Right, so a MMO with >1 million players (200,000 online at the same time) would require 200,000 graphics cards at the server end, which would be rendering the game and converting that into a video stream, hopefully at least a 20mbps HD stream to be playable and worth using. So we're talking about a 4 TERABIT network connection from this super computer. We're talking about venting 16MW of heat just from the GPUs, or probably 80MW in total from the servers.
Me? I suggest that the player buys their own graphics card, and that the servers just stay in the world of mathematics - the servers know where the users are (or can interpolate mostly accurately), the algorithms for physics and deformation are known - just send a message to the clients saying "Bunker Bomb exploded at x,y,z with force n".
It doesn't solve the problem of the user with GMA950 graphics and a 1.6GHz Atom, but that would probably have problems decoding ta 20mbps video stream anyway.
The problem with the Ribbon is that each Ribbon (Ribbon-Tab?) is Function Oriented, rather than Task Oriented. Each Ribbon is effectively a fancy GUI version of a pull down menu for that function.
The Flex example in the article appears to be Task Oriented, so even though it may seem to have some Ribbon-like features in it, it actually could turn out to be quite different. It will be interesting to see how it finally gets implemented, I suspect it will end up Function Oriented because it's easier to implement.
Build a load near London near a tube line. They'll be £200,000 within months, even for the 300 square foot plans.
However these houses/apartments should be sold as plans for social housing in first world countries. In the UK there is too little social housing, so these people are put in expensive private rental properties - costing the country a huge amount in benefits. These people should be housed, etc, but not in such places - if they're not working, they can deal with a cheaper out of town location with a bus stop. They can deal with a bare minimum floor plan. Simples.
I think when you read through all the comments, there is a common theme:
1) Buses are suckingly poor 2) Trains are good 3) Poor public transport integration makes trains fail from a user perspective
For (3) I mean that you need large, cheap, accessible car parks at train stations for the train to be desirable.
Don't expect people to get a bus to the train station, because buses suck, they're late, they get caught in the same traffic snarl-ups that cars get caught in, and they often don't go to the train station!
Cambridge, UK: They've turned a (disused) train line into a "guided bus". This at least gets rid of the problem with busy road... until the bus comes off the route into Cambridge itself, and hits all the busy traffic. If they had kept the train line, built a couple more stations, they would have had a direct link to the train station, and hence London. They have built a lot of park and rides around Cambridge however, again with the problem of buses on busy roads and non-integration with the train station. FFS! The problem is that clueless public servants don't understand other people's needs, they only see their need to go shopping in town on a Saturday morning.
I think we're all agreed (including the article) that counting sales / downloads / activations is clearly not a good method, it doesn't measure actual usage, or replacements, multiple installs for free OSes, etc.
Therefore measuring web accesses across a wide variety of websites seems reasonable, at least for measuring web users. However what if you want to count home users only? What about professional non-web use (all those people with Linux or BSD servers that are on the home/work network but rarely go online).
I think that at some point you accept that you might not get an accurate count, but if you can get within 1% of reality then you have a useful metric. If the web tracking uses a vast wide spread of websites to gather statistics then you can probably get a reasonably accurate snapshot of web users. Maybe some error bars on the results would make the data more understandable.
(oddly enough my answer to you was meant to be a response to myself higher up but I must have cocked up)
1) He mentions a marketshare statistic that uses web share to estimate usage share of desktop machines. Not downloads (that might be used multiple times, or not at all). So why diss downloads immediately? The questions should be "Do Linux users visit the sites that contribute to these statistics more or less than Windows users? Do Linux users use the web more or less than Windows users?" and so on.
2) Android runs on Linux, the Linux kernel. But Android actually runs within its own environment that uses a custom VM that borrows heavily from Java, and very little of the Linux system underneath is available. What share of the desktop does Gnome or KDE have? That's the equivalent (and arguably in terms of desktop use, a more viable question that what share of the market does Linux have). I think it is right that the Android figures are accounted for separately.
Munging together servers and clients is a pointless benchmark. Linux could have 30% of the server ecosystem, but that would make a 0.001% indent on client share.
Regardless, 1.02% is a far cry from 5 or 6 percent, never mind 10%. Who would even say that a Linux machine makes up 1 in 10 machines on the web, haven't they seen all the Windows machines, all the business machines, etc?
I don't think it's a problem that the game used a text mode - in 1975, given the price of memory, any real graphical mode would have been very expensive to implement, even in black and white.
It's just that by using a text mode, it's not really a graphical game. However how different is a text mode from a graphical tiled mode, e.g., the C64. Redefinability is key, and arbitrary positioning capability (on the C64 this was via sprites). The graphical hardware on this machine couldn't do that (it appears). On the other hand, it could cope with character-mapped Snake I expect. Or you could drop in a ROM with game-specific characters.
I also fail to see why this is being called a graphical game, when it is clearly a character-mapped display. Maybe if the characters were redefinable in software (or even dropping in a ROM with game-specific symbols), but seeing as they were using standard letters it looks like they weren't.
This is meant to be emulating 70s CRT *televisions*. Not computer displays (CGA, etc) which even in CRT form in the early 80s had a lot tighter phosphor mask (or whatever the term is) for crisp 80 column text display.
Like that person talking about their (for the time good quality) Amiga 1080 monitor below connected via an RGB connection to their modern console remake... not a 70s television connected via composite with bleeding from other signals.
Yeah, because when I played computer games as a kid in the 80s, I played it on a brand new late-90s CRT display. Not an 8 year old hand-me-down 14" TV in the bedroom (albeit mine was a 14" Trinitron which was surprisingly competent and not as grainy as this emulator, and even good for the old Amiga). Then again they're emulating a late 70s CRT.
If there is something this emulator doesn't do, it doesn't emulate what the poor signal quality on the cheap cable between the console and the TV that would create artifacts like shadowing. Nor does it emulate that CRT weirdness where the image is bright then dark, where it wobbles too dark, then a little too bright before fully changing to the correct colour.
The interesting thing comes with retro-game writers (who write games for the old machines, today) and the graphical styles of the games, which due to being designed and written in emulators on LCD monitors have changed. Old games used to stipple a lot to simulate shades of colours between what the hardware could actually achieve, whilst the newer games seem to have a more flat colour scheme - arguably this could be because the LCDs make the stippling look awful, whereas the CRT would make it look blended.
Yeah, really, because we all liked straight lines that were always slightly curved, squares that weren't, huge hefty monitors that barely exceeded 19" (I'd like to see you hefting a 30" CRT monitor around or finding the space to put a keyboard in front of it on your average desk). CRT made a shit input signal (either the signal being crap, or the source being primitive like early consoles) look better by its blurry glowy nature. We don't need that now.
People have a right to reasonable privacy in their gardens, not just within their houses. A drone that is effectively infra-red (and other wavelengths?) CCTV that can also see into back gardens is surely an invasion of the privacy of the innocent people that it will also record.
I agree that it's just flying CCTV, Half Life 2 style, to keep the masses in check. Even if it starts out being used at high altitudes for this purpose, someone will abuse it down the line to get detailed imagery of people where they expect reasonable privacy.
That's where this hair based sheeting (why not just use the hair from highland cattle?) and other ground cover works - it's permeable.
Plastic + Wax Paper sounds like a recipe for not letting water get to the ground, and hence the roots. Might be good as a base for your driveway/patio/shed floor.
Well for showing tabular data, HTML tables are what you should use. I think a lot of people hear "don't use tables" and don't hear the "for layout" at the end.
CSS layouts are actually quite easy until you want a non-trivial layout (there are some gotchas that have simple answers, and there are cross-browser compatibility issues with IE6 usually), and the latter is unlikely to happen for a person who isn't doing web design as the primary role. It's worth investing a few hours to learn it, it makes the HTML so much quicker to write, and easier to read.
I remember buying my Radeon 9500 when it first came out because it was the cheap option at the time - it meant I could play then-current games at medium resolution and medium settings. It cost twice as much as this card that can play most current games at very high resolution and high settings. Even two years ago the 9600GT upon release couldn't achieve that. Whilst TSMC's 40nm process isn't the best, it allows for great die shrinkage and hence for a competitive price. This is definitely the best $100 card upon release given the state of the games at the time for a long time, if not ever.
Wow, it's not 2001 anymore. ATI/AMD have monthly driver releases, you very rarely hear about issues on the tech websites, and they're opening up the hardware specifications for open source drivers, which will take time to arrive but at least it's a good move for people who want an open source only desktop.
Slashdot Mirror
Right, so a MMO with >1 million players (200,000 online at the same time) would require 200,000 graphics cards at the server end, which would be rendering the game and converting that into a video stream, hopefully at least a 20mbps HD stream to be playable and worth using. So we're talking about a 4 TERABIT network connection from this super computer. We're talking about venting 16MW of heat just from the GPUs, or probably 80MW in total from the servers.
Me? I suggest that the player buys their own graphics card, and that the servers just stay in the world of mathematics - the servers know where the users are (or can interpolate mostly accurately), the algorithms for physics and deformation are known - just send a message to the clients saying "Bunker Bomb exploded at x,y,z with force n".
It doesn't solve the problem of the user with GMA950 graphics and a 1.6GHz Atom, but that would probably have problems decoding ta 20mbps video stream anyway.
It was the TLB bug in the Barcelona Opterons / Phenom 1s I imagine. Trivial to workaround, but with a minor performance hit.
The problem with the Ribbon is that each Ribbon (Ribbon-Tab?) is Function Oriented, rather than Task Oriented. Each Ribbon is effectively a fancy GUI version of a pull down menu for that function.
The Flex example in the article appears to be Task Oriented, so even though it may seem to have some Ribbon-like features in it, it actually could turn out to be quite different. It will be interesting to see how it finally gets implemented, I suspect it will end up Function Oriented because it's easier to implement.
Sharing one wall is semi-detached.
If you share two walls it's a terrace.
Most houses in England are semi-detached or terraced.
Build a load near London near a tube line. They'll be £200,000 within months, even for the 300 square foot plans.
However these houses/apartments should be sold as plans for social housing in first world countries. In the UK there is too little social housing, so these people are put in expensive private rental properties - costing the country a huge amount in benefits. These people should be housed, etc, but not in such places - if they're not working, they can deal with a cheaper out of town location with a bus stop. They can deal with a bare minimum floor plan. Simples.
I think when you read through all the comments, there is a common theme:
1) Buses are suckingly poor
2) Trains are good
3) Poor public transport integration makes trains fail from a user perspective
For (3) I mean that you need large, cheap, accessible car parks at train stations for the train to be desirable.
Don't expect people to get a bus to the train station, because buses suck, they're late, they get caught in the same traffic snarl-ups that cars get caught in, and they often don't go to the train station!
Cambridge, UK: They've turned a (disused) train line into a "guided bus". This at least gets rid of the problem with busy road ... until the bus comes off the route into Cambridge itself, and hits all the busy traffic. If they had kept the train line, built a couple more stations, they would have had a direct link to the train station, and hence London. They have built a lot of park and rides around Cambridge however, again with the problem of buses on busy roads and non-integration with the train station. FFS! The problem is that clueless public servants don't understand other people's needs, they only see their need to go shopping in town on a Saturday morning.
Same here - Virgin's Cable service has been overall very good for the past 8 years that I have had it.
Yes, there have been incidents.
Yes, their TV service and boxes are shittier than a sewage farm on a hot sunny day.
But I've always got the advertised speeds.
Damn right. It's "fishies". I saw it on Red Dwarf: "Come here little fishies".
Is it some kind of virtualisation OS that runs over multiple machines, hence making data centre deployments easy?
Certainly nothing to do with LEGO which are little plastic bricks, that aren't good for halon delivery systems.
I think we're all agreed (including the article) that counting sales / downloads / activations is clearly not a good method, it doesn't measure actual usage, or replacements, multiple installs for free OSes, etc.
Therefore measuring web accesses across a wide variety of websites seems reasonable, at least for measuring web users. However what if you want to count home users only? What about professional non-web use (all those people with Linux or BSD servers that are on the home/work network but rarely go online).
I think that at some point you accept that you might not get an accurate count, but if you can get within 1% of reality then you have a useful metric. If the web tracking uses a vast wide spread of websites to gather statistics then you can probably get a reasonably accurate snapshot of web users. Maybe some error bars on the results would make the data more understandable.
(oddly enough my answer to you was meant to be a response to myself higher up but I must have cocked up)
I guess Taco Bell would qualify as a "reinforced rubber product".
Right, so now I'm reading the article.
1) He mentions a marketshare statistic that uses web share to estimate usage share of desktop machines. Not downloads (that might be used multiple times, or not at all). So why diss downloads immediately? The questions should be "Do Linux users visit the sites that contribute to these statistics more or less than Windows users? Do Linux users use the web more or less than Windows users?" and so on.
2) Android runs on Linux, the Linux kernel. But Android actually runs within its own environment that uses a custom VM that borrows heavily from Java, and very little of the Linux system underneath is available. What share of the desktop does Gnome or KDE have? That's the equivalent (and arguably in terms of desktop use, a more viable question that what share of the market does Linux have). I think it is right that the Android figures are accounted for separately.
It's just tipped above 1% for consumer systems that are used for internet usage. http://techreport.com/discussions.x/16860
Munging together servers and clients is a pointless benchmark. Linux could have 30% of the server ecosystem, but that would make a 0.001% indent on client share.
Regardless, 1.02% is a far cry from 5 or 6 percent, never mind 10%. Who would even say that a Linux machine makes up 1 in 10 machines on the web, haven't they seen all the Windows machines, all the business machines, etc?
I don't think it's a problem that the game used a text mode - in 1975, given the price of memory, any real graphical mode would have been very expensive to implement, even in black and white.
It's just that by using a text mode, it's not really a graphical game. However how different is a text mode from a graphical tiled mode, e.g., the C64. Redefinability is key, and arbitrary positioning capability (on the C64 this was via sprites). The graphical hardware on this machine couldn't do that (it appears). On the other hand, it could cope with character-mapped Snake I expect. Or you could drop in a ROM with game-specific characters.
For 1975 I'll let it slide.
I also fail to see why this is being called a graphical game, when it is clearly a character-mapped display. Maybe if the characters were redefinable in software (or even dropping in a ROM with game-specific symbols), but seeing as they were using standard letters it looks like they weren't.
Is that Sun using SPARC or x86? Can it also predict the future?
Makes me want to get a parrot, but I don't think it would mix well with my cat.
This is meant to be emulating 70s CRT *televisions*. Not computer displays (CGA, etc) which even in CRT form in the early 80s had a lot tighter phosphor mask (or whatever the term is) for crisp 80 column text display.
Like that person talking about their (for the time good quality) Amiga 1080 monitor below connected via an RGB connection to their modern console remake... not a 70s television connected via composite with bleeding from other signals.
Yeah, because when I played computer games as a kid in the 80s, I played it on a brand new late-90s CRT display. Not an 8 year old hand-me-down 14" TV in the bedroom (albeit mine was a 14" Trinitron which was surprisingly competent and not as grainy as this emulator, and even good for the old Amiga). Then again they're emulating a late 70s CRT.
If there is something this emulator doesn't do, it doesn't emulate what the poor signal quality on the cheap cable between the console and the TV that would create artifacts like shadowing. Nor does it emulate that CRT weirdness where the image is bright then dark, where it wobbles too dark, then a little too bright before fully changing to the correct colour.
The interesting thing comes with retro-game writers (who write games for the old machines, today) and the graphical styles of the games, which due to being designed and written in emulators on LCD monitors have changed. Old games used to stipple a lot to simulate shades of colours between what the hardware could actually achieve, whilst the newer games seem to have a more flat colour scheme - arguably this could be because the LCDs make the stippling look awful, whereas the CRT would make it look blended.
Yeah, really, because we all liked straight lines that were always slightly curved, squares that weren't, huge hefty monitors that barely exceeded 19" (I'd like to see you hefting a 30" CRT monitor around or finding the space to put a keyboard in front of it on your average desk). CRT made a shit input signal (either the signal being crap, or the source being primitive like early consoles) look better by its blurry glowy nature. We don't need that now.
People have a right to reasonable privacy in their gardens, not just within their houses. A drone that is effectively infra-red (and other wavelengths?) CCTV that can also see into back gardens is surely an invasion of the privacy of the innocent people that it will also record.
I agree that it's just flying CCTV, Half Life 2 style, to keep the masses in check. Even if it starts out being used at high altitudes for this purpose, someone will abuse it down the line to get detailed imagery of people where they expect reasonable privacy.
You also want water to get through.
That's where this hair based sheeting (why not just use the hair from highland cattle?) and other ground cover works - it's permeable.
Plastic + Wax Paper sounds like a recipe for not letting water get to the ground, and hence the roots. Might be good as a base for your driveway/patio/shed floor.
Well for showing tabular data, HTML tables are what you should use. I think a lot of people hear "don't use tables" and don't hear the "for layout" at the end.
CSS layouts are actually quite easy until you want a non-trivial layout (there are some gotchas that have simple answers, and there are cross-browser compatibility issues with IE6 usually), and the latter is unlikely to happen for a person who isn't doing web design as the primary role. It's worth investing a few hours to learn it, it makes the HTML so much quicker to write, and easier to read.
I remember buying my Radeon 9500 when it first came out because it was the cheap option at the time - it meant I could play then-current games at medium resolution and medium settings. It cost twice as much as this card that can play most current games at very high resolution and high settings. Even two years ago the 9600GT upon release couldn't achieve that. Whilst TSMC's 40nm process isn't the best, it allows for great die shrinkage and hence for a competitive price. This is definitely the best $100 card upon release given the state of the games at the time for a long time, if not ever.
Wow, it's not 2001 anymore. ATI/AMD have monthly driver releases, you very rarely hear about issues on the tech websites, and they're opening up the hardware specifications for open source drivers, which will take time to arrive but at least it's a good move for people who want an open source only desktop.