The replacement for the MS system is not without its own issues.
By saying that you miss the major point: what are the risks?
Issues might and do happen in every piece of software different from the helloworld.c.
If you go with closed source solution, you are at the mercy of a 3rd party. If you go with open source system you have an option to replace the 3rd party if it is not delivering. Or even hire full-rime people to run the operations.
N.B. Note that "closed source" != "proprietary". Some commercial systems used to run infrastructures of the scale often come with the source code. Users are not allowed to distribute it, but patching/optimizing/etc is perfectly fine. As user concerned, they are open source.
In particular case of LSE, operational costs of the system itself are much higher than the cost of software used to run it. And on top of that, you have number of third parties you have to deal with and coordinate constantly. It was only natural for LSE to -eventually- come to the point where they have full control over system.
That way, LSE might be actually paying more for running open source system, yet risks are much more manageable. (*) Something with closed source software you can achieve only with exceptionally good 3rd party partners. And such partners are very very very rare.
(*) Manageable: no, not necessarily lower risks, but manageable. Downtimes per se are not the problem - the unplanned downtimes are.
As for myself, I don't think Linux needs world domination on the desktop, it does need interoperability though.
You do understand that as long as MSFT has a desktop dominance, it would do anything to make sure that there would be no interoperability with any other competing OS?
Because interoperability (through truly open standards) is what gives people choice.
[...] I don't mean the OOXML farce that was pulled through the ISO.
And MSFT many times exemplified that in their opinion a "de facto" standard (they have complete control over like OOXML) is just as good as a "de jure" standard.
That's why as long as MSFT has >50% of market, there would be neither interoperability nor open standards.
What really amazes me is the Nazi in distro development circles and forums screaming at new users about "security", "security", "never log in as root", "never do xyz", "why don't need a root x session" yet they insist on loading up their distro with every possible piece of crap script, new program, extension, plugin, and so on by default.
Thanks for sharing. Glad to know that I'm not alone who sees that as at best moronic.
You normally do not have to take every kernel update they push down the pipe, and after a while you learn the hard way that updating your kernel (and a lot of other things) is something you only do if you really really have to do on a production system.
Frankly, it's beyond me why Ubuntu does what it does the way it does. (Often kernel updates also screw people like me who have to have a separate/boot file system: because Ubuntu can't recognize that mkinitrd failed due to a full file system. And there are no automatic file system monitoring by default nor clean up of unused kernels/etc.)
I had similar experience to yours when I was running few servers under Debian Stable. 230+ days uptime was my record (disrupted by a janitor's mop). Heck, some people even did "apt-get update && apt-get upgrade" on a *week-end* from a *cronjob*. And it did work.
Rush after the bleeding edge makes me sad: it essentially ruins what Linux is best known for - stability. Because in the end it doesn't matter why OS interrupted you - was it a crash or was it an "strongly suggested" update. It still interrupts your work.
The funny thing is, The *ENTIRE* fucking F/OSS movement is built around making cheap knock-offs of existing successful proprietary products and giving it away.
Proprietary competition is no different: make a similar product and sell it cheaper than competitors.
Because digital distribution sheds many usual to other industries costs, software prices went down to literally zero.
And believe me, proprietary "cheap knock-offs" are much worse than the general F/LOSS treat.
Its funny how Linux has only succeeded when you place it far away from the user (servers which require maintenance by professional system admins) or lock it down so that the user cant interact with it (embedded devices).
It's not funny. It's logical (provided that you have brains of course).
On desktop, Linux is a newcomer. (Even Windows needed more than decade to become a monopoly.)
On servers, Linux replaces older *NIX systems to which it is largely compatible.
But on the desktop, haha. 1% market share. A resounding success.
Biggest problem is mutimedia and it is a problem because due to patent regime in U.S., one pretty much can't distribute anything video/audio playing without fear to be sued. As long as distros are going to disregard multimedia, no real concerted development would occur to fix that. And multimedia, as name implies, requires the concerted development effort to resolve often conflicting requirements of different components (audio driver v. video driver v. X server v. disk I/O v. demux v. codec v. GUI).
Most distros now push kernel updates more often - and they require reboots. Never measured, but my Ubuntu 8.04 LTS system asks for a reboot about every month.
Nowadays 100+ days probably only achievable on Debian Stable. But I heard Debian want to fix that too...
P.S. Heck, even Mac OS X now asks for a restart for pretty much every update.
bittorrent is GREAT for helping distribute some official file.. but for anything 'underground' it's awful - even with the trackerless torrent hack.
Nobodies talking about something underground. There is nothing 'underground' about how WikiLeaks operates - it is just way to political meaning that it might become a target.
Why has everyone moved to torrents when something like gnutella would be far more appropriate?
Gnutella is good for small stuff, but even at MB sizes, lack of proper check-summing oftentimes results in a corrupt download. Also it is susceptible to poisoning attacks most of all P2P networks. And depends on central specilized servers to do searches.
It's like eDonkey/eMule/Kad and BitTorrent learned from Gnutella how NOT to do P2P.
I don't really see how it's ugly, it's simply function over form;...
I presume because most commenters here are from U.S. where cars long time ago stopped being a transportation media and have become a way to boost your ego. But I guess that is the case all over the G7 countries.
... most of the downloading endpoints really don't have a lot of upstream bandwidth...
Who cares. The point is that whether free web can be user support or it has to rely on large businesses (which operate backbones).
Problem of CDNs in the context is obvious: WikiLeaks can't use them for political reasons and CDNs likely to refuse serve WikiLeaks for the vary same political reasons.
The difference P2P (or "user-supported cloud") makes is that the CDN is fully user supported thus can be apolitical.
... has many awkward deployment issues...
Yeah. I figured it. The problem in the case with BitTorrent is that it doesn't have decentralized search like eMule's Kademlia. I do not think they support file modification timestamps, but that is trivial technicality. Otherwise, as long as a P2P network supports decentralized search and clients can do automatic content expiration, it can be used as a simple mirror service for web pages.
For best performance, Freenet will run continually. It should not interfere with your computer usage, as it requires around 200MB of RAM and 10% of one CPU core, plus some disk access.
And no wonder considering that it is written in Java...
Not all PCs have Java installed. First. Second. With that kind of resource utilization, I do not see Freenet catching with average consumers.
Probably they should invest into a lightweight C/C++ client. That even I would let run on my systems.
It may be easier to support; it may mean greater security; but if it doesn't offer enough stuff that those pesky users want then it will be resisted and then subverted. IT Departments cannot impose there will on their clients in the long term.
Or in other words, the IT becomes "single point of failure." While PC can run without any infrastructure or IT involvement.
That has been always the biggest argument against thin clients.
Plus few people actually want to work on the thin clients, so you rarely see any grass root support for them coming from employees. And "better PC" is often used by management as an incentive for subordinates too.
Look at it for a moment from different prospective: how many books you have read which had many intervened story lines. How many you actually liked? Writing such book is by no mean an easy task.
Same applies to game scripts. The fact that Fallout had a flexible story structure, amendable by accomplished quests, doesn't mean it is easy to make one.
Rather the fact itself that Fallout is still remembered, highlights that it is rather difficult to do.
But highly customizable characters are pain for story designers. In story player encounters the monsters with not some random stats - but the stats which reflect the role the monster plays in story.
Just recall the NWN1 where they had the magic button "Recommend". During character creation and level up screen, hitting "Recommend" was defaulting to properly leveled warrior and guaranteeing that player wouldn't hit an obstacle s/he can't overcome. Because playing custom character means that there might be monsters you can't defeat alone. And for some exotic classes there were even specialized modules, allowing you to play in full force, because default campaign was designed for a warrior. (Even playing as barbarian, due to its low persuasion, one would miss many interesting side quests.)
Essentially, exotic/custom classes increase game complexity on both sides. Making a campaign becomes more complicated as many classes has to be taken into account. Playing with a custom class requires quite a skill of knowing and expoloiting strengths and weaknesses - your own and monsters. That's not something average gamers might expect - many are way too used to bashing stuff with a sword or annihilating everything with a magic.
So in their view, a clean-room implementation of a libmysql work-alike would have been somehow subject to their GPL claims. That was the evil part.
Oh. That's new to me.
Honestly, though I was on the receiving end of the problem, the MySQL AB stance on that forced me (and my employer) to switch to Linux: MySQL's Windows up-to-date version was commercial. And I came to be very thankful to that very painful at the time decisions (painful == I was a principal windows software developer at the times).
I would like to have seen that challenged in court(s).
No need. IANAL but IMO the argument fails at the interoperability clauses, commonly found in the copyright law exceptions. IIRC even DMCA has such clause, allowing to break DRM in some cases for interoperability reasons. Essentially, copyright/etc can't be used to deny the users access to their information.
Oracle's bread and butter is support contracts, not license revenue.
That's right. And I'm already terrified with prospects that to deploy MySQL I would need the same amount of pains (and two/more certifications) as with Oracle.
MySQL is simple and slick RDBMS. In some part it caught in market because it it very easy to deploy, rather easy to develop for and quite easy to maintain.
Monty is a liar, Groklaw caught him in the lie and he shouldn't be trusted.
As a techie who built a RDBMS competing with Oracle, I would pay anytime Monty much more attention than some over-religious zealots over at Groklaw.
Because as a techie too, I understand that in real life it is not all that black and white. And depending on to whom you talk to, you pick different - often apparently (but only apparently) contradictory - arguments. Because you know different people would listen to different arguments.
He's abusing the EU approval process for personal gain.
No amount of ad nominem attacks would change the fact that Oracle can easily (and likely would) screw up whole commercial eco-system around MySQL. Monty now probably only a small part of it and has much less to loose compared to its other participants.
That is nothing new. The problem is that Monty now found himself on the other side of fence and he is faced with the same choice as MySQL AB customers were in past: get a free GPLed MySQL fork or buy a license for a commercial MySQL variant.
GPL played the evil trick that you can't link commercial applications against libmysql*. IOW, to develop proprietary closed-source MySQL based product, you have to buy a license for the commercial fork of MySQL. And that is to my understanding the matter of his objection. And it is a rather valid objection, since Oracle now has a way to kill completely (not only Monty but) whole commercial infrastructure surrounding MySQL .
On one side I'm sadistically happy that Monty himself got the taste of it. On another side I also recognize that building something like MySQL completely open source might have been impossible and some revenue stream is much required. (Even much touted PostgreSQL, thanks to BSD license, has quite a number of proprietary applications around it.)
IMO both console and PC gaming are dying - in a figure of speech.
"Rise" of casual gaming now only further highlighted that gaming as many of us relate to it is a rather smallish niche of modern entertainment. For many people the idea that one has to spend hours and hours before screen (TV or PC display) only to learn controls - before s/he could enjoy a game - is simply absurd. E.g. I take my (now decade old) FPS skills for granted and even though I write this it is still hard for me to imagine why many people just can't "get it".
The point is that we do not need skills to enjoy movies or TV shows. And the learning divide, for as long as it would exist, would be making gaming business vulnerable.
Console gaming in that respect is even more vulnerable: PC gaming in a way is self-sustainable thanks to the fact that a gamer can become game developer rather easily. Console gaming on other side creates only consumers who depend completely on a business to provide entertainment to them.
Though thanks to Nintendo's WiiMote (and following it Sony's Wand and MS' Natal) I'd say console business has a pretty good shot at making games more accessible. As business and market, they are still pretty strong and still manage to come up with something new.
99% of me agrees with you wholeheartedly. It's rather useless.
Yet there is 1% of me which also knows that it is an integral part of science to make up some silly theories and models about stuff which we would never know for sure. After all, pretty much everything in the today's science started some long time ago from a silly theory. It was silly and unknowable in the past - while now it is treated as an fact.
... just like the rest of the science: it is all based on observations made by rather imperfect human eyes and generalizations delivered by our rather interpretive brains.
The crucial difference is whether scientists do understand the shaky foundation or they foolishly insist on objectivity of their research.
What is also right as to remind everybody with whom we are dealing here.
M$' "embrace and extend" (and extinguish) strategy is probably worst kept secret ever. If they would start taking competition with Google more seriously, I think user's privacy would end up in even greater jeopardy.
Slashdot Mirror
The replacement for the MS system is not without its own issues.
By saying that you miss the major point: what are the risks?
Issues might and do happen in every piece of software different from the helloworld.c.
If you go with closed source solution, you are at the mercy of a 3rd party. If you go with open source system you have an option to replace the 3rd party if it is not delivering. Or even hire full-rime people to run the operations.
N.B. Note that "closed source" != "proprietary". Some commercial systems used to run infrastructures of the scale often come with the source code. Users are not allowed to distribute it, but patching/optimizing/etc is perfectly fine. As user concerned, they are open source.
In particular case of LSE, operational costs of the system itself are much higher than the cost of software used to run it. And on top of that, you have number of third parties you have to deal with and coordinate constantly. It was only natural for LSE to -eventually- come to the point where they have full control over system.
That way, LSE might be actually paying more for running open source system, yet risks are much more manageable. (*) Something with closed source software you can achieve only with exceptionally good 3rd party partners. And such partners are very very very rare.
(*) Manageable: no, not necessarily lower risks, but manageable. Downtimes per se are not the problem - the unplanned downtimes are.
As for myself, I don't think Linux needs world domination on the desktop, it does need interoperability though.
You do understand that as long as MSFT has a desktop dominance, it would do anything to make sure that there would be no interoperability with any other competing OS?
Because interoperability (through truly open standards) is what gives people choice.
[...] I don't mean the OOXML farce that was pulled through the ISO.
And MSFT many times exemplified that in their opinion a "de facto" standard (they have complete control over like OOXML) is just as good as a "de jure" standard.
That's why as long as MSFT has >50% of market, there would be neither interoperability nor open standards.
What really amazes me is the Nazi in distro development circles and forums screaming at new users about "security", "security", "never log in as root", "never do xyz", "why don't need a root x session" yet they insist on loading up their distro with every possible piece of crap script, new program, extension, plugin, and so on by default.
Thanks for sharing. Glad to know that I'm not alone who sees that as at best moronic.
You normally do not have to take every kernel update they push down the pipe, and after a while you learn the hard way that updating your kernel (and a lot of other things) is something you only do if you really really have to do on a production system.
Frankly, it's beyond me why Ubuntu does what it does the way it does. (Often kernel updates also screw people like me who have to have a separate /boot file system: because Ubuntu can't recognize that mkinitrd failed due to a full file system. And there are no automatic file system monitoring by default nor clean up of unused kernels/etc.)
I had similar experience to yours when I was running few servers under Debian Stable. 230+ days uptime was my record (disrupted by a janitor's mop). Heck, some people even did "apt-get update && apt-get upgrade" on a *week-end* from a *cronjob*. And it did work.
Rush after the bleeding edge makes me sad: it essentially ruins what Linux is best known for - stability. Because in the end it doesn't matter why OS interrupted you - was it a crash or was it an "strongly suggested" update. It still interrupts your work.
The funny thing is, The *ENTIRE* fucking F/OSS movement is built around making cheap knock-offs of existing successful proprietary products and giving it away.
Proprietary competition is no different: make a similar product and sell it cheaper than competitors.
Because digital distribution sheds many usual to other industries costs, software prices went down to literally zero.
And believe me, proprietary "cheap knock-offs" are much worse than the general F/LOSS treat.
Its funny how Linux has only succeeded when you place it far away from the user (servers which require maintenance by professional system admins) or lock it down so that the user cant interact with it (embedded devices).
It's not funny. It's logical (provided that you have brains of course).
On desktop, Linux is a newcomer. (Even Windows needed more than decade to become a monopoly.)
On servers, Linux replaces older *NIX systems to which it is largely compatible.
But on the desktop, haha. 1% market share. A resounding success.
Biggest problem is mutimedia and it is a problem because due to patent regime in U.S., one pretty much can't distribute anything video/audio playing without fear to be sued. As long as distros are going to disregard multimedia, no real concerted development would occur to fix that. And multimedia, as name implies, requires the concerted development effort to resolve often conflicting requirements of different components (audio driver v. video driver v. X server v. disk I/O v. demux v. codec v. GUI).
Thanks God XMMS doesn't look like WinAmp...
Though as switching sound cards in GUI is PITA and still generally broken, I stick most of the time with mpg123...
Most distros now push kernel updates more often - and they require reboots. Never measured, but my Ubuntu 8.04 LTS system asks for a reboot about every month.
Nowadays 100+ days probably only achievable on Debian Stable. But I heard Debian want to fix that too...
P.S. Heck, even Mac OS X now asks for a restart for pretty much every update.
What the fuck is with this need to bash countries? Especially when it's so far off the mark.
What the fuck is with this habit to slap "ugly" on everything what is purely functional?
I live in Nepal and you would be hard pressed to find one person to buy that based upon its looks.
That was precisely my point.
bittorrent is GREAT for helping distribute some official file.. but for anything 'underground' it's awful - even with the trackerless torrent hack .
Nobodies talking about something underground. There is nothing 'underground' about how WikiLeaks operates - it is just way to political meaning that it might become a target .
Why has everyone moved to torrents when something like gnutella would be far more appropriate?
Gnutella is good for small stuff, but even at MB sizes, lack of proper check-summing oftentimes results in a corrupt download. Also it is susceptible to poisoning attacks most of all P2P networks. And depends on central specilized servers to do searches.
It's like eDonkey/eMule/Kad and BitTorrent learned from Gnutella how NOT to do P2P.
I don't really see how it's ugly, it's simply function over form; ...
I presume because most commenters here are from U.S. where cars long time ago stopped being a transportation media and have become a way to boost your ego. But I guess that is the case all over the G7 countries.
It's called a Content Delivery Network
Who cares. The point is that whether free web can be user support or it has to rely on large businesses (which operate backbones).
Problem of CDNs in the context is obvious: WikiLeaks can't use them for political reasons and CDNs likely to refuse serve WikiLeaks for the vary same political reasons.
The difference P2P (or "user-supported cloud") makes is that the CDN is fully user supported thus can be apolitical.
Yeah. I figured it. The problem in the case with BitTorrent is that it doesn't have decentralized search like eMule's Kademlia. I do not think they support file modification timestamps, but that is trivial technicality. Otherwise, as long as a P2P network supports decentralized search and clients can do automatic content expiration, it can be used as a simple mirror service for web pages.
Oops. Was a reply to GP.
From the site:
For best performance, Freenet will run continually. It should not interfere with your computer usage, as it requires around 200MB of RAM and 10% of one CPU core, plus some disk access.
And no wonder considering that it is written in Java...
Not all PCs have Java installed. First. Second. With that kind of resource utilization, I do not see Freenet catching with average consumers.
Probably they should invest into a lightweight C/C++ client. That even I would let run on my systems.
I'm surprised nobody yet thought up a BitTorrent analogue for HTTP - to offload/share traffic from busy sites.
I guess latencies are the problem, but faced with information being not available at all, higher latencies are probably a good compromise.
Sites like Wikipedia or WikiLeaks could definitely benefit from such technology.
It may be easier to support; it may mean greater security; but if it doesn't offer enough stuff that those pesky users want then it will be resisted and then subverted. IT Departments cannot impose there will on their clients in the long term.
Or in other words, the IT becomes "single point of failure." While PC can run without any infrastructure or IT involvement.
That has been always the biggest argument against thin clients.
Plus few people actually want to work on the thin clients, so you rarely see any grass root support for them coming from employees. And "better PC" is often used by management as an incentive for subordinates too.
Example of Fallout is a good one.
Look at it for a moment from different prospective: how many books you have read which had many intervened story lines. How many you actually liked? Writing such book is by no mean an easy task.
Same applies to game scripts. The fact that Fallout had a flexible story structure, amendable by accomplished quests, doesn't mean it is easy to make one.
Rather the fact itself that Fallout is still remembered, highlights that it is rather difficult to do.
True. That's not the problem for the programmers.
But highly customizable characters are pain for story designers. In story player encounters the monsters with not some random stats - but the stats which reflect the role the monster plays in story.
Just recall the NWN1 where they had the magic button "Recommend". During character creation and level up screen, hitting "Recommend" was defaulting to properly leveled warrior and guaranteeing that player wouldn't hit an obstacle s/he can't overcome. Because playing custom character means that there might be monsters you can't defeat alone. And for some exotic classes there were even specialized modules, allowing you to play in full force, because default campaign was designed for a warrior. (Even playing as barbarian, due to its low persuasion, one would miss many interesting side quests.)
Essentially, exotic/custom classes increase game complexity on both sides. Making a campaign becomes more complicated as many classes has to be taken into account. Playing with a custom class requires quite a skill of knowing and expoloiting strengths and weaknesses - your own and monsters. That's not something average gamers might expect - many are way too used to bashing stuff with a sword or annihilating everything with a magic.
So in their view, a clean-room implementation of a libmysql work-alike would have been somehow subject to their GPL claims. That was the evil part.
Oh. That's new to me.
Honestly, though I was on the receiving end of the problem, the MySQL AB stance on that forced me (and my employer) to switch to Linux: MySQL's Windows up-to-date version was commercial. And I came to be very thankful to that very painful at the time decisions (painful == I was a principal windows software developer at the times).
I would like to have seen that challenged in court(s).
No need. IANAL but IMO the argument fails at the interoperability clauses, commonly found in the copyright law exceptions. IIRC even DMCA has such clause, allowing to break DRM in some cases for interoperability reasons. Essentially, copyright/etc can't be used to deny the users access to their information.
Oracle's bread and butter is support contracts, not license revenue.
That's right. And I'm already terrified with prospects that to deploy MySQL I would need the same amount of pains (and two/more certifications) as with Oracle.
MySQL is simple and slick RDBMS. In some part it caught in market because it it very easy to deploy, rather easy to develop for and quite easy to maintain.
Monty is a liar, Groklaw caught him in the lie and he shouldn't be trusted.
As a techie who built a RDBMS competing with Oracle, I would pay anytime Monty much more attention than some over-religious zealots over at Groklaw.
Because as a techie too, I understand that in real life it is not all that black and white. And depending on to whom you talk to, you pick different - often apparently (but only apparently) contradictory - arguments. Because you know different people would listen to different arguments.
He's abusing the EU approval process for personal gain.
No amount of ad nominem attacks would change the fact that Oracle can easily (and likely would) screw up whole commercial eco-system around MySQL. Monty now probably only a small part of it and has much less to loose compared to its other participants.
Small niche?
Compare the number of TV/FM/cellphone owners to the number of console/PC owners.
Reach of gaming industry pales in comparison.
P.S. "the biggest games are making more than movie releases" - picking at ailing movie industry is a bad taste.
That is nothing new. The problem is that Monty now found himself on the other side of fence and he is faced with the same choice as MySQL AB customers were in past: get a free GPLed MySQL fork or buy a license for a commercial MySQL variant.
GPL played the evil trick that you can't link commercial applications against libmysql*. IOW, to develop proprietary closed-source MySQL based product, you have to buy a license for the commercial fork of MySQL. And that is to my understanding the matter of his objection. And it is a rather valid objection, since Oracle now has a way to kill completely (not only Monty but) whole commercial infrastructure surrounding MySQL .
On one side I'm sadistically happy that Monty himself got the taste of it. On another side I also recognize that building something like MySQL completely open source might have been impossible and some revenue stream is much required. (Even much touted PostgreSQL, thanks to BSD license, has quite a number of proprietary applications around it.)
IMO both console and PC gaming are dying - in a figure of speech.
"Rise" of casual gaming now only further highlighted that gaming as many of us relate to it is a rather smallish niche of modern entertainment. For many people the idea that one has to spend hours and hours before screen (TV or PC display) only to learn controls - before s/he could enjoy a game - is simply absurd. E.g. I take my (now decade old) FPS skills for granted and even though I write this it is still hard for me to imagine why many people just can't "get it".
The point is that we do not need skills to enjoy movies or TV shows. And the learning divide, for as long as it would exist, would be making gaming business vulnerable.
Console gaming in that respect is even more vulnerable: PC gaming in a way is self-sustainable thanks to the fact that a gamer can become game developer rather easily. Console gaming on other side creates only consumers who depend completely on a business to provide entertainment to them.
Though thanks to Nintendo's WiiMote (and following it Sony's Wand and MS' Natal) I'd say console business has a pretty good shot at making games more accessible. As business and market, they are still pretty strong and still manage to come up with something new.
99% of me agrees with you wholeheartedly. It's rather useless.
Yet there is 1% of me which also knows that it is an integral part of science to make up some silly theories and models about stuff which we would never know for sure. After all, pretty much everything in the today's science started some long time ago from a silly theory. It was silly and unknowable in the past - while now it is treated as an fact.
The crucial difference is whether scientists do understand the shaky foundation or they foolishly insist on objectivity of their research.
Mods, how is this flame bait?
duh anything positive to m$ is flamebait
What is also right as to remind everybody with whom we are dealing here.
M$' "embrace and extend" (and extinguish) strategy is probably worst kept secret ever. If they would start taking competition with Google more seriously, I think user's privacy would end up in even greater jeopardy.