I'm really puzzled about the (Surprise) remark in the article.
The "(surprise)" remark is about Photoshop. I think the author of the article means that, since Photoshop is a very important app for Macs (they being the platform of choice for artists, at least traditionally), it is often expected that it should run better there than on Windows.
I don't know how true it is today that graphics professionnals prefer Macs overs Windows boxes, or whether Adobe et al put more effort in optimizing their products for either system, but it certainly seems odd that the G5, which the article finds to be faster than the Opteron in many cases, performs badly (relatively speaking) on an application as crucial as Photoshop.
I suppose it's good news that there are people who do care about Wifi security.
However, I'm wondering: how much security does SSID-based blocking add (could individuals forge SSIDs, or would they have to be organizations with cash and determination?)? Shouldn't all connections on a wireless network use a strong encoding (SSH or such)?
How do real people provide and use services that are normally insecure (NFS comes to mind) over Wifi?
Standard ABIs can be good; for example the fact that the Intel compilers and GCC use the same ABI for C++ on x86/Linux means that a program can successfully have parts compiled by either compiler. Although I can't think of an example right now, there are probably situations where it makes perfect sense to use several different compilers.
In the case discussed here, the absence of a (stable) ABI means that a module compiled against a given kernel version is not guaranteed to work with another kernel without recompiling.
It is interesting to note that Linus, although he clearly is hostile to proprietary, binary-only modules, said he didn't plan to make gratuitous ABI changes.
Also, it has been said that Linux doesn't provide an API either (the P is for "programming"), which means that even modules that come with their source code (be it complete or, as with nVidia, just "glue" between a binary and the kernel headers) should not expect to work with any future kernel without some changes. I don't think it would be wise to drastically change the API, but you never know...
They use an opensource wrapper just to comply with the GPL
I think the reason isn't GPL compliance (after all, for many "stock" kernels, they provide a ready-to-load binary module), but rather the fact that Linux doesn't provide an ABI. This is on purpose; so as to discourage closed-source drivers.
Sorry, I didn't know they used the metric system for minutes too! In the US of A, six minutes are only 360 seconds:-).
if you read the article rather than skimmed
Since their figures don't seem to be consistent, I had to discard some of them; 20km at 65kph should take about 18 minutes, not 6. Otherwise, you're of course right: they wouldn't do it if it weren't cheaper than the actual network connection they can get.
I don't know of any 1MBps (8 Mbps) ISP that is cheap outside of Korea...
Japan maybe? But indeed, I goofed; buying several megabit links may not be cheap in most countries (and is apparently impractical in the case explained in the article).
Wireless (802.11) links over 20km get a lot of attention on Slashdot but are very uncommon
What? Are you implying that what is said on Slashdot doesn't always closely reflect reality? Man, I'm disappointed!
Also, assuming fifty falcons, carrying three one-gig memory sticks each, we get up to a whopping 4 gigabits per second. I don't know if I'd let birds handle that many memory cards (they are expensive), though.
And don't get me started on the ping times, as others have mentioned already.
Nick Andreef's pigeons are faster than the internet, but no match for falcons.
Faster than the internet? Let's see:
- The picture shows a Memory Stick. That's at most 1GB (the blue MS like the one on the picture are at most 256MB, but let's be generous),
- the pigeons go at 65 kph,
- they have to travel over 20 km.
That's 20km/65kph = 1107 seconds. Which converts to slightly less than 942 KBps. Now, I don't know what kind of ISPs they have in New Zealand, but 1MBps shouldn't be that hard to achieve! Even using the announced 3GB capacity of a single bird, that's just 3MBps.
Maybe if they had used one of those wireless networks they wouldn't need to feed pigeons (and clean up after them)? Even the falcons go at a mere 10MBps bandwidth!
Put as many windows codec's under/usr/win32 and let mplayer have a shot at them. I've not seen one AVI that it couldnt play correctly.
Same here. Additionally, I'd mention that MPEGs should just work in every player there is. If you have MPEGs that don't play, they're most likely broken, and that's it.
You may want to use the file utility to see whether it things that your MPEGs are indeed MPEGs.
Now if they think I'm going to go click the link to confirm that my e-mail address exists, then they would surely be willing to buy some property on Mars I have for sale. Radiation-free. Really.
Or, more likely, if someone clicks their link, they can be added to the list of gullible people who can be expected to buy land on Mars.
While I don't see this as a problem, I can understand it comes as a schock to new users. Especially if they come from a "where do I click?" world.
But distributions usually allow one to install a package simply by clicking on its icon. One has to give the root password to complete the operations, but that's usually it.
*Missing packages and circular dependencies - a wont install without b, b wont install without c, c wont install without a.
Just install all dependent packages simultaneously... Although I must recognize I don't know how to do this other than with the command line.
*Maybe no.DLL hell but certainly library hell
I wonder how a non-programmer could really suffer from this (certainly programmers don't mind and likely prefer the command line), hm?
*Plug and play works ok during the initial setup but not very well after that. Try changing your video card after already doing the initial install.
True, I guess. One has to click the HardDrake thingie and ask for the new hardware to be recognized. Granted, it's automatic on Windows, but from time to time, Windows will also decide that you don't have a mouse, or that you have a new monitor.
*Cut and paste doesnt work most of the time.
I can't comment on this. I'm under the impression that copy-pasting is incredibly easy under Unix, since one doeasn't have to popup a menu or learn a keayboard short cut. Typically, whatever is selected goes to the X clipboard, and the second mouse button pastes. But then again, maybe I've been wrong all these years and it's true that cut and paste doesn't work.
*Crappy fonts - most web pages look like crap, even slashdot. I dont consider stealing fonts from windows a fix to this problem.
*X and Gnome and Kde are just as bloated as XP.
That's subjective, I suppose. Although I do use an old version of Gnome so as to enjoy short loading times.
*Everything is a beta. Wheres all the version 1.0+ software?
You mean the XP Extreme Edition software:-)? Well, anyway, in my experience, software marked "beta" is, for many projects, quite stable. Don't be afraid a priori to use pre-1.0 software; just see what the developers have to say about the stability of their current version. If developpers know their program isn't ready for prime-time, they will usually be honest about rather than risking making their users unhappy.
The idea of using a language like Java to build webpages seems interesting, at least because Java is designed to scale well on big programs. Perl at least scales just fine; I don't know about PHP though.
It's interesting to note that Linux (probably with Apache's Tomcat) is a very popular system for JavaServer Pages (where one would have expected Sun's own Solaris platform to be popular). There is an article at Netcraft (from July) that summarizes the situation:
JSP continues to enjoy fast growth with a 94% increase in ip addresses running JSP based sites to over 44,000 ip addresses running some 105,000 active sites.
More surprising is the composition of these sites choice of operating systems. One might expect that by far the most common operating system amongst JSP based sites would be Solaris, (...). However, Solaris is only placed 3rd with 17% behind Linux with 40% and Windows with 26%.
And if you wanted hardware accel on one of the currently-supported platforms, but Nvidia didn't release any drivers at all, you'd still be on your own.
Absolutely correct (last time I tried, the free "nv" drivers didn't work at all with with my current graphics card). I don't think we exactly want nVidia to simply give us the source to their drivers, but would rather like them to release information about how to interface with their hardware.
It seems that cards manufaturers feel that it is important to keep their hardware/kernel interfaces secret. I'm not sure how much competitors could hope to gain from knowing them, though.
I understand where you're coming from, but I'm firmly in the "I just want my expensive hardware to work properly" camp.
To clarify the situation: I'm very happy with the nVidia drivers, but I understand that they don't make the life of kernel maintainers any easier.
One of the main 'comments' I get when I recommend Debian GNU/Linux to people, is 'Debian is difficult to install'
I think it can be argued that the Debian installer asks many questions that may not be easy to answer for a Linux newbie.
But, as you say, there is hope: I remember someone saying, a few years ago, that a RedHat had formatted their drives without clearly mentioning that it would be destructive (oops!). Today, Mandrake can be installed after just a few minutes worth of clicking "OK". It generally makes the right choices for the user, clearly shows what partitions will be created, and warns if it's about to blank an existing windows partition. If it finds some unsupported hardware, it mentions what it knows about it, so that the user can simply ask their local guru for help.
I think it's no exaggeration to say that someone who already installed Windows can safely install e.g. a Mandrake.
I'm not a coder. (...) I get no benefits from having the code available.
Actually, I think you would get benefits from having the source code.
For example, it doesn't seem like nVidia is in a hurry to release drivers for new kernel versions, and just renaming their current driver doesn't work in Linux 2.6. The day you want to move to 2.6 (and you may have reasons; for example if it brings support for a new device you buy), you'll find that you can't just go to the manufacturer's site and download a binary.
(Note: it is possible to use the current nVidia drivers with 2.6, but with some patching.)
No AGP slot on RISC computers? nVidia's cards come with some Apple machines, but nVidia doesn't provide any driver for Linux/PPC.
Drivers for Linux (on 386, Intel-64, AMD-64) and FreeBSD/386 are available, but if you want to get hardware acceleration on another architecture, you're on your own.
So far, no system has been able to match faces better than 60-80% in real life tests.
That's interesting. I seem to remember that biometrics were used to control access to certain sites of the Atlanta Olympics back in 1996.
They scanned hands (since they didn't want people walking around the city with stolen eyes in a plastic bag or at the top of an ice-pick... but I digress), at the time.
I wonder if this was a major failure? Or did they somehow get a "sufficient" success rate?
Warning: mysql_connect() [function.mysql-connect]: Too many connections in/usr/local/httpd/htdocs/kunden/web9/html/_dx/_dx_s ettings.php on line 9 Eine Verbindung zum Datenbankserver ist zur Zeit nicht moglich...
<clueless> If think this is the ultimate proof that MySQL can't be used for high-traffic websites:-) </clueless>
i will get a 2.6.xx kernel when most the major distros start packageing them with their distros
Well, Linus recommends that distros start offering 2.6.0 kernels as an installation option. It may happen quite quickly (I remember having the option to use either a rock-solid 2.2 or a "not tested as thoroughly" 2.4) some time ago.
There is an nForce configuration option at least in 2.6.0-test10, but, according to discussions that took place this week, there are stability problems. So I guess that yes, this board is going to be supported somehow.
It is interesting to note that Linus doesn't like nVidia releasing binary-only, proprietary drivers, and thus doesn't plan to make it easy for graphics card makers to distribute modules in binary form. To quote him: It's a two-way street: if you don't help me, I don't help you.
So, I don't know whether the nForce documentation comes with hidden "trade secrets" or such, but if nVidia's attitude is the same as for graphic cards, it may not motivate developpers to support their products.
I'll be modded down as flamebait or offtopic, when this is stuff that matters (...) Still stick there at 1.25GHz on your G4s apple? Tsk.
May not be that much off-topic, actually. It's blatantly obvious nowadays that clock frequency isn't closely linked to performance, especially when comparing different architectures like PPC and 386.
I don't think it would be nonsensical to run a benchmark comparing PPC, Intel x86 and AMD x86: if you read a few of the articles at Ars Technica, you will see how incredibly complex the P4 is. In these conditions, it wouldn't be surprising that Intel's chip is as different from PPC as from AMD's chips.
Slashdot Mirror
I'm really puzzled about the (Surprise) remark in the article.
The "(surprise)" remark is about Photoshop. I think the author of the article means that, since Photoshop is a very important app for Macs (they being the platform of choice for artists, at least traditionally), it is often expected that it should run better there than on Windows.
I don't know how true it is today that graphics professionnals prefer Macs overs Windows boxes, or whether Adobe et al put more effort in optimizing their products for either system, but it certainly seems odd that the G5, which the article finds to be faster than the Opteron in many cases, performs badly (relatively speaking) on an application as crucial as Photoshop.
I suppose it's good news that there are people who do care about Wifi security.
However, I'm wondering: how much security does SSID-based blocking add (could individuals forge SSIDs, or would they have to be organizations with cash and determination?)? Shouldn't all connections on a wireless network use a strong encoding (SSH or such)?
How do real people provide and use services that are normally insecure (NFS comes to mind) over Wifi?
ABI: application binary interface.
Standard ABIs can be good; for example the fact that the Intel compilers and GCC use the same ABI for C++ on x86/Linux means that a program can successfully have parts compiled by either compiler. Although I can't think of an example right now, there are probably situations where it makes perfect sense to use several different compilers.
In the case discussed here, the absence of a (stable) ABI means that a module compiled against a given kernel version is not guaranteed to work with another kernel without recompiling.
It is interesting to note that Linus, although he clearly is hostile to proprietary, binary-only modules, said he didn't plan to make gratuitous ABI changes.
Also, it has been said that Linux doesn't provide an API either (the P is for "programming"), which means that even modules that come with their source code (be it complete or, as with nVidia, just "glue" between a binary and the kernel headers) should not expect to work with any future kernel without some changes. I don't think it would be wise to drastically change the API, but you never know...
They use an opensource wrapper just to comply with the GPL
I think the reason isn't GPL compliance (after all, for many "stock" kernels, they provide a ready-to-load binary module), but rather the fact that Linux doesn't provide an ABI. This is on purpose; so as to discourage closed-source drivers.
Among the systems that his laptop does not run:
SCO
SCO seems to be everybody's favorite company these days. (...)
Absolutely true! Just see how often SCO makes it to the frontpage of Slashdot...
six minues (600 seconds)
:-).
Sorry, I didn't know they used the metric system for minutes too! In the US of A, six minutes are only 360 seconds
if you read the article rather than skimmed
Since their figures don't seem to be consistent, I had to discard some of them; 20km at 65kph should take about 18 minutes, not 6. Otherwise, you're of course right: they wouldn't do it if it weren't cheaper than the actual network connection they can get.
I don't know of any 1MBps (8 Mbps) ISP that is cheap outside of Korea...
Japan maybe? But indeed, I goofed; buying several megabit links may not be cheap in most countries (and is apparently impractical in the case explained in the article).
Wireless (802.11) links over 20km get a lot of attention on Slashdot but are very uncommon
What? Are you implying that what is said on Slashdot doesn't always closely reflect reality? Man, I'm disappointed!
Also, assuming fifty falcons, carrying three one-gig memory sticks each, we get up to a whopping 4 gigabits per second. I don't know if I'd let birds handle that many memory cards (they are expensive), though.
And don't get me started on the ping times, as others have mentioned already.
Nick Andreef's pigeons are faster than the internet, but no match for falcons.
Faster than the internet? Let's see:
- The picture shows a Memory Stick. That's at most 1GB (the blue MS like the one on the picture are at most 256MB, but let's be generous),
- the pigeons go at 65 kph,
- they have to travel over 20 km.
That's 20km/65kph = 1107 seconds. Which converts to slightly less than 942 KBps. Now, I don't know what kind of ISPs they have in New Zealand, but 1MBps shouldn't be that hard to achieve! Even using the announced 3GB capacity of a single bird, that's just 3MBps.
Maybe if they had used one of those wireless networks they wouldn't need to feed pigeons (and clean up after them)? Even the falcons go at a mere 10MBps bandwidth!
Definitely not the same if you ask me
How comes? Do Sun deliberately package their x86 version differently? If so, what's different?
> > SMP (symbol manipulation program) architecture. :-)
> A case of acronym confusion, I guess.
Indeed. Surely they meant "Service Mediation Platform" or possibly "Sex, Money, Power"...
Put as many windows codec's under /usr/win32 and let mplayer have a shot at them. I've not seen one AVI that it couldnt play correctly.
Same here. Additionally, I'd mention that MPEGs should just work in every player there is. If you have MPEGs that don't play, they're most likely broken, and that's it.
You may want to use the file utility to see whether it things that your MPEGs are indeed MPEGs.
Now if they think I'm going to go click the link to confirm that my e-mail address exists, then they would surely be willing to buy some property on Mars I have for sale. Radiation-free. Really.
Or, more likely, if someone clicks their link, they can be added to the list of gullible people who can be expected to buy land on Mars.
*Installing new apps from the command line
.DLL hell but certainly library hell
:-)? Well, anyway, in my experience, software marked "beta" is, for many projects, quite stable. Don't be afraid a priori to use pre-1.0 software; just see what the developers have to say about the stability of their current version. If developpers know their program isn't ready for prime-time, they will usually be honest about rather than risking making their users unhappy.
While I don't see this as a problem, I can understand it comes as a schock to new users. Especially if they come from a "where do I click?" world.
But distributions usually allow one to install a package simply by clicking on its icon. One has to give the root password to complete the operations, but that's usually it.
*Missing packages and circular dependencies - a wont install without b, b wont install without c, c wont install without a.
Just install all dependent packages simultaneously... Although I must recognize I don't know how to do this other than with the command line.
*Maybe no
I wonder how a non-programmer could really suffer from this (certainly programmers don't mind and likely prefer the command line), hm?
*Plug and play works ok during the initial setup but not very well after that. Try changing your video card after already doing the initial install.
True, I guess. One has to click the HardDrake thingie and ask for the new hardware to be recognized. Granted, it's automatic on Windows, but from time to time, Windows will also decide that you don't have a mouse, or that you have a new monitor.
*Cut and paste doesnt work most of the time.
I can't comment on this. I'm under the impression that copy-pasting is incredibly easy under Unix, since one doeasn't have to popup a menu or learn a keayboard short cut. Typically, whatever is selected goes to the X clipboard, and the second mouse button pastes. But then again, maybe I've been wrong all these years and it's true that cut and paste doesn't work.
*Crappy fonts - most web pages look like crap, even slashdot. I dont consider stealing fonts from windows a fix to this problem.
*X and Gnome and Kde are just as bloated as XP.
That's subjective, I suppose. Although I do use an old version of Gnome so as to enjoy short loading times.
*Everything is a beta. Wheres all the version 1.0+ software?
You mean the XP Extreme Edition software
The idea of using a language like Java to build webpages seems interesting, at least because Java is designed to scale well on big programs. Perl at least scales just fine; I don't know about PHP though.
It's interesting to note that Linux (probably with Apache's Tomcat) is a very popular system for JavaServer Pages (where one would have expected Sun's own Solaris platform to be popular). There is an article at Netcraft (from July) that summarizes the situation:
JSP continues to enjoy fast growth with a 94% increase in ip addresses running JSP based sites to over 44,000 ip addresses running some 105,000 active sites.
More surprising is the composition of these sites choice of operating systems. One might expect that by far the most common operating system amongst JSP based sites would be Solaris, (...). However, Solaris is only placed 3rd with 17% behind Linux with 40% and Windows with 26%.
And if you wanted hardware accel on one of the currently-supported platforms, but Nvidia didn't release any drivers at all, you'd still be on your own.
Absolutely correct (last time I tried, the free "nv" drivers didn't work at all with with my current graphics card). I don't think we exactly want nVidia to simply give us the source to their drivers, but would rather like them to release information about how to interface with their hardware.
It seems that cards manufaturers feel that it is important to keep their hardware/kernel interfaces secret. I'm not sure how much competitors could hope to gain from knowing them, though.
I understand where you're coming from, but I'm firmly in the "I just want my expensive hardware to work properly" camp.
To clarify the situation: I'm very happy with the nVidia drivers, but I understand that they don't make the life of kernel maintainers any easier.
One of the main 'comments' I get when I recommend Debian GNU/Linux to people, is 'Debian is difficult to install'
I think it can be argued that the Debian installer asks many questions that may not be easy to answer for a Linux newbie.
But, as you say, there is hope: I remember someone saying, a few years ago, that a RedHat had formatted their drives without clearly mentioning that it would be destructive (oops!). Today, Mandrake can be installed after just a few minutes worth of clicking "OK". It generally makes the right choices for the user, clearly shows what partitions will be created, and warns if it's about to blank an existing windows partition. If it finds some unsupported hardware, it mentions what it knows about it, so that the user can simply ask their local guru for help.
I think it's no exaggeration to say that someone who already installed Windows can safely install e.g. a Mandrake.
I'm not a coder. (...) I get no benefits from having the code available.
Actually, I think you would get benefits from having the source code.
For example, it doesn't seem like nVidia is in a hurry to release drivers for new kernel versions, and just renaming their current driver doesn't work in Linux 2.6. The day you want to move to 2.6 (and you may have reasons; for example if it brings support for a new device you buy), you'll find that you can't just go to the manufacturer's site and download a binary.
(Note: it is possible to use the current nVidia drivers with 2.6, but with some patching.)
No AGP slot on RISC computers? nVidia's cards come with some Apple machines, but nVidia doesn't provide any driver for Linux/PPC.
Drivers for Linux (on 386, Intel-64, AMD-64) and FreeBSD/386 are available, but if you want to get hardware acceleration on another architecture, you're on your own.
So far, no system has been able to match faces better than 60-80% in real life tests.
That's interesting. I seem to remember that biometrics were used to control access to certain sites of the Atlanta Olympics back in 1996.
They scanned hands (since they didn't want people walking around the city with stolen eyes in a plastic bag or at the top of an ice-pick... but I digress), at the time.
I wonder if this was a major failure? Or did they somehow get a "sufficient" success rate?
Warning: mysql_connect() [function.mysql-connect]: Too many connections in /usr/local/httpd/htdocs/kunden/web9/html/_dx/_dx_s ettings.php on line 9 ...
:-)
Eine Verbindung zum Datenbankserver ist zur Zeit nicht moglich
<clueless>
If think this is the ultimate proof that MySQL can't be used for high-traffic websites
</clueless>
I don't know this is meant to be humorous... so I'll assume it isn't.
It's a "well-known fact" that 2.4 wasn't completely ready until 2.4.10 or 2.4.13.
i will get a 2.6.xx kernel when most the major distros start packageing them with their distros
Well, Linus recommends that distros start offering 2.6.0 kernels as an installation option. It may happen quite quickly (I remember having the option to use either a rock-solid 2.2 or a "not tested as thoroughly" 2.4) some time ago.
There is an nForce configuration option at least in 2.6.0-test10, but, according to discussions that took place this week, there are stability problems. So I guess that yes, this board is going to be supported somehow.
It is interesting to note that Linus doesn't like nVidia releasing binary-only, proprietary drivers, and thus doesn't plan to make it easy for graphics card makers to distribute modules in binary form. To quote him: It's a two-way street: if you don't help me, I don't help you.
So, I don't know whether the nForce documentation comes with hidden "trade secrets" or such, but if nVidia's attitude is the same as for graphic cards, it may not motivate developpers to support their products.
Sorry, but we cannot sell this cell phone jammer to UK customers
:)
Apparently, it's not very legal in the UK either
I'll be modded down as flamebait or offtopic, when this is stuff that matters (...) Still stick there at 1.25GHz on your G4s apple? Tsk.
May not be that much off-topic, actually. It's blatantly obvious nowadays that clock frequency isn't closely linked to performance, especially when comparing different architectures like PPC and 386.
I don't think it would be nonsensical to run a benchmark comparing PPC, Intel x86 and AMD x86: if you read a few of the articles at Ars Technica, you will see how incredibly complex the P4 is. In these conditions, it wouldn't be surprising that Intel's chip is as different from PPC as from AMD's chips.