Use a social netowrking site, don't use one. Use MySpace, Facebook, or don't. Is this really a problem? No. Is it bothering anyone else? No. Is this news? No. Nothing to see here -- move along.
Well, obviously, you don't care. A lot of people do. In the (admittedly perhaps dubious) judgement of the Slashdot editors, enough of them care to make this relevant to the homepage. Perhaps, even if people here didn't care, it'd be relevant because it could be the beginning of a big trend in the social networking sphere.
Can anyone think of anything about parent's comment that distinguishes it from "Slashdot ran a story which doesn't interest me personally; therefore, I'm going to claim it doesn't interest anyone except the submitter and whine loudly" in all but form? If not, would someone mod it down from Insightful to Troll, Offtopic, or Flamebait (I think my comment here turned into a flame...), please? (I'd do it myself, but I've already posted.)
On LiveJournal however (which was created by the author of this article)
The summary seems to imply that the article was written by "some random guy", which, IMO, is misleading given that it was written by the founder of one of the most popular social networking sites.
Geez, your going on & on about tons of total non-issues. The point is "automated downloading of songs" is just better than streaming.
Please don't assume everyone has the same preferences as you do.
I, personally, a perfectly fine with Internet radio as-is, and I'd rather fill my hard drive with other things. If I wanted to download songs, I'm perfectly capable of downloading them myself; I listen to Internet radio because I want a mix of songs delivered to me without the space or time overhead required to maintain a large collection.
That's a buffer in the wrong direction. My point is that the main use of a p2p system would be that the main server doesn't need to transmit the latest streaming content to everyone.
That would imply that it only transmits data to some people. They need to transmit the data to other people, but there's no incentive for them to do so, because the other party would be unlikely to have data which the first party would want.
The only way there would be an incentive to share would be if you gave some people immediate data and some people data for.. say.. 30s from now. But, unless you have an extremely active swarm, you'd need a very large buffer (measured in time) to give data enough chance to propagate before everyone needs it.
It would also be even harder than it currently is in order to get people close to synchronized in terms of their playback position.
Then, you're going to want the system to be reasonably resistant to client disconnects. Certainly to a few clients, and, if you're trying to put "near" (either by network or geographical proximity) clients together, you're going to want to be resistant to if an entire block (say, a campus with a number of listeners) drops. If you have a persistent graph of who streams to who, you're going to need more time to readjust it; if you compute which pieces go where on a continuous basis (per BitTorrent) you'll also need more time, since there will be less predictability.
Once you get into a buffer window the size of a song, you no longer have Internet radio, you have "automated downloading of songs".
Store the logs in a standard on-disk format on a highly secured machine ("machines"?) with backups and such. Then, just print hashes of the logfiles on the printer.
The only reason my doctor could give for denying my medication was the risk of developing addiction. But I had showed no signs of addiction, and when I stopped the medication I experienced no withdrawal. I just experienced a lot of pain.
Just to play devil's advocate, a lack of evidence of addiction thusfar does not preclude you from becoming addicted later.
You accuse me of a straw man argument, but then choose a set of like 10 words that isn't even a complete sentence (Hell, the sentence stands without the phrase.). Thank you for ignoring one part of my attack, choosing a single, semi-related point and beating on it. Oh wait, that IS a straw man!
Quoth Wikipedia, "To 'set up a straw man' or 'set up a straw man argument' is to create a position that is easy to refute, then attribute that position to the opponent." I was criticizing your argument because you discuss flexibility of RDP when the person to which you originally replied spoke of ease of use and not of flexibility. My statement, whether phrased as an inequality or a sentence, was not a straw man.
Really, because the anecdotal evidence I've heard from people seems to agree with me. My own experience seems to agree with me as well. And this was testing both X and RDP on an internal network where load was consistent for both platforms. Not to mention you can greatly reduce the X load by using single Windows instead of creating a whole desktop. Goes back to my mentioning of flexibility. It gives the appearance of better integration then RDP ever has.
So you find RDP more efficient. I don't see how that contradicts my claim that people around me tend to find the opposite.
His whole argument was praise for MS. I was stating praise for MS is unfounded with regards to RDP since the best work on it was never even done by them.
How is this relevant?
However, it is nice of you to yell this why making some straw men yourself.
However, it is nice of you to not make sense while criticizing me for using non-sentences. (Where did I yell?)
I had zero problem with my very new desktop and the hardware for it. Also, since the only evidence anyone is presenting is anecdotal, it is quite important how many problems you or him may or may not have had. His argument was poor in presenting problems from the 1990s and ignoring the current state of the situation. Tell me, which problem(s) have you had? And what is your definition of "specialty" hardware?
Specialty hardware: hardware which would be used by a small fraction of computer users. Like rare webcams or odd MIDI interfaces.
New hardware: for example, recently released wireless chipsets.
As for exactly what problems I've had in 2007, I stand by what I said earlier: "how many problems both or either of us have had is irrelevant."
Except that this isn't truly the case. There are package management systems that with take random.dem and.rpm files and properly install them. Most websites containing these softwares offer separate repositories if the distributions do not include their software in them. I would challenge that installing most rpm and deb packages is no hardware on their respective distributions then installing some.exe. There are still programs in Windows that will have prerequisites that are required for a program to be installed.
Please take my argument in context: I was addressing specifically "apt-get or one of the various front-ends available or yum".
I can't speak in detail about yum because I haven't used any rpm-based distributions in years. APT, however, does not install random.debs. dpkg itself will install arbitrary.debs, but then you don't get all the benefits of APT. Adding a new repository works only if the software you want to install is publicly-available and distributed via the Internet.
I had in mind commercial software, which tends not to be freely available.
I really do not see (nor have I seen) this great disparity that you speak of.
That's because commercial software vendors don't tend to distribute any software for Linux in the first place. I'd argue that the lack of a simple way to distribute software for "Linux" (as opposed to some small subset of distributions) is a reason why.
Not only does remotely using X offer far more flexibility then RDP
flexibility != ease of use
I believe it is a better bandwidth user then RDP.
Anecdotal evidence disagrees with you.
It should be noted we shouldn't praise MS for RDP either, their original TS implementation sucked. Citrix licensed their stuff off to MS so MS could make a better product. Look at old TS and compare it with Citrix of that time period, you will see who was the leader.
Straw man.
This is just a poor argument. Tell me how many issues you've had in 2007? I will also say that there is better legacy support in Linux then Windows. I can still find devices that won't install drivers from the base Windows install but can in Linux, even if only well enough for me to get drivers that work well. The gap is practically closed with the largest problem being the quality of some video drivers.
And, of course, newly released hardware. And specialty hardware.
(Incidentally, I've had three problems, and how many problems both or either of us have had is irrelevant.)
Really? Use apt-get or one of the various front-ends available or yum and then tell me this. Both of these do great at handling dependencies and make installations rather painless.
Only for software which is freely available in the repository. There's no sane way for users to install other software.
It was probably your latency which was killing things (and, indeed, as you say, was simply insufficient for remote desktoping purposes regardless of the protocol.)
If you want VNC or X to be secure, you need to tunnel it. IIRC, RDP supports SSL natively. (Yes, I know some of the VNC variants support SSL, but that's just some.)
The methods you describe for X forwarding are way too technical for the average user to access his computer from home. In addition, you gloss over some things... e.g. "select the machine you want to log into and log into it".
You're also restricting yourself to using computers on the same LAN; X degrades Real Fast over slower links. (At least it does in my experience.)
Finally, you're comparing apples and oranges: RFB forwards an entire desktop-- desktop, window manager, windows, input, sound, disk, etc. X11 forwards the content of specific windows.
Maybe an operating system which doesn't crash [1] whenever supposedly hot-pluggable devices go away or it gets angry?
[1] e.g. develop unkillable processes
(If only Apple would make kill -9, umount -f, and macx_swapoff() actually work, and fix critical apps (at first glance: Finder, Dock, WindowServer, SystemUIServer, diskarbitrartiond, the printing system,...) to either be multithreaded or not block on hardware/network access, I'd be *much* more convinced of OSX's ability to be used long-term as a stable OS.)
As somebody pointed out, decision theory can be useful here.
Secondly, I've decided recently that mail and fax are truly underapreciated by us nerds: I'd spend a few minutes to write out a one-page fax and drop it to an arbitrary fax number of the bank's. From my perspective, the problem is solved (and if they chose to ignore the fax, then it's their problem.)
For large amounts (or when the error was not in my favor), a snail-mail or a fax sent to a more carefully selected number can be incredibly useful. (Example: a few months ago, the farebox on a local bus ate $5 of mine. I submitted a claim form using the traditional procedures, waited, waited, called, etc.... nothing. I sent a fax to the MBTA main offices at about 5pm on a Tuesday, and the Postal Service delivered me an apology and $21.50 of service credit on the following Thursday)
That depends. The things that would strike me as "Linux" things would be cases where I wanted high performance and/or stability.
Performance: obviously at least as good when not subject to virtualization
Stability: if a virtualized machine crashes, it hurts. If the host crashes, it hurts more. (This matters to me because most things that force me to restart under OSX are related to IO getting fucked up when devices disappear, so I'd much rather have OSX running inside the shelter of a VM.)
No, what we really need is some sort of system like this, but for Slashdot. Some sort of database that we can organize all the articles with tags through a distributed effort, so when a new story comes along, the editors will be able to find its dupe easily.
Slashdot Mirror
Sell the router.
:-) [1].
If it was sent via the postal service, at least, you're legally permitted to do whatever the hell you want with it
(If it was sent via UPS/FedEx/DHL/etc., YMMV; IANAL.)
Imagine a Beowulf cluster of hard drives!
...who cares?
Use a social netowrking site, don't use one. Use MySpace, Facebook, or don't. Is this really a problem? No. Is it bothering anyone else? No. Is this news? No. Nothing to see here -- move along.
Well, obviously, you don't care. A lot of people do. In the (admittedly perhaps dubious) judgement of the Slashdot editors, enough of them care to make this relevant to the homepage. Perhaps, even if people here didn't care, it'd be relevant because it could be the beginning of a big trend in the social networking sphere.
Can anyone think of anything about parent's comment that distinguishes it from "Slashdot ran a story which doesn't interest me personally; therefore, I'm going to claim it doesn't interest anyone except the submitter and whine loudly" in all but form? If not, would someone mod it down from Insightful to Troll, Offtopic, or Flamebait (I think my comment here turned into a flame...), please? (I'd do it myself, but I've already posted.)
On LiveJournal however (which was created by the author of this article)
The summary seems to imply that the article was written by "some random guy", which, IMO, is misleading given that it was written by the founder of one of the most popular social networking sites.
And this is exactly what's provided by automated downloading and deleting of songs coupled with semi-live mixing instructions broadcast by stations.
Except that (according to you), that would require a buffer whose size was measured in gigabytes instead of in kilobytes.
This will be my last reply.
Geez, your going on & on about tons of total non-issues. The point is "automated downloading of songs" is just better than streaming.
Please don't assume everyone has the same preferences as you do.
I, personally, a perfectly fine with Internet radio as-is, and I'd rather fill my hard drive with other things. If I wanted to download songs, I'm perfectly capable of downloading them myself; I listen to Internet radio because I want a mix of songs delivered to me without the space or time overhead required to maintain a large collection.
That's a buffer in the wrong direction. My point is that the main use of a p2p system would be that the main server doesn't need to transmit the latest streaming content to everyone.
That would imply that it only transmits data to some people. They need to transmit the data to other people, but there's no incentive for them to do so, because the other party would be unlikely to have data which the first party would want.
The only way there would be an incentive to share would be if you gave some people immediate data and some people data for.. say.. 30s from now. But, unless you have an extremely active swarm, you'd need a very large buffer (measured in time) to give data enough chance to propagate before everyone needs it.
It would also be even harder than it currently is in order to get people close to synchronized in terms of their playback position.
Then, you're going to want the system to be reasonably resistant to client disconnects. Certainly to a few clients, and, if you're trying to put "near" (either by network or geographical proximity) clients together, you're going to want to be resistant to if an entire block (say, a campus with a number of listeners) drops. If you have a persistent graph of who streams to who, you're going to need more time to readjust it; if you compute which pieces go where on a continuous basis (per BitTorrent) you'll also need more time, since there will be less predictability.
Once you get into a buffer window the size of a song, you no longer have Internet radio, you have "automated downloading of songs".
How exactly do you plan to use a p2p system for distributing a stream, without a ridiculously large buffer and/or timing oddities on the clients?
I have no idea how those software deal with installation, so I can't reply further.
Store the logs in a standard on-disk format on a highly secured machine ("machines"?) with backups and such. Then, just print hashes of the logfiles on the printer.
Of course. I'm advocating for the devil after all. I doubt he'd be sympathetic.
The only reason my doctor could give for denying my medication was the risk of developing addiction. But I had showed no signs of addiction, and when I stopped the medication I experienced no withdrawal. I just experienced a lot of pain.
Just to play devil's advocate, a lack of evidence of addiction thusfar does not preclude you from becoming addicted later.
You accuse me of a straw man argument, but then choose a set of like 10 words that isn't even a complete sentence (Hell, the sentence stands without the phrase.). Thank you for ignoring one part of my attack, choosing a single, semi-related point and beating on it. Oh wait, that IS a straw man!
.dem and .rpm files and properly install them. Most websites containing these softwares offer separate repositories if the distributions do not include their software in them. I would challenge that installing most rpm and deb packages is no hardware on their respective distributions then installing some .exe. There are still programs in Windows that will have prerequisites that are required for a program to be installed.
.debs. dpkg itself will install arbitrary .debs, but then you don't get all the benefits of APT. Adding a new repository works only if the software you want to install is publicly-available and distributed via the Internet.
Quoth Wikipedia, "To 'set up a straw man' or 'set up a straw man argument' is to create a position that is easy to refute, then attribute that position to the opponent." I was criticizing your argument because you discuss flexibility of RDP when the person to which you originally replied spoke of ease of use and not of flexibility. My statement, whether phrased as an inequality or a sentence, was not a straw man.
Really, because the anecdotal evidence I've heard from people seems to agree with me. My own experience seems to agree with me as well. And this was testing both X and RDP on an internal network where load was consistent for both platforms. Not to mention you can greatly reduce the X load by using single Windows instead of creating a whole desktop. Goes back to my mentioning of flexibility. It gives the appearance of better integration then RDP ever has.
So you find RDP more efficient. I don't see how that contradicts my claim that people around me tend to find the opposite.
His whole argument was praise for MS. I was stating praise for MS is unfounded with regards to RDP since the best work on it was never even done by them.
How is this relevant?
However, it is nice of you to yell this why making some straw men yourself.
However, it is nice of you to not make sense while criticizing me for using non-sentences. (Where did I yell?)
I had zero problem with my very new desktop and the hardware for it. Also, since the only evidence anyone is presenting is anecdotal, it is quite important how many problems you or him may or may not have had. His argument was poor in presenting problems from the 1990s and ignoring the current state of the situation. Tell me, which problem(s) have you had? And what is your definition of "specialty" hardware?
Specialty hardware: hardware which would be used by a small fraction of computer users. Like rare webcams or odd MIDI interfaces.
New hardware: for example, recently released wireless chipsets.
As for exactly what problems I've had in 2007, I stand by what I said earlier: "how many problems both or either of us have had is irrelevant."
Except that this isn't truly the case. There are package management systems that with take random
Please take my argument in context: I was addressing specifically "apt-get or one of the various front-ends available or yum".
I can't speak in detail about yum because I haven't used any rpm-based distributions in years. APT, however, does not install random
I had in mind commercial software, which tends not to be freely available.
I really do not see (nor have I seen) this great disparity that you speak of.
That's because commercial software vendors don't tend to distribute any software for Linux in the first place. I'd argue that the lack of a simple way to distribute software for "Linux" (as opposed to some small subset of distributions) is a reason why.
Not only does remotely using X offer far more flexibility then RDP
flexibility != ease of use
I believe it is a better bandwidth user then RDP.
Anecdotal evidence disagrees with you.
It should be noted we shouldn't praise MS for RDP either, their original TS implementation sucked. Citrix licensed their stuff off to MS so MS could make a better product. Look at old TS and compare it with Citrix of that time period, you will see who was the leader.
Straw man.
This is just a poor argument. Tell me how many issues you've had in 2007? I will also say that there is better legacy support in Linux then Windows. I can still find devices that won't install drivers from the base Windows install but can in Linux, even if only well enough for me to get drivers that work well. The gap is practically closed with the largest problem being the quality of some video drivers.
And, of course, newly released hardware. And specialty hardware.
(Incidentally, I've had three problems, and how many problems both or either of us have had is irrelevant.)
Really? Use apt-get or one of the various front-ends available or yum and then tell me this. Both of these do great at handling dependencies and make installations rather painless.
Only for software which is freely available in the repository. There's no sane way for users to install other software.
Take a look at the description of NT's original design and of its architecture. Both were originally quite elegant.
(Of course, Microsoft did fuck up the implementation, and compromised the design more and more in each successive release.)
It was probably your latency which was killing things (and, indeed, as you say, was simply insufficient for remote desktoping purposes regardless of the protocol.)
(My experience. YMMV)
You're both arguing extremes.
If you want VNC or X to be secure, you need to tunnel it. IIRC, RDP supports SSL natively. (Yes, I know some of the VNC variants support SSL, but that's just some.)
The methods you describe for X forwarding are way too technical for the average user to access his computer from home. In addition, you gloss over some things... e.g. "select the machine you want to log into and log into it".
You're also restricting yourself to using computers on the same LAN; X degrades Real Fast over slower links. (At least it does in my experience.)
Finally, you're comparing apples and oranges: RFB forwards an entire desktop-- desktop, window manager, windows, input, sound, disk, etc. X11 forwards the content of specific windows.
Maybe an operating system which doesn't crash [1] whenever supposedly hot-pluggable devices go away or it gets angry?
...) to either be multithreaded or not block on hardware/network access, I'd be *much* more convinced of OSX's ability to be used long-term as a stable OS.)
:-)
[1] e.g. develop unkillable processes
(If only Apple would make kill -9, umount -f, and macx_swapoff() actually work, and fix critical apps (at first glance: Finder, Dock, WindowServer, SystemUIServer, diskarbitrartiond, the printing system,
Oh, and I'm not bitter in the least.
As somebody pointed out, decision theory can be useful here.
Secondly, I've decided recently that mail and fax are truly underapreciated by us nerds: I'd spend a few minutes to write out a one-page fax and drop it to an arbitrary fax number of the bank's. From my perspective, the problem is solved (and if they chose to ignore the fax, then it's their problem.)
For large amounts (or when the error was not in my favor), a snail-mail or a fax sent to a more carefully selected number can be incredibly useful. (Example: a few months ago, the farebox on a local bus ate $5 of mine. I submitted a claim form using the traditional procedures, waited, waited, called, etc.... nothing. I sent a fax to the MBTA main offices at about 5pm on a Tuesday, and the Postal Service delivered me an apology and $21.50 of service credit on the following Thursday)
That depends. The things that would strike me as "Linux" things would be cases where I wanted high performance and/or stability.
Performance: obviously at least as good when not subject to virtualization
Stability: if a virtualized machine crashes, it hurts. If the host crashes, it hurts more. (This matters to me because most things that force me to restart under OSX are related to IO getting fucked up when devices disappear, so I'd much rather have OSX running inside the shelter of a VM.)
Someone please tell me I'm talking out of my ass
Oh, I'd say you're full of hot air.
Indeed; the boringness of magical battles is the only point with which I agree with GP.
Do not want.
50 GB/s = 50*8 Gbit/s = 400 Gbit/s.
400 is nowhere near the order of magnitude different from 500 as your choice of units imply.
No, what we really need is some sort of system like this, but for Slashdot. Some sort of database that we can organize all the articles with tags through a distributed effort, so when a new story comes along, the editors will be able to find its dupe easily.
Oh, wait...