A brief read of the paper tells me that it's simply a brute force method of checksumming; they send all possible bit combinations in the checksum field (using the same actual message bits), and the server only responds to the correct checksum. They don't actually compute the checksum locally.
While it does work, it's basically trading a (relatively) small amount of actual computation for a large amount of bandwidth. Actually sending those packets out in the first place may take more computation that the actual checksum would, so I'm not sure if this is entirely useful...
I already did, all I'm saying is for every codec I can think of except Sorenson there is already Free software than can play it. If this is the first that can do Sorenson, great, but personally I'd just rather see Sorenson stop being used than 'finally' have support for it.
That's not even starting to talk about the fact they want $19.95 for their thing.
Sorry, I should have added they can play some quicktime movies (since that's the topic here). I've been able to play every QT movie I've encountered except Sorenson codec'd movies.
...unless they specifically mean they can play the Sorenson codec. There are Free programs out there already that can play AVI, ASF, etc files; some I know about are:
if just one of those less-than-top-of-the-line PCs fails while performing a mission critical severity-one application
Ummm...who the hell said this was for mission-critical applications?
And as far as companies not caring about saving money, I'm not sure what companies you've been working for but all the ones I've seen do care about saving money. They do savings analysis, and in the cases I was talking about (like where I specifically mention deployments where the users aren't resource-intensive) they'd probably (but it of course depends on the specific circumstances) approve the X terminal solution.
I see the value for home and small business applications (maybe even a department-wide deployment, particularly in with smarter IT users) but that's it.
A home application? Anyone using a central server with several X terminals throughout their house (first) has a pretty big house (and can afford a full system in each room), and (second) probably is a software programmer who's going to want a full system in each room.
Small business may find it useful. But not as useful as a bigger business (or government as I said before).
From the general direction of your comment, I don't think you quite understand exactly what's meant by a X terminal.
An X terminal (at least, those discussed in the article) runs only X. Nothing else. It gets all X content from the server machine(s). So, opening email and starting Netscape will be quite fast, not '30 minutes' as you describe...
In fact, this configuration most likely will appear faster than what the user has at home, since the server(s) typically are quite fast.
Well, the rule of thumb in large corporations is that labor costs greatly exceed hardware costs (by ratios of 10-to-1 or more!). If you had a large group of techs running around debugging and fixing your garbage hardware, any cost savings would be eaten almost immedately.
You're assuming the hardware will be failing on a regular basis; I don't agree with that assumption. I don't think anyone's talking about buying used hardware here, just old hardware. Big, big difference.
One place I worked at had a sign up: "Technicians Shall Not Open The Case". Meaning that if it couldn't be fixed by minor tweaking or re-imaging, the machine was pulled, replaced, and donated immedately.
So what's the problem? Tech spends 5 minutes (10 tops) re-imaging the machine; if it still doesn't work replace it. Keeping spares would not be expensive.
For home/small business users I think the second choice is a valid one, but for large business and corporations I just don't think they'll ever see the value in it.
If you're talking about a large business, that savings could be quite sizable...say, $500 per terminal (which is probably less than the real savings) times 1000 terminals...that's half a million. I'd say you could pay one or two sysadmin's salaries to make sure all that different hardware worked with that kind of cost savings. And really, once X is set up, you pretty much don't have to touch the system after that unless the hardware fails (since the user is not really using software locally).
Plus, when you need to upgrade in 2-3 years, you really only have to upgrade the main servers, which is a massive cost savings.
I'd say that large businesses are exactly the ones who can benefit most from this. Especially places where the user base does not do system-intensive things (the government comes to mind here...;-)
Microsoft's response...
on
Hotmail Hacked
·
· Score: 5, Funny
...is priceless:
"However," Microsoft said, "we recognize the concerns raised in the computational infeasibility of this mechanism and are investigating ways that we can raise this bar even higher."
Like Taco said...you just can't make this stuff up. That response is just too funny.
A company that recently went bankrupt was ECW. They filed for Chapter 11. Were dead and gone three months later, only to be "resurrected" by the Vin man in name only.
Wait, are you saying that the ECW and WCW aren't the greatest threat the WWF has ever encountered?:-)
People should not confuse this with a Chapter 7 liquidation, where you close the doors and sell off the assets. That is not what we have done.
We filed a Chapter 11 reorganization. This will allows us to deal with our creditors fairly and equitably and at the same time continue to operate the company. We are still shipping products and porting new games and expect to be doing so for a long, long time.
Don't worry, my post to lkml is public record (and made kernel traffic too! I didn't expect that...!). I don't think I'll get in any trouble, but if I do it was my fault anyway;-)
Anyway, I'm very interested to hear Scott's answer to your question!
Witness this recent thread on the linux kernel mailing list in which one Dan Streetman from IBM states plainly:
Mmm..ouch! Didn't think that would make it onto/.!
As I replied to Mike, keep in mind that IBM is huge and there has not been a mandate from Lou stating "everything is Open!". There still has to be a business reason to Open code, especially if that code has been closed for a long time and is now getting 'ported'. I tried to provide a business reason, but it didn't work.
After having said that, I do wish IBM educated their employees on the benefits of Open Source, and encouraged more internal software to be Open. From what I've seen so far, the primary focus is on what Open software is already out there, and available. In fact the process of getting internal Open Source approval is not simple.
And (in case my manager is reading this;-) getting approval to work on Open Source in your 'personal time' is a long process too.
This obviously makes linux developers doubtful about the real motivations behind IBMs open source efforts
No, no, no - IBM does have an interest and good motivation in Open Source software. I won't speculate on (i.e. don't know) the actual motivation of the real decision makers, but I know that there are a lot of people internally that are very pro-OSS, for the right reasons; and that's what really matters. Keep in mind that IBM is allowing internally created code to be Open Sourced! There are plenty of projects at DW, and any IBMer can get approval to release certain code as Open Source. It might be a little more difficult than if you were doing it on your own, but hey - IBM is paying people to write Open Source software! Just because there are some cases were employees couldn't get approval to Open certain code, doesn't mean that IBM has bad intentions or motivations.
Given that there are a wealth of Linux distributions already available in the marketplace, it seems that IBM has to choose one (or more), or create one (or more).
So, will IBM create a Linux distribution (or multiple distributions)?
If not, why? And what distributions will/has IBM choose to support/market?
If so, why? And will that/those distribution(s) compete with other distributions (i.e. be directly available) or only (or at least primarily) available as part of an IBM 'solution'?
Can't you just fire a couple of EMPs at the disk first? Or would that damage the
hardware?
Yes the EMPs would destroy all electrical equipment, turning the system into a hunk of scrap metal.
Placing the drive in a very large and powerful magnetic field would (disputably) erase the drive, but it's certainly possible that there would be residual data left...
Sometimes the end does justify the means. If the evil combatted is so extra-ordinarily bad, and if the only way to bring it down is a slightly unethical action, I'd opt
for the slightly unethical action, rather than the unspeakable evil.
Translation:
When fighting those without ethics the only way to win is abandon you own ethics.
Not that I disagree, my inital post was trying to convey that declaring we must fight wars in an 'ethical' manner is ridiculous. See "Apocalypse Now" for exactly my opinion on the matter.
After all, how many/.ers have old 486 or Pentium machines running as
firewalls/routers?
I suspect that 486 does not have enough processing power to be a full G. Lite DSL modem (i.e. 1.5Mbps downstream). It's much, much easier routing ethernet packets than doing full handling of a DSL analog signal line at 1.5Mbps and routing...even if the routing was left to another system, I'll bet the modem, at full load, will take a considerable amount of CPU resources...(i.e. maybe too much for a 486).
Maybe it will work...but especially since DSL modems are dropping fast in price I don't see this really being useful.
Re:Not that impressive
on
DSLBlaster?
·
· Score: 1
8 bits x 8KHz sampling rate =
64Kb/sec
8 bits * 8kHz = 64kHz, not 64kbps.
'bandwidth' is always misused in Comp Sci, it really is measured in Hz (or kHz/mHz), not in bits per second. And 8kHz will not necessarily result in 8kbps (it's not a 1Hz = 1bps relationship).
Slashdot Mirror
While it does work, it's basically trading a (relatively) small amount of actual computation for a large amount of bandwidth. Actually sending those packets out in the first place may take more computation that the actual checksum would, so I'm not sure if this is entirely useful...
Flamebait? No thanks, I'm not hungry.
I already did, all I'm saying is for every codec I can think of except Sorenson there is already Free software than can play it. If this is the first that can do Sorenson, great, but personally I'd just rather see Sorenson stop being used than 'finally' have support for it.
That's not even starting to talk about the fact they want $19.95 for their thing.
Sorry, I should have added they can play some quicktime movies (since that's the topic here). I've been able to play every QT movie I've encountered except Sorenson codec'd movies.
Note that the avifile project has links to many other players...
XAnim is (AFAIK) the oldest player. It supports some AVIs but (IIRC) not ASFs...
Most of the ASF et. al. support comes from using the Windows binary codecs...
If you don't like what happens at /. they why don't you stop coming here?
Just change 'www.nytimes.com' to 'archive.nytimes.com' for any URL (I think).
So here, it's. html
S VAL.html
http://www.nytimes.com/2001/08/26/business/26SVAL
to
http://archive.nytimes.com/2001/08/26/business/26
Ummm...who the hell said this was for mission-critical applications?
And as far as companies not caring about saving money, I'm not sure what companies you've been working for but all the ones I've seen do care about saving money. They do savings analysis, and in the cases I was talking about (like where I specifically mention deployments where the users aren't resource-intensive) they'd probably (but it of course depends on the specific circumstances) approve the X terminal solution.
I see the value for home and small business applications (maybe even a department-wide deployment, particularly in with smarter IT users) but that's it.
A home application? Anyone using a central server with several X terminals throughout their house (first) has a pretty big house (and can afford a full system in each room), and (second) probably is a software programmer who's going to want a full system in each room.
Small business may find it useful. But not as useful as a bigger business (or government as I said before).
An X terminal (at least, those discussed in the article) runs only X. Nothing else. It gets all X content from the server machine(s). So, opening email and starting Netscape will be quite fast, not '30 minutes' as you describe...
In fact, this configuration most likely will appear faster than what the user has at home, since the server(s) typically are quite fast.
You're assuming the hardware will be failing on a regular basis; I don't agree with that assumption. I don't think anyone's talking about buying used hardware here, just old hardware. Big, big difference.
One place I worked at had a sign up: "Technicians Shall Not Open The Case". Meaning that if it couldn't be fixed by minor tweaking or re-imaging, the machine was pulled, replaced, and donated immedately.
So what's the problem? Tech spends 5 minutes (10 tops) re-imaging the machine; if it still doesn't work replace it. Keeping spares would not be expensive.
If you're talking about a large business, that savings could be quite sizable...say, $500 per terminal (which is probably less than the real savings) times 1000 terminals...that's half a million. I'd say you could pay one or two sysadmin's salaries to make sure all that different hardware worked with that kind of cost savings. And really, once X is set up, you pretty much don't have to touch the system after that unless the hardware fails (since the user is not really using software locally).
Plus, when you need to upgrade in 2-3 years, you really only have to upgrade the main servers, which is a massive cost savings.
I'd say that large businesses are exactly the ones who can benefit most from this. Especially places where the user base does not do system-intensive things (the government comes to mind here...;-)
"However," Microsoft said, "we recognize the concerns raised in the computational infeasibility of this mechanism and are investigating ways that we can raise this bar even higher."
Like Taco said...you just can't make this stuff up. That response is just too funny.
Wait, are you saying that the ECW and WCW aren't the greatest threat the WWF has ever encountered? :-)
Vince wouldn't lie like that!
Scott Draeker (President of Loki) sent a comment to linuxgames.com which said:
People should not confuse this with a Chapter 7 liquidation, where you close the doors and sell off the assets. That is not what we have done.
We filed a Chapter 11 reorganization. This will allows us to deal with our creditors fairly and equitably and at the same time continue to operate the company. We are still shipping products and porting new games and expect to be doing so for a long, long time.
But, you never know what's going to happen...
Anyway, I'm very interested to hear Scott's answer to your question!
Mmm..ouch! Didn't think that would make it onto /.!
As I replied to Mike, keep in mind that IBM is huge and there has not been a mandate from Lou stating "everything is Open!". There still has to be a business reason to Open code, especially if that code has been closed for a long time and is now getting 'ported'. I tried to provide a business reason, but it didn't work.
After having said that, I do wish IBM educated their employees on the benefits of Open Source, and encouraged more internal software to be Open. From what I've seen so far, the primary focus is on what Open software is already out there, and available. In fact the process of getting internal Open Source approval is not simple.
And (in case my manager is reading this ;-) getting approval to work on Open Source in your 'personal time' is a long process too.
This obviously makes linux developers doubtful about the real motivations behind IBMs open source efforts
No, no, no - IBM does have an interest and good motivation in Open Source software. I won't speculate on (i.e. don't know) the actual motivation of the real decision makers, but I know that there are a lot of people internally that are very pro-OSS, for the right reasons; and that's what really matters. Keep in mind that IBM is allowing internally created code to be Open Sourced! There are plenty of projects at DW, and any IBMer can get approval to release certain code as Open Source. It might be a little more difficult than if you were doing it on your own, but hey - IBM is paying people to write Open Source software! Just because there are some cases were employees couldn't get approval to Open certain code, doesn't mean that IBM has bad intentions or motivations.
Given that there are a wealth of Linux distributions already available in the marketplace, it seems that IBM has to choose one (or more), or create one (or more).
So, will IBM create a Linux distribution (or multiple distributions)?
If not, why? And what distributions will/has IBM choose to support/market?
If so, why? And will that/those distribution(s) compete with other distributions (i.e. be directly available) or only (or at least primarily) available as part of an IBM 'solution'?
Why isn't this in the BSD section of /.?
Can't you just fire a couple of EMPs at the disk first? Or would that damage the hardware?
Yes the EMPs would destroy all electrical equipment, turning the system into a hunk of scrap metal.
Placing the drive in a very large and powerful magnetic field would (disputably) erase the drive, but it's certainly possible that there would be residual data left...
people who cripple themselves into inaction through excessive contemplation of 'ethics' = dead
Is there an echo in here? I could have sworn I just said that...
Sometimes the end does justify the means. If the evil combatted is so extra-ordinarily bad, and if the only way to bring it down is a slightly unethical action, I'd opt for the slightly unethical action, rather than the unspeakable evil.
Translation:
When fighting those without ethics the only way to win is abandon you own ethics.
Not that I disagree, my inital post was trying to convey that declaring we must fight wars in an 'ethical' manner is ridiculous. See "Apocalypse Now" for exactly my opinion on the matter.
Col. Kurtz : "The Horror...the horror..."
stopping genocide: good
If 'stopping genocide' requires unethical actions, why pretend that war is ethical at all?
qualified ethics : pointless
After all, how many /.ers have old 486 or Pentium machines running as
firewalls/routers?
I suspect that 486 does not have enough processing power to be a full G. Lite DSL modem (i.e. 1.5Mbps downstream). It's much, much easier routing ethernet packets than doing full handling of a DSL analog signal line at 1.5Mbps and routing...even if the routing was left to another system, I'll bet the modem, at full load, will take a considerable amount of CPU resources...(i.e. maybe too much for a 486).
Maybe it will work...but especially since DSL modems are dropping fast in price I don't see this really being useful.
8 bits x 8KHz sampling rate = 64Kb/sec
8 bits * 8kHz = 64kHz, not 64kbps.
'bandwidth' is always misused in Comp Sci, it really is measured in Hz (or kHz/mHz), not in bits per second. And 8kHz will not necessarily result in 8kbps (it's not a 1Hz = 1bps relationship).