Re:If they can do it to Google, they can do it to
on
The Case For Oracle
·
· Score: 1
I heard that Sun used to charge royaltees for mobile phones that licensed the J2ME platform - complaint or certified implementations. And this kinda proves this:
From http://java.sun.com/j2se/embedded/faq.html#faq10:
Q: Does deploying Java SE for embedded devices or purposes require a royalty?
A: Yes. Sun's license for Java SE enables it to be freely used for general purpose desktops or servers. If Java SE is bundled as part of a dedicated solution that involves or controls hardware of some kind, then it's likely an embedded application and is subject to modest royalty payments.
I believe Google java implementation is custom - not part of SE or ME - so they avoid having pay any royaltees to Oracle. This is why I think Oracle is pissed.
The audio layer (audiodg.exe) was ripped from kernel and moved to user space in Vista. This required audio drivers to be re-written and in my case my Creative Audigy sucks a lot of CPU because the driver guys had bugs with incorrect usage of buffers.
This means playing sounds sucks 30-40% of my 1.2GhZ CPU! Playing videos with sound suck 100% CPU on Vista!
When I moved back to XP, playing sounds suck 5% of CPU and Playing videos/w sound suck 25-30% of CPU! Now what the fuck is vista doing? And who to blame? Microsoft or Creative?
Exactly -- they are just APIs. Fan-boys get into stupid flame-wars about what API is better. Real guys would talk about implementation of pthreads under linux vs winthreads. Heck, the linux NPTL implementation got reader/writer locks, spinlocks, and a lot better performance compared to the sucky LinuxThreads. For example, grabbing a lock on a mutex is just a couple of instructions (if no one is holding the lock), and thread context switch time is 1us.
But if ppl just resort to comparing APIs, it like saying "mine is bigger than yours":)
You had no issues because a) performance was good enough and b) the rate of incoming client requests was relatively small compared to a loaded webserver.
Now if you had thousands of client requests-per-second, the fork will show why-you-should-not-use-it-in-such-situations. In such cases you can use a pool of forked processes or simply write the whole damn thing using threads. For example, apache gives you the ability to change 'worker' modules...and you can experiment with that to get an idea of all these request processors.
btw, threads don't complicate your code as long as you minimize data-sharing between threads and write the thread the same way as you were writing a forked process, except that you put all global variables within the thread's local data area.
Kashif
Good point, but I find brainstorming at a meeting to be far more effective when all peers are at the same level of understanding. In other words, having a meeting with just engineers or just directors is better than a mix of those groups, simply because the two groups don't collide on objectives (i.e. problems/solutions vs time-to-ship).
I mean if you're at a meeting where the boss says, I want this done by next month or you're fired really restricts your options of creativity. Then it's simply 'fuck it, whatever works'.
"It would be *really* nice if Dell would do some basic work to document device-driver compatibility for their systems. So if I was configuring a Linux system online, I'd like to see something like:"
Do you know how HARD that is to do? Essentially Dell would have to QUALIFY every permutation of kernel (2.6.17,.18,.19), open source drivers, closed sourced drivers, etc. It's not just figuring out which "drivers work", but also verifying they work through test cycles. The last thing anyone wants is some driver that 'works', but under some stress or certain combination of I/Os causes hangs/stability problems.
Then your guy working 10-hours a day can do NOTHING, because of some blocker bug and has to wait from either the a) open source community, b) project maintainers, c) consultant, d) fix it himself, or e) find commercial vendor.
It's MUCH easier if Dell partners with a distro vendor and get customized drivers/patches needed to have the distro run on their boxes.
"In general, OS X packages have to bundle libraries that are not included by default in the OS. This leads to anywhere from a little to a lot of wasted disk space and memory."
First point, lots of people of have plenty of disk space...you can buy 500G hd under $300. Second, the total amount of disk space used by all of your applications (OS, msoffice, itunes, etc) is going to be 10G. Most of the space will be eaten up by your endless archive of pictures, movies, music, and porn. My windows OS, itunes, msn messenger, yahoo messenger, firefox, msoffice, nero, azureus + another 10 or so small shitty programs take up a mesealy 5.40GB of space.
Finally how many programs do you have that ship with custom libraries? And even if they ship with 3rd party libraries, as long as all other applications ship with the same library, the OS should load only a single instance of that library.
The Linux vserver sounds very similar to UserModeLinux in both capabilities and features. Both run linux with linux, but UML has been included within the kernel. Care to describe why vserver is better/different than UML?
None of the concepts of computer science are new, but what is ground breaking is Google touching all aspects of computer science to solve a problem. Distributed Databases, Replicated Filesystems, Clustering, Learning algorithms, job scheduling, map/reduce languages, etc. are not new. But they applied each of these sub-domains to 'searching' and 'lots of data'.
Using old ideas is _new_ ways is ground breaking. That what everyone does(like Carmack and DOOM3).
If that's so, then if 1000 of your computers get infected, then that $50,000. So for us small guys, it's nothing, but for the big guys, they can atleast hire more admins just to handle the fucking spyware crap.
I had a Windows XP Professional original CD, and had lost the paper-back insert with CD-key. And I needed to format my windows. Of course, I could have gone to kazaa and tried to find a crack -- but that takes just as long as phoning.
So this is how the process went: - I asked them I lost my insert, but had the original CD - They said fine. But then they need to verify I had the original CD, so they made me describe how the CD looked in detail - I told them it had holographic pictures, it was shiny, it has "unauthorized copies prohibited", I described all the little writing around the CD and the back, it says Microsoft here and there, what it says on the concentric rings, etc. - They said OK. And proceeded to give me a brand-new CD key and scribbled it down - Now they said, "we have to wait for you to install Windows XP upto the point where it asks you to input the key". - I said OK, "but it will take a while" - They said fine, "we can wait for a few minutes" - So I proceeded to install Windows XP, and finally after 20 or so minutes, it asked me for the key. I put the key in, and it was accepted. - I told them CD key was accepted, and we did our goodbyes and hanged up.
Two things that surprised me from this ordeal: 1) from what I heard, I didn't know you had to describe the CD so they knew you didn't have a burnt copy 2) They actually waited for me to install Windows XP and try out the key. I could have just lied and said, "yeah it worked", but If I recorded the key incorrectly, I would have been an idiot calling again saying, "it worked, but then it didn't". They would say, "WTF?"
Yes, it is asking for permission. Also making it clear to the user what the software is and that it is _not_ spyware(when in fact it is adware). This protects the anti-spyware authors ass from litigation, but also allowing user to delete based on information available about program.
So in a nutshell anti-spyware should rename themselves to "RemoveCrap Now! 2.0"
"I've found that Netflix "throttles" my rentals after a period when I rent too many movies for them to make a profit."
If you rent too-many movies(like 10 or 20), the postage cost should be incremental for them. Assuming they have the same agreement as Blockbuster with their movie suppliers(i.e. get a cut of the profits, no cost on DVD copies). After all each DVD weighs very little, and to copy them is less than a dollar.
The only strain I see is management-wise and title availability. I.e. one guy starving everyone else of rentals.
Which is why Anti-Spyware software have to rename themselves, or put something in bold letters when something found isn't spy-based software. Something desclaimer like:
WARNING: 'ChangeYourToolBar.exe' isn't spyware. It is a program that changes your toolbar with its own custom version, without asking your permission. If you think this is an illegitimate program, please click on the check box to delete.
Now this removes the burden-of-proof from anti-spyware programs, and lets the user decide whether he should keep it or not based on the description.
Just today I was thinking about how google and other search engines should provide you with information on a particular domain. For example, if your searching for "lastest windows patches" you get a result "windowsupdate.microsoft.com". Now while the site is very relevant for what I am _looking_ for, it doesn't _directly_ give me the results I want. I have to go thru various links and reading just to get the list of patches to download.
But if google was smart, it would actually give me the list of the "lastest" patch files for all versions of windows. THAT is what I want. How you do that without using meta-data is another subject all together. A simple algorithm will be:
- return list of all relevant web pages with "windows patches" - generate a list of objects on these pages - find all objects that are patch files and ignore all other objects(.jpg,.gif, etc) - return list of these patch files - perform "latest" operation on result and output final result in HTML
The hardest part will be identifying objects that are patches, but searching thru relevant pages you _hope_ to get the objects needed.
Google does something similar to this with images.
SELinux -- even if apache has holes, linux is _more_ securable. You can run apache with fine-grained mode telling what it can or cannot do. So even if someone finds a buffer-overrun vunerability, the most they can do is whatever apache is _allowed_ to do.
The point with Linux is: the choice is there to make it MORE secure. Windows doesn't even have any type of MAC security. And last-time I checked, SELinux was supported for Redhat Linux ES.
And those who want to give SELinux a run can try it out on Fedora Core 3, before shelling out for RHES.
This is like releasing a disease out in the wild and selling the disease in the market. It causes disease to get maximum exposure.
Of course, they only give you the antidote if you acually bought the disease in the market. If you want the antidote, buy the legitimized version of the disease.
Which means all bastardized versions of the disease will have to legitimize themselves or be denied the Antidote.
So why not call your new distro, "RedHat-compatible". Meaning, redhat certified software can run on your new distro.
In one case, I can see RedHat being concerned with "here is the Free version of RedHat Linux!", they don't want some free-product thinking it is _the_ redhat linux product.
On the otherhand, one should be able to mention that this new distro is a fork of RedHat AS 3.x. It is redhat-compatible. Also if RedHat is releasing a GPL software(i.e. kernel) you should be able to say kernel-2.4-3.2as by RedHat. Meaning you have to give credit to the authors. But according to their terms you can't say "contains Red Hat Enterprise Linux X.X.".
To confuse you even more, there is a Windows Messenger and MSN Messenger. When you install MSN Messenger it simply disables start-up of Windows Messenger.
So when you uninstall MSN Messenger, it may be automatically enabled Windows Messenger. I don't know if that's what you are seeing.
My father used to work for Inglis - an appliance maker. One of the problems Inglis saw was that people weren't buying new appliances as the old "versions" were working perfectly fine for the past 10-20 years. Why buy a new one if the old one ain't broke?
Now Inglis manufactures things like washing machines such as one part inside is _designed_ to fail _after_ the warranty period and the whole product is only designed to last for at most 5 years. Also the whole cost cutting approach(trying to use plastic, cheaper parts, etc) is making appliances crappier day by day.
What I'm trying to say, hardware makers design their machines to fail after x amount of time.
And this is evident with games. Heck, everyone knows that games are pushing hardware more than anything else. There's no way a game like Doom3 will run faster on a older hardware. Which thus proves, games are using every cycle possible otherwise games would run at 10 frames-per-second.
Why not just emulate IE's broken behaviour(with active X of course). Sure it would be 'non-standard' but there will be a button, but we're all pragmatic bastards and what things to just work.
Simple way: Implement a new checkbox, "Use IE Emulation for this web-site" and save resultant web-site to list. Now I don't know how hard this will be i.e. you may need two different HTML renderers.
Kashif
Slashdot Mirror
I heard that Sun used to charge royaltees for mobile phones that licensed the J2ME platform - complaint or certified implementations. And this kinda proves this: From http://java.sun.com/j2se/embedded/faq.html#faq10: Q: Does deploying Java SE for embedded devices or purposes require a royalty? A: Yes. Sun's license for Java SE enables it to be freely used for general purpose desktops or servers. If Java SE is bundled as part of a dedicated solution that involves or controls hardware of some kind, then it's likely an embedded application and is subject to modest royalty payments. I believe Google java implementation is custom - not part of SE or ME - so they avoid having pay any royaltees to Oracle. This is why I think Oracle is pissed.
The audio layer (audiodg.exe) was ripped from kernel and moved to user space in Vista. This required audio drivers to be re-written and in my case my Creative Audigy sucks a lot of CPU because the driver guys had bugs with incorrect usage of buffers.
/w sound suck 25-30% of CPU! Now what the fuck is vista doing? And who to blame? Microsoft or Creative?
This means playing sounds sucks 30-40% of my 1.2GhZ CPU! Playing videos with sound suck 100% CPU on Vista!
When I moved back to XP, playing sounds suck 5% of CPU and Playing videos
Kashif
Exactly -- they are just APIs. Fan-boys get into stupid flame-wars about what API is better. Real guys would talk about implementation of pthreads under linux vs winthreads. Heck, the linux NPTL implementation got reader/writer locks, spinlocks, and a lot better performance compared to the sucky LinuxThreads. For example, grabbing a lock on a mutex is just a couple of instructions (if no one is holding the lock), and thread context switch time is 1us.
:)
But if ppl just resort to comparing APIs, it like saying "mine is bigger than yours"
kashif
You had no issues because a) performance was good enough and b) the rate of incoming client requests was relatively small compared to a loaded webserver. Now if you had thousands of client requests-per-second, the fork will show why-you-should-not-use-it-in-such-situations. In such cases you can use a pool of forked processes or simply write the whole damn thing using threads. For example, apache gives you the ability to change 'worker' modules...and you can experiment with that to get an idea of all these request processors. btw, threads don't complicate your code as long as you minimize data-sharing between threads and write the thread the same way as you were writing a forked process, except that you put all global variables within the thread's local data area. Kashif
Good point, but I find brainstorming at a meeting to be far more effective when all peers are at the same level of understanding. In other words, having a meeting with just engineers or just directors is better than a mix of those groups, simply because the two groups don't collide on objectives (i.e. problems/solutions vs time-to-ship). I mean if you're at a meeting where the boss says, I want this done by next month or you're fired really restricts your options of creativity. Then it's simply 'fuck it, whatever works'.
"It would be *really* nice if Dell would do some basic work to document device-driver compatibility for their systems. So if I was configuring a Linux system online, I'd like to see something like:"
.18, .19), open source drivers, closed sourced drivers, etc. It's not just figuring out which "drivers work", but also verifying they work through test cycles. The last thing anyone wants is some driver that 'works', but under some stress or certain combination of I/Os causes hangs/stability problems.
Do you know how HARD that is to do? Essentially Dell would have to QUALIFY every permutation of kernel (2.6.17,
Then your guy working 10-hours a day can do NOTHING, because of some blocker bug and has to wait from either the a) open source community, b) project maintainers, c) consultant, d) fix it himself, or e) find commercial vendor.
It's MUCH easier if Dell partners with a distro vendor and get customized drivers/patches needed to have the distro run on their boxes.
"In general, OS X packages have to bundle libraries that are not included by default in the OS. This leads to anywhere from a little to a lot of wasted disk space and memory."
First point, lots of people of have plenty of disk space...you can buy 500G hd under $300. Second, the total amount of disk space used by all of your applications (OS, msoffice, itunes, etc) is going to be 10G. Most of the space will be eaten up by your endless archive of pictures, movies, music, and porn. My windows OS, itunes, msn messenger, yahoo messenger, firefox, msoffice, nero, azureus + another 10 or so small shitty programs take up a mesealy 5.40GB of space.
Finally how many programs do you have that ship with custom libraries? And even if they ship with 3rd party libraries, as long as all other applications ship with the same library, the OS should load only a single instance of that library.
Quit from the Company then rake $$$ in consulting fees. As time goes on increase your fees as you have the upper hand.
The Linux vserver sounds very similar to UserModeLinux in both capabilities and features. Both run linux with linux, but UML has been included within the kernel. Care to describe why vserver is better/different than UML?
Kashif
None of the concepts of computer science are new, but what is ground breaking is Google touching all aspects of computer science to solve a problem. Distributed Databases, Replicated Filesystems, Clustering, Learning algorithms, job scheduling, map/reduce languages, etc. are not new. But they applied each of these sub-domains to 'searching' and 'lots of data'. Using old ideas is _new_ ways is ground breaking. That what everyone does(like Carmack and DOOM3).
If that's so, then if 1000 of your computers get infected, then that $50,000. So for us small guys, it's nothing, but for the big guys, they can atleast hire more admins just to handle the fucking spyware crap.
I had a Windows XP Professional original CD, and had lost the paper-back insert with CD-key. And I needed to format my windows. Of course, I could have gone to kazaa and tried to find a crack -- but that takes just as long as phoning.
So this is how the process went:
- I asked them I lost my insert, but had the original CD
- They said fine. But then they need to verify I had the original CD, so they made me describe how the CD looked in detail
- I told them it had holographic pictures, it was shiny, it has "unauthorized copies prohibited", I described all the little writing around the CD and the back, it says Microsoft here and there, what it says on the concentric rings, etc.
- They said OK. And proceeded to give me a brand-new CD key and scribbled it down
- Now they said, "we have to wait for you to install Windows XP upto the point where it asks you to input the key".
- I said OK, "but it will take a while"
- They said fine, "we can wait for a few minutes"
- So I proceeded to install Windows XP, and finally after 20 or so minutes, it asked me for the key. I put the key in, and it was accepted.
- I told them CD key was accepted, and we did our goodbyes and hanged up.
Two things that surprised me from this ordeal:
1) from what I heard, I didn't know you had to describe the CD so they knew you didn't have a burnt copy
2) They actually waited for me to install Windows XP and try out the key. I could have just lied and said, "yeah it worked", but If I recorded the key incorrectly, I would have been an idiot calling again saying, "it worked, but then it didn't". They would say, "WTF?"
Yes, it is asking for permission. Also making it clear to the user what the software is and that it is _not_ spyware(when in fact it is adware). This protects the anti-spyware authors ass from litigation, but also allowing user to delete based on information available about program.
So in a nutshell anti-spyware should rename themselves to "RemoveCrap Now! 2.0"
Kashif
"I've found that Netflix "throttles" my rentals after a period when I rent too many movies for them to make a profit."
If you rent too-many movies(like 10 or 20), the postage cost should be incremental for them. Assuming they have the same agreement as Blockbuster with their movie suppliers(i.e. get a cut of the profits, no cost on DVD copies). After all each DVD weighs very little, and to copy them is less than a dollar.
The only strain I see is management-wise and title availability. I.e. one guy starving everyone else of rentals.
Kashif
WARNING: 'ChangeYourToolBar.exe' isn't spyware. It is a program that changes your toolbar with its own custom version, without asking your permission. If you think this is an illegitimate program, please click on the check box to delete.
Now this removes the burden-of-proof from anti-spyware programs, and lets the user decide whether he should keep it or not based on the description.
Kashif
You can also have smart humans reading email addresses instead of botz. Kashif
Just today I was thinking about how google and other search engines should provide you with information on a particular domain. For example, if your searching for "lastest windows patches" you get a result "windowsupdate.microsoft.com". Now while the site is very relevant for what I am _looking_ for, it doesn't _directly_ give me the results I want. I have to go thru various links and reading just to get the list of patches to download.
.gif, etc)
But if google was smart, it would actually give me the list of the "lastest" patch files for all versions of windows. THAT is what I want. How you do that without using meta-data is another subject all together. A simple algorithm will be:
- return list of all relevant web pages with "windows patches"
- generate a list of objects on these pages
- find all objects that are patch files and ignore all other objects(.jpg,
- return list of these patch files
- perform "latest" operation on result and output final result in HTML
The hardest part will be identifying objects that are patches, but searching thru relevant pages you _hope_ to get the objects needed.
Google does something similar to this with images.
Kashif
SELinux -- even if apache has holes, linux is _more_ securable. You can run apache with fine-grained mode telling what it can or cannot do. So even if someone finds a buffer-overrun vunerability, the most they can do is whatever apache is _allowed_ to do.
The point with Linux is: the choice is there to make it MORE secure. Windows doesn't even have any type of MAC security. And last-time I checked, SELinux was supported for Redhat Linux ES.
And those who want to give SELinux a run can try it out on Fedora Core 3, before shelling out for RHES.
Kashif
This is like releasing a disease out in the wild and selling the disease in the market. It causes disease to get maximum exposure.
Of course, they only give you the antidote if you acually bought the disease in the market. If you want the antidote, buy the legitimized version of the disease.
Which means all bastardized versions of the disease will have to legitimize themselves or be denied the Antidote.
Kashif
So why not call your new distro, "RedHat-compatible". Meaning, redhat certified software can run on your new distro.
In one case, I can see RedHat being concerned with "here is the Free version of RedHat Linux!", they don't want some free-product thinking it is _the_ redhat linux product.
On the otherhand, one should be able to mention that this new distro is a fork of RedHat AS 3.x. It is redhat-compatible. Also if RedHat is releasing a GPL software(i.e. kernel) you should be able to say kernel-2.4-3.2as by RedHat. Meaning you have to give credit to the authors. But according to their terms you can't say "contains Red Hat Enterprise Linux X.X.".
To confuse you even more, there is a Windows Messenger and MSN Messenger. When you install MSN Messenger it simply disables start-up of Windows Messenger.
So when you uninstall MSN Messenger, it may be automatically enabled Windows Messenger. I don't know if that's what you are seeing.
Kashif
Plural form: Octopussies. :)
My father used to work for Inglis - an appliance maker. One of the problems Inglis saw was that people weren't buying new appliances as the old "versions" were working perfectly fine for the past 10-20 years. Why buy a new one if the old one ain't broke? Now Inglis manufactures things like washing machines such as one part inside is _designed_ to fail _after_ the warranty period and the whole product is only designed to last for at most 5 years. Also the whole cost cutting approach(trying to use plastic, cheaper parts, etc) is making appliances crappier day by day. What I'm trying to say, hardware makers design their machines to fail after x amount of time.
And this is evident with games. Heck, everyone knows that games are pushing hardware more than anything else. There's no way a game like Doom3 will run faster on a older hardware. Which thus proves, games are using every cycle possible otherwise games would run at 10 frames-per-second.
Why not just emulate IE's broken behaviour(with active X of course). Sure it would be 'non-standard' but there will be a button, but we're all pragmatic bastards and what things to just work. Simple way: Implement a new checkbox, "Use IE Emulation for this web-site" and save resultant web-site to list. Now I don't know how hard this will be i.e. you may need two different HTML renderers. Kashif