The problem with reversible computers is that they are liable to suddenly do a U-turn in the middle of a computation and instead start sucking all the intelligence out of your head, leaving you feeling completely dumb.
Actually if my wife is to be believed, this is how our computer works.
it'll take six months to a year before I feel comfortable pitching either of these to buisnesses.
That's just crazy talk. From a corporate perspective, the Red Hat decision and this Novell aquisition are the best possible things that could have happened to these distributions. In both cases there will be more resources to pour into R&D for limited-scope corporate server and desktop configurations.
The potential losers in either of these scenarios are the home users.
The distinction between GPL and Public Domain is hardly a moot one. It forms a significant part of SCO's legal case against IBM and the rest of the world (well, it did *last* week, anyway. Who knows what nonsense they'll come up with next week).
Remember that the GPL hasn't been properly tested in court yet. If it does ever get in front of a judge, you'd better hope that judge is a good one because a poorly informed decision that doesn't uphold the copyleft principle could blow the open source community apart. I believe there are many, many developers who would just give up releasing free software altogether if the GPL was invalidated.
What have Novell actually bought for their 200 mill?
Is it SuSE's products? To a small degree, I suppose, since they have a brand new retail distribution out there. But it won't make anywhere near $200 million. And the Enterprise products are about a year old, and they sell in much smaller numbers.
Is it SuSE's' technology? I doubt it. The value-added technology in a Linux distribution lasts only about four months I guess, since that is about the average lifetime for any package before a new version needs to be tweaked, tested, rpm'ed and put up for download. What about YaST? Well, how much does it cost to develop a configuration and package management tool? Not $200 million, that's for sure.
What they have actually bought is SuSE's customer base. And that's been the reason behind most major computer corporation aquisitions in the past decade, as the focus of the industry moves away from hardware as a profit centre (just as it now moves away from software too) towards support, consultancy and other services.
The question for dyed-in-the-wool SuSE users now is, what will be left of their favourite distribution after Novell have remoulded it in their own image? There are bound to be major changes to the character of the distribution, whatever happens.
Since they also own Ximian, I'd guess SuSE's traditional emphasis on KDE is likely to wither away. That would be a blow to the KDE developer community since SuSE have always been one of the few distributors to contribute binary rpms for download soon after each release.
Novell will surely want to focus on integrating their own Directory Services as a key component so support for LDAP and NIS/YP may suffer (I'm not suggesting they will stop offering those packages, just that less effort will go into their configuration and integration, and the packages may lag behind the latest version somewhat).
More worrying is the prospect that Novell may not share SuSE's sentimental attachment to their traditional non-corporate user base and they may cut down this remarkable distribution drastically in terms of the number of packages supported, losing many that are not seen as useful to business. Or they may even follow Red Hat's lead and drop the home desktop product altogether. I even think this is the most likely scenario since it is the home user business that has so conspicuously failed to turn a profit for *any* Linux distributor. That is, after all, most likely the reason why Novell are buying SuSE and not the other way round; SuSE has what is quite possibly the most impressive Linux distribution on the market, and yet they are *still* takeover fodder.
There's always the remote possibility that Novell plans to go head-to-head with Microsoft on the home desktop and will pump money into polishing up the "user experience" but they've made no mention of anything like this in their press release. Their stated intent is to concentrate on developing and marketing it as an enterprise platform.
If RH's abandonment of supported consumer releases teaches us anything, it's the importance of financially supporting our favorite distros and tolerating the bit rot of boxed sets.
I wholeheartedly agree. I fail to understand how it is that so many supposedly intelligent people seem to think it's wrong to pay for an open source OS. Every time a new one comes out there are numerous posts on Slashdot either moaning about the lack of free ISO's or complaining bitterly about the $50 price of the boxed set and boasting that they can get it for $5 at Cheapbytes or whatever. Do these people think that distributions just spring out of the ground fully formed?
If we can't support the distribution publishers with a few dollars now and again, then eventually they will all follow Red Hat and all we'll have left is the perpetually out-of-date Debian, and good old Slackware. If we're lucky.
This is hardly groundless doom-mongering. Just look at the cashflow problems Mandrake have had, and that's supposed to be one of the popular ones.
I'm hardly rolling in cash myself (As a contractor in the UK I get taxed to the hilt and I've been out of work for the last five months too) but I do at least understand the potential danger of the "tragedy of the commons". If it's going to come down to a choice between quality packaging and "roll your own distribution from sources", then even though I have 1 megabit broadband and can easly download the whole thing, I'm more than happy to stump up 50 notes each year for another pretty box. Given the voluntary aspect of this I kind of consider it a charitable donation.
Would that more people thought the same. But so many don't. And if it isn't software, it's music, or online movie downloads. Several people I know - neighbours, friends, work colleagues - on getting a modern computer and broadband internet installed for the first time, they came to me for advice, but all they wanted to know was "where is all the ripped off stuff I can steal?" I truly find this kind of attitude offensive. It's nothing short of parasitism.
Yeah but, when 9.1 comes out (and assuming SuSE don't follow Red Hat's lead and just can the consumer releases) what are the chances it'll be closely followed by the release of XFree86 4.3.1, or glibc 2.3, or a radically new and backwards-incompatible gcc compiler, or some such...
Mind you it was probably 5.1 that put me off. It was a dog.
Ever since then I've been using SuSE. Not that SuSE's distros are perfect and absent all inconsistency, but they do pretty well considering the phenomenal amount of software that you get in the box, ready-to-run. It's always been a *much* better value proposition than Red Hat.
Great. Just bloody great. I only just took delivery of SuSE 9.0 and now I have to upgrade half of it already. So begins the long decline into terminal bit rot.
Your final suggestion is correct: the way biological neurons encode memory in the formation of synaptic networks results in a dynamic model which bears little in common with silicon hardware. The most important difference is that unlike DRAM memory, in the course of normal operation synapses do degrade over time, and entire brain cells die.
Incidentally, you suggested that the brain's memory capacity itself is finite. Well sure it must be, but taken in isolation the maximum information density of vertebrate cortex is much higher than you appear to think and I don't think any human has ever approached that limit.
What I mean is, the neural architecture of the cortex is in principle capable of encoding much more information than has ever been achieved in practice because of limits on the dynamic processes working for and against memory retention. These are:
1) the tendency of a memory to decay over time (individual synapses deteriorate, whole brain cells die).
2) the ability to refresh memory, counteracting the above deterioration, through the act of recall (Long Term Potentiation, or LTP).
As you correctly pointed out, the more time one spends on collecting new memories, the less time one has to reminisce and refresh the old.
And, even if we devote our lives purely to remembering and rehearsing the memories we already collected, given that the rate of memory recall must be as finite as the rate of memory decay there has to be some upper limit to the amount of memory that can be maintained through constant and deliberate maintenance.
This upper limit can be calculated roughly if we know the decay rate of synaptic connections and the amount of time necessary to spend in the act of recall to restore them. This would yield an answer in terms of the number of synaptic connections that can be maintained. Divide that by the average number of synapses per memory (though that's likely to be arbitrary: just what constitutes a "memory" anyway - its rather a vague term) and the answer is the maximum number of memories that can be maintained.
So, the practical limit on neural memory capacity isn't so much like a bucket of fixed size which gets full and overflows; it's more like a very large bucket constantly being filled with a dribble of water from a tap, but which has a significant hole in the bottom.
The water level never gets above a relatively low level for most people because doing so would require the tap to be kept on full blast for extended periods of time (which would require a lot of effort) and also the tap's maximum flow rate isn't terribly high.
On the basis of that analogy you might conclude that if it were possible to beam memories directly into your mind, like with the "Teacher" device Dr McCoy used in that Star Trek episode "Spock's Brain", you could then suddenly fill the bucket right up and maintain a truly prodigious amount of memory by expending the same amount of effort as anybody else, because the rate of loss from the bucket is still the same.
However, that's where the analogy breaks down. If you have a memory ten times as full then you have ten times the rate of synapse decay and you need to spend ten times the effort replacing decayed synapses. Individual synapses will decay just as frequently and the total rate of loss will be proportional to the total number of synapses. So it's more like the fuller the bucket gets, the bigger the hole in the bottom gets. The bucket has a thin rubber bottom with a *stretchy* hole.
Consequently the number of retrievable memories in that big full brain would decline until reaching whatever maximum level is sustainable on the owner's usual memory maintenance schedule.
If you want to expand human memory capacity then there are two ways.
The first is commonly practiced already: building synaptic networks that take less work to maintain through being more densely interconnected. This is what happens when you learn a subject well! Up to a certain point when you are lea
Well I'd love to have read that article but unfortunately the myriad flashing gifs blink tags and huge animated flash ads right at top centre of each page made me feel dizzy. Hell, it even made my browser sick. I've left the page already and my browser still hasn't recovered. What kind of idiot thinks that sort of page design is OK?
Yes indeed, and if the US had gone in and taken Baghdad back then, it would at least have been just, and many more people would have supported it. But that was then and this is now. The last gulf war just wasn't supported by such logic, it was all too clearly done for the monetary gain of US corporations and the richest and most powerful American families.
Speaking from a neuroanatomical standpoint I don't know that such dumping would occur. Not in such stark black and white terms, anyway. Surely a few vivid memories from one's youth would survive at least.
I believe memory persists mainly through recall and rehearsal; as one tends to rehearse recent memories more, and remote memories less and less, the old memories simply fade while fresher ones are added. In which case it's not "capacity" (as so famously outlined by Conan Doyle) as such that underlies the memory problem, but merely the lack of persistence of memory engrams. If some biological or biochemical means could be found of preventing these from fading? Unlike Sherlock Holmes, I don't believe that the attic would ever be full; better memory retention wouldn't prevent the brain of a centenarian from continuing to accumulate new memories with the same ease as a thirty year old.
Even if, as we've both noted, memory deposition is lower in that thirty year old than in a twenty-five year old. The difference might not after all signify a progressive decline throughout one's lifetime: it might just be due to a fairly rapid switch from juvenile brain mode to adult brain mode, more or less complete by the time one is forty.
Kim Stanley Robinson wrote about this in the last book of his Mars Trilogy "Blue Mars" and in some of the additional short stories published elsewhere which served as postscripts to the trilogy.
One of the most disturbing aspects of his take on extreme old age was his assumption that by the time you were a couple of hundred years old you would have completely forgotten *all* of your childhood and even most of their first century of adulthood, consigning all memories of what we turn-of-the-millennium humans would think of as a normal human lifespan to the dustbin.
Also, although medical advances and a healthy lifestyle could enable one to retain one's health and a degree of fitness, it couldn't completely preserve the elasticity of skin or prevent all bone erosion so that you would end up spending most of your life *looking* kind of turtle-like. This was therefore the default appearance of humans in KSR's vision of utopia.
I want to add something to that. My own experience of ageing tells me that the brains of the young and the old are very different indeed. In general, the young are creative and adaptable. The mature brain is much less agile but benefits from a store house of experience not accessible to the young. So while young people are generally much more conscious than older people in the sense that they spend more of their time having genuinely original and creative thoughts, older people are more like extremely sophisticated robots.
It's not as devastating as it sounds. For most tasks, including many very sophisticated tasks such as in various fields of engineering, experience trumps creativity.
The rate at which the creative gives way to the robotic is obviously going to vary widely from person to person. But I'll bet most people can see the truth in what I'm saying when they consider either their own declining mental agility vs. their increasing experience and their tendency to sink into habit as they gets older, or the increasing tendency of their parents to repeat the same increasingly limited behaviour patterns, speeches and anecdotes, blissfully unaware that the needle is stuck in a groove.
Now what was I saying? Oh yeah...Kim Stanley Robinson wrote about this in the last book of his...what? What the hell are you laughing at?
Slashdot Mirror
That's a godo story. You could even enter it in a short story competition.
does everybody lie about what they're doing on the Internet?"
Ye... er, no!
The problem with reversible computers is that they are liable to suddenly do a U-turn in the middle of a computation and instead start sucking all the intelligence out of your head, leaving you feeling completely dumb.
Actually if my wife is to be believed, this is how our computer works.
The goatse guy could do that scene...
Yeah the view is fucked up, all right. But it's hardly js7a's fault.
That's just crazy talk. From a corporate perspective, the Red Hat decision and this Novell aquisition are the best possible things that could have happened to these distributions. In both cases there will be more resources to pour into R&D for limited-scope corporate server and desktop configurations.
The potential losers in either of these scenarios are the home users.
The distinction between GPL and Public Domain is hardly a moot one. It forms a significant part of SCO's legal case against IBM and the rest of the world (well, it did *last* week, anyway. Who knows what nonsense they'll come up with next week).
Remember that the GPL hasn't been properly tested in court yet. If it does ever get in front of a judge, you'd better hope that judge is a good one because a poorly informed decision that doesn't uphold the copyleft principle could blow the open source community apart. I believe there are many, many developers who would just give up releasing free software altogether if the GPL was invalidated.
What have Novell actually bought for their 200 mill?
Is it SuSE's products? To a small degree, I suppose, since they have a brand new retail distribution out there. But it won't make anywhere near $200 million. And the Enterprise products are about a year old, and they sell in much smaller numbers.
Is it SuSE's' technology? I doubt it. The value-added technology in a Linux distribution lasts only about four months I guess, since that is about the average lifetime for any package before a new version needs to be tweaked, tested, rpm'ed and put up for download. What about YaST? Well, how much does it cost to develop a configuration and package management tool? Not $200 million, that's for sure.
What they have actually bought is SuSE's customer base. And that's been the reason behind most major computer corporation aquisitions in the past decade, as the focus of the industry moves away from hardware as a profit centre (just as it now moves away from software too) towards support, consultancy and other services.
The question for dyed-in-the-wool SuSE users now is, what will be left of their favourite distribution after Novell have remoulded it in their own image? There are bound to be major changes to the character of the distribution, whatever happens.
Since they also own Ximian, I'd guess SuSE's traditional emphasis on KDE is likely to wither away. That would be a blow to the KDE developer community since SuSE have always been one of the few distributors to contribute binary rpms for download soon after each release.
Novell will surely want to focus on integrating their own Directory Services as a key component so support for LDAP and NIS/YP may suffer (I'm not suggesting they will stop offering those packages, just that less effort will go into their configuration and integration, and the packages may lag behind the latest version somewhat).
More worrying is the prospect that Novell may not share SuSE's sentimental attachment to their traditional non-corporate user base and they may cut down this remarkable distribution drastically in terms of the number of packages supported, losing many that are not seen as useful to business. Or they may even follow Red Hat's lead and drop the home desktop product altogether. I even think this is the most likely scenario since it is the home user business that has so conspicuously failed to turn a profit for *any* Linux distributor. That is, after all, most likely the reason why Novell are buying SuSE and not the other way round; SuSE has what is quite possibly the most impressive Linux distribution on the market, and yet they are *still* takeover fodder.
There's always the remote possibility that Novell plans to go head-to-head with Microsoft on the home desktop and will pump money into polishing up the "user experience" but they've made no mention of anything like this in their press release. Their stated intent is to concentrate on developing and marketing it as an enterprise platform.
I wholeheartedly agree. I fail to understand how it is that so many supposedly intelligent people seem to think it's wrong to pay for an open source OS. Every time a new one comes out there are numerous posts on Slashdot either moaning about the lack of free ISO's or complaining bitterly about the $50 price of the boxed set and boasting that they can get it for $5 at Cheapbytes or whatever. Do these people think that distributions just spring out of the ground fully formed?
If we can't support the distribution publishers with a few dollars now and again, then eventually they will all follow Red Hat and all we'll have left is the perpetually out-of-date Debian, and good old Slackware. If we're lucky.
This is hardly groundless doom-mongering. Just look at the cashflow problems Mandrake have had, and that's supposed to be one of the popular ones.
I'm hardly rolling in cash myself (As a contractor in the UK I get taxed to the hilt and I've been out of work for the last five months too) but I do at least understand the potential danger of the "tragedy of the commons". If it's going to come down to a choice between quality packaging and "roll your own distribution from sources", then even though I have 1 megabit broadband and can easly download the whole thing, I'm more than happy to stump up 50 notes each year for another pretty box. Given the voluntary aspect of this I kind of consider it a charitable donation.
Would that more people thought the same. But so many don't. And if it isn't software, it's music, or online movie downloads. Several people I know - neighbours, friends, work colleagues - on getting a modern computer and broadband internet installed for the first time, they came to me for advice, but all they wanted to know was "where is all the ripped off stuff I can steal?" I truly find this kind of attitude offensive. It's nothing short of parasitism.
Unless you're in an episode of Star Trek.
Yeah but, when 9.1 comes out (and assuming SuSE don't follow Red Hat's lead and just can the consumer releases) what are the chances it'll be closely followed by the release of XFree86 4.3.1, or glibc 2.3, or a radically new and backwards-incompatible gcc compiler, or some such...
Mind you it was probably 5.1 that put me off. It was a dog.
Ever since then I've been using SuSE. Not that SuSE's distros are perfect and absent all inconsistency, but they do pretty well considering the phenomenal amount of software that you get in the box, ready-to-run. It's always been a *much* better value proposition than Red Hat.
Great. Just bloody great. I only just took delivery of SuSE 9.0 and now I have to upgrade half of it already. So begins the long decline into terminal bit rot.
*sigh*.
Your final suggestion is correct: the way biological neurons encode memory in the formation of synaptic networks results in a dynamic model which bears little in common with silicon hardware. The most important difference is that unlike DRAM memory, in the course of normal operation synapses do degrade over time, and entire brain cells die.
Incidentally, you suggested that the brain's memory capacity itself is finite. Well sure it must be, but taken in isolation the maximum information density of vertebrate cortex is much higher than you appear to think and I don't think any human has ever approached that limit.
What I mean is, the neural architecture of the cortex is in principle capable of encoding much more information than has ever been achieved in practice because of limits on the dynamic processes working for and against memory retention. These are:
1) the tendency of a memory to decay over time (individual synapses deteriorate, whole brain cells die).
2) the ability to refresh memory, counteracting the above deterioration, through the act of recall (Long Term Potentiation, or LTP).
As you correctly pointed out, the more time one spends on collecting new memories, the less time one has to reminisce and refresh the old.
And, even if we devote our lives purely to remembering and rehearsing the memories we already collected, given that the rate of memory recall must be as finite as the rate of memory decay there has to be some upper limit to the amount of memory that can be maintained through constant and deliberate maintenance.
This upper limit can be calculated roughly if we know the decay rate of synaptic connections and the amount of time necessary to spend in the act of recall to restore them. This would yield an answer in terms of the number of synaptic connections that can be maintained. Divide that by the average number of synapses per memory (though that's likely to be arbitrary: just what constitutes a "memory" anyway - its rather a vague term) and the answer is the maximum number of memories that can be maintained.
So, the practical limit on neural memory capacity isn't so much like a bucket of fixed size which gets full and overflows; it's more like a very large bucket constantly being filled with a dribble of water from a tap, but which has a significant hole in the bottom.
The water level never gets above a relatively low level for most people because doing so would require the tap to be kept on full blast for extended periods of time (which would require a lot of effort) and also the tap's maximum flow rate isn't terribly high.
On the basis of that analogy you might conclude that if it were possible to beam memories directly into your mind, like with the "Teacher" device Dr McCoy used in that Star Trek episode "Spock's Brain", you could then suddenly fill the bucket right up and maintain a truly prodigious amount of memory by expending the same amount of effort as anybody else, because the rate of loss from the bucket is still the same.
However, that's where the analogy breaks down. If you have a memory ten times as full then you have ten times the rate of synapse decay and you need to spend ten times the effort replacing decayed synapses. Individual synapses will decay just as frequently and the total rate of loss will be proportional to the total number of synapses. So it's more like the fuller the bucket gets, the bigger the hole in the bottom gets. The bucket has a thin rubber bottom with a *stretchy* hole.
Consequently the number of retrievable memories in that big full brain would decline until reaching whatever maximum level is sustainable on the owner's usual memory maintenance schedule.
If you want to expand human memory capacity then there are two ways.
The first is commonly practiced already: building synaptic networks that take less work to maintain through being more densely interconnected. This is what happens when you learn a subject well! Up to a certain point when you are lea
They already said what they wanted it for. To generate wealth for themselves and their employees. In other words, they just want to be rich.
Yes, it's called "selling out". Nothing to see here (any more), move along, go back to your homes.
Back to your soon-to-be-useless internet connection.
Well I'd love to have read that article but unfortunately the myriad flashing gifs blink tags and huge animated flash ads right at top centre of each page made me feel dizzy. Hell, it even made my browser sick. I've left the page already and my browser still hasn't recovered. What kind of idiot thinks that sort of page design is OK?
Lay off. At least he didn't write "virii".
Yes indeed, and if the US had gone in and taken Baghdad back then, it would at least have been just, and many more people would have supported it. But that was then and this is now. The last gulf war just wasn't supported by such logic, it was all too clearly done for the monetary gain of US corporations and the richest and most powerful American families.
I'm running a vanilla mozilla 1.5 build from SuSE on my old SuSE 8.1 box and the applet worked fine for me. I didn't have to change a thing.
Speaking from a neuroanatomical standpoint I don't know that such dumping would occur. Not in such stark black and white terms, anyway. Surely a few vivid memories from one's youth would survive at least.
I believe memory persists mainly through recall and rehearsal; as one tends to rehearse recent memories more, and remote memories less and less, the old memories simply fade while fresher ones are added. In which case it's not "capacity" (as so famously outlined by Conan Doyle) as such that underlies the memory problem, but merely the lack of persistence of memory engrams. If some biological or biochemical means could be found of preventing these from fading? Unlike Sherlock Holmes, I don't believe that the attic would ever be full; better memory retention wouldn't prevent the brain of a centenarian from continuing to accumulate new memories with the same ease as a thirty year old.
Even if, as we've both noted, memory deposition is lower in that thirty year old than in a twenty-five year old. The difference might not after all signify a progressive decline throughout one's lifetime: it might just be due to a fairly rapid switch from juvenile brain mode to adult brain mode, more or less complete by the time one is forty.
I hope!
I think he was trying to spell "serious" but he was typing with his feet.
That's because it's in German, you insensitive clod!
Kim Stanley Robinson wrote about this in the last book of his Mars Trilogy "Blue Mars" and in some of the additional short stories published elsewhere which served as postscripts to the trilogy.
One of the most disturbing aspects of his take on extreme old age was his assumption that by the time you were a couple of hundred years old you would have completely forgotten *all* of your childhood and even most of their first century of adulthood, consigning all memories of what we turn-of-the-millennium humans would think of as a normal human lifespan to the dustbin.
Also, although medical advances and a healthy lifestyle could enable one to retain one's health and a degree of fitness, it couldn't completely preserve the elasticity of skin or prevent all bone erosion so that you would end up spending most of your life *looking* kind of turtle-like. This was therefore the default appearance of humans in KSR's vision of utopia.
I want to add something to that. My own experience of ageing tells me that the brains of the young and the old are very different indeed. In general, the young are creative and adaptable. The mature brain is much less agile but benefits from a store house of experience not accessible to the young. So while young people are generally much more conscious than older people in the sense that they spend more of their time having genuinely original and creative thoughts, older people are more like extremely sophisticated robots.
It's not as devastating as it sounds. For most tasks, including many very sophisticated tasks such as in various fields of engineering, experience trumps creativity.
The rate at which the creative gives way to the robotic is obviously going to vary widely from person to person. But I'll bet most people can see the truth in what I'm saying when they consider either their own declining mental agility vs. their increasing experience and their tendency to sink into habit as they gets older, or the increasing tendency of their parents to repeat the same increasingly limited behaviour patterns, speeches and anecdotes, blissfully unaware that the needle is stuck in a groove.
Now what was I saying? Oh yeah...Kim Stanley Robinson wrote about this in the last book of his...what? What the hell are you laughing at?
I've been upgrading every time a new version comes out - I'm on 1.5 now. It's no better.
Thanks, I'll try that before I put 9.0 on.