MByte/S (8bit 8MHz = 64Mbit/S) to 32MByte/S (16bit 16MHz = 256Mbit/S) which provided a far better solution in these setups than a 10base-* NIC that was going to be plugged into ISA or EISA *anyway*!
And PCI can carry data at 133MBytes/S (32bit 33.333MHz = 1Gbit/S)
These cards are usually slaves to the host motherboard, not the other way around. This way they're easier to make and the assumption can be that whatever they're plugged into will be the master of the PCI bus, so no need to fiddle with master/slave config. For usage with a dumb PCI backplane, PCI master cards (one per backplane please) can also be purchased. Though I have'nt looked at this companies offerings.
A Linux machine set up as a web server, accelerated with khttpd and one of these cards running FreeBSD serving the db would be an awesome setup. Nice and real fast. Especially with a server mobo with multiple PCI buses (buses != slots) to seperate the 100/1000Mb NIC interfaces and the db card.
Speaking of *BSD... If Linux continues to have a development VM system in the stable kernel, causing it to lock up hard, the people who came to Linux to get away from that kind of crap in the first place, (who are now more Unix aware) might move to something truely stable like FreeBSD.
Of course, according to Linus, there is nothing interesting in FreeBSD. ; ) Maybe he ought to steal Free's VM, or maybe Red Hat will do that for him.
Software RAID in 2.4 is crap and now the VM... more and more I want to move completely the FreeBSD and hopefully soon get some OSX happening too.
"I don't think this was related to anything I was doing as I wasn't actually doing a compile run at the time - probably just a random occurance, but worth mentioning. "
What is this guy smoking!? A Ferrari might take the least amount of time to get somewhere, *IF* it ever gets there that is!
I've hard-rebooted my Linux machine a few times ever, and those few times were mostly since moving to 2.4.10+2.4.13.
Under high load, "the 2.2 kernel gave the best responsiveness".
Under low load, "2.2 again was the most responsive" AND WHILE IT "used the most swap ".
"Also, after one 2.4.13 test run, the system locked up entirely, requiring a hard reboot."
Another "random occurance" to be sure.
"2.4 VM systems are undeniably better". Say what? Better because it is quicker or better because it brings Linux in-line with Windows 95 stability?
Is it just me, or does this look like a guy trying to advertise himself and his company as smart guys who know Linux and want work, through some long winded test that claims that the VM in 2.4 is much better that 2.2 because it is a quicker, ignoring the glaring fact that 2.4 locked up hard for him more than once in his tests?
George?
/usr, I'd move completely over to 4.4 Release. If the Linux kernel does'nt get the VM and software RAID sorted out before the next major release, then I might upgrade my PII-300/BX using Linux software RAID-0 to maybe an AMD with hardware RAID-0 and move to FreeBSD that way. Software RAID-0 in Linux is also pretty sucky with 2.4.x, I get a reduction in speed over the single drives! : (
Actually, as you mention, it's when I close Acroreader that the trashing begins.
From what you say, what I've heard from others and from that article, it seems both VM systems are kinda sucky in the 2.4 kernels.
If FreeBSD supported RAID-0 with Vinum for / and
If he says so.
I have never experienced such "system freeze" swapping with Linux, since I started about 4 years ago. Felt like my Win95 days with 8Mb RAM were back.
With 256Mb RAM, I only need to open up a ~1Mb pdf with Acroreader to experience some extreme disk thrashing that takes an age to give back my system. Something I've never seen happen with Linux.
"...but now that I've been using Linux on those same systems for about 4 years..."
You haven't though. Linux uses differnent drivers for that hardware (as I'm sure you know).
I have'nt what? Been using Linux? Are you trying to say that Microsoft OSes crashing regularly versus Linux never crashing, is an indication that Microsoft crashes are due to poor MS 3rd party drivers?
Ever heard of MS certified drivers? Loads of crashes occur with them too. Also, if MS were more open with their specs and source code, we could expect more stable drivers and thus a more stable user experience.
The Linux drivers on my system are mostly from the Linux community, the WARP code for my various Matrox cards besides.
It's the driver in most cases causing the crash.
If my NIC crashes, ifconfig eth0 down;rmmod tulip;insmod tulip;ifconfig eth0 up, NIC back up and system still going.
Graphics card driver crashes, Ctrl-Alt-Backspace, then start X again. If that does'nt work, I can shh (telnet disabled) from another machine and kill X. System keeps going.
Same for sound... etc.
Attribute MS instability to what you like, but don't dare try to claim that MS is somehow not responsible for the instability of their products, driver issue or not. When fresh systems come from Dell or Gateway, with MS blessed everything, crashing here and there is who's fault?
Regardless of where the blame lies, I will continue to use the systems that have never crashed on me, with their mostly hacker written drivers. As opposed to shit OS code further hampered by shit driver code. MS warez are a dogs breakfast.
Yeah, SBM is great. I've been using it since late 1999 early 2000.
.edu environment where students (who thought they were leet haxors) would like to stuff things up on the work stations. Combined with getting the BIOS to automatically switch the PC's ON late at night (when nobody is there), I could get SBM to automatically boot a hidden partition, which in turn runs Ghost to restore the SOE.
It has some great features, one in particular I like is the boot scheduling. You can configure it to automatically boot a particular partition based on the time of the day and get it to hide partitions also, depending on what partition you want booted.
What on Earth would you want with Boot Scheduling?... I supported a medium size
It was pretty cool. During the day, students could switch the PC ON, PC would boot straight to NT (NOT see any boot manager) and use it. But at 11pm that night, all the PC's would suddenly come ON, Ghost would come up automatically and restore the student workstation image to all the PC's, then about 3 minutes later (after the Ghosting has finished) the PC's would all power OFF.
My boss and his was pretty amazed at this. It allowed us to remove warez, mp3's, bad print jobs stuck in spools, viruses (mostly macro:), leet haxor settings and generally restore the machines to a state that is exactly the same as every other students, the teachers and all the other classes. Doing this manually on around 300 student PC's was a nightmare.
Another neato use for SBM's boot scheduling is getting all the PC's to automatically start some networked number crunching after business hours. Render farm, search for alien life or a cure for cancer anyone?
I'm pretty sure it messes up a LAN to have duplicate MAC addresses on it.
It does. Duped MAC addresses serviced by the same router are going to cause problems for those devices.
Also some places base their security on MAC addrs, probably not a good idea,
I took care of a site with DEC hubs that allowed you to use a particular MAC address to be a form of 'authentication' for a particular hub port. If someone else plugged their PC into that port, the hub would disable the port, assuming they did'nt change the MAC address on their card.
Oh, I have heard of vendors accidently duplicating MAC addresses (years ago).
I don't know about that 100 figure, but I have heard of some dodgy no-name cards from Taiwan that have MACs that are not unique.
Me thinks that they were no mistake. ; ) They just want to churn out $20 10/100 cards en mass, but don't have the vendor range of MAC's to do it. Or worse, any vendor range of MAC addresses at all.
No two of their cards have a duped MAC though. Their theory is, what are the chances of a customer purchasing a card from us with the same MAC of another device on their network. The MAC is 48bit so there are a fair few to choose from. ; )