I'm 34 - and have also worked in software my entire life. I've a CS degree and I make regular professional use of material I learned.
I consider myself lucky never to have worked in the military, and I've managed to avoid working for banks so far, too. I prefer not to think of what I do as IT, either - that conjures visions of installing Microsoft products and fixing printers.... there's little need for an academic qualification in that field.
I'm amused that you assumed I was criticising your standing - rather than asking about the seniority of the qualified people to which you purjoratively referred. While I've no evidence to suggest that qualified people are more competent or pleasant, I've none to the contrary either. I have encountered circumstances where over-qualified people feel undervalued - and wonder if this might explain. My suspicion, now, is that, possibly subconsciously, you're anxious about your own credentials - and this colours your views on the 'educated' demographic.
I can only wish you all the best, and comment that - in my experience - acquiring a degree does not make someone less affable, competent or capable than they otherwise were.
I needed to make room to upgrade, because the 4 Gb SSD in the EEE was close to full.
Ditto... I spannered my 9.04 and then decided to swap the 16GB and 4GB partitions - since I'll likely only ever need a few megs for my own documents. It took several attempts - but, eventually it worked... (and I'm not sure what went wrong when the re-install failed... it might have been my fault)
That said, I'm fairly impressed. I love PAN support for 3G via a bluetooth phone in NetworkManager - but I am frustrated that VPN support seems intermittently buggy. I've had to start my VPN up-to a dozen times... it's not clear what's the problem - but there's an open ticket that seems to describe my issue.
Very impressed with the release in general...
I'm curious about "Ubuntu One"... on the surface it looks incredibly useful - but, conversely, I definitely do not want to put my personal documents (perhaps including banking details, say) into the cloud unencrypted.
I've done some careful thinking about a related situation recently, and I've come to a conclusion that might, at first, sound surprising. It is sure to be contested by many...
I think you're looking at the situation all wrong... it isn't something intrinsic about people or the public perception of them that defines popularity... quite the opposite. What is most important is how people see others.
One extreme has internalised that others deserve respect and assumes that they are likely intelligent and decent people - entitled to whatever views they hold. The other extreme assumes others are likely fundamentally stupid selfish people - a potential threat where it is best to put in the metaphorical boot first.
I think this dimension to personality best explains popular cliques. Far from having established the 'best' friends, it is more likely high anxiety and fear of exclusion that binds the 'popular' groups. The 'popular' are more likely to hold others in contempt - and assume others "stupid". The unpopular are more likely self-confident and and making the error that they over estimate others' intelligence relative to their own faculties.
I think it's all about what you think others (should/do) think of you that defines how you are perceived. If you show contempt for the unpopular, you're likely to be more popular.
Really? Only last week I was looking at NetworkManager - and it didn't support this - even in the development version... based upon the information I could find.
Is the problem with Visio one of compatibility with existing Visio files, or in creating Visio-style diagrams?
If it's the latter, I was recently extremely imprssed with Dia, and prefer it to Visio for my purposes. If it doesn't suit yours, I'd be interested to know what features are missing.
LOL, very Dawkinsian. The problem with the divine is that, by definition, it defies scientific investigation. Sure, you might be able to set up these experiments - but you won't be able to overcome my objection that correlation is not causation.
This relates to other ideas I've been pondering of late (of a more earthly variety...) relating to economic policy. Essentially, central bankers are faced with something of a 'divine' problem in the task they're set. There are 'levers' that can be pulled, and there are statistics that can be watched - but it is impossible to say, with any scientific certainty, that any intervention had the desired effect... since there is no control-experiment... no parallel universe in which a different decision can be made with which to compare. This is rather like the prayer experiments... where, if you insist on a scientific-sounding explanation, the effect of praying might be subtle and indirect - like the proverbial butterfly wings in chaos theory. Until you can provide a parallel universe with a different praying behaviour... it is impossible to confirm or refute the value of the activity to the extent that any thinking scientist should be happy to agree.
Divine questions are about establishing what should be - independent of what is or what was. Simply thinking about these things has an impact on our environment - even if that is restricted to a few neurons in a skull... I think it would be very foolish to rule out any wider influence... irrespective of whether existing scientific theories can explain a mechanism. Our theories approximate reality - not the other way around.
The study about whether or not prayer helps the sickly followed this mindset, and since the sick humans had no idea what the study was about, when they were told that people were praying for them they thought they were much sicker and actually recovered more slowly.
I've recently read about this - in anticipation of seeing Richard Dawkins speak next month. In spite of 'Christian origins' I've always considered evolution to be as rock-solid a theory as mankind has discovered... but, on reading this prominent Atheist's work, I find myself aligning more and more with the perspectives of those Dawkins dismisses as 'Creationists'. Dawkins' arguments are so unbelievably poor, it seems to me that - in effect, if not by intention - he serves as a fifth-columnist for religious cults in undermining the intellectual integrity I associate with science.
I strongly suspect that the benefit of prayer is predominantly for those who pray as opposed to those for whom prayers are directed. I definitely see benefits from 'meditation' - and expect that ritual observance is beneficial in its own right... allowing people to 'do something' when they feel helpless... it likely also has subconscious influences which are significant. Perhaps the act influences the praying to subtly alter their practical behaviour? Similarly, being prayed for by a bunch of strangers who clearly have no particular reason to care must feel depressing - the idea that "it's come to this" could easily be a negative influence - especially for those who aren't immersed in the culture.
I consider myself comfortable with software - and I'm fascinated by the prospect of synthetic DNA. A glitch is that I don't have a degree-level background in either chemistry or biology - and this makes it very hard for me to appreciate what sort of things are viable. I would definitely like to know more about this field...
Can you recommend a text-book that explains the relevant biochemistry for this from-the-ground-up?
More to the point, an electronic copy is a sale to a different demographic - especially in the context of textbooks... A novel will likely be read cover-to-cover and most novels are inexpensive to purchase - making them fair value. Conversely, a technical text may only be referenced by index and only a single paragraph ever read. The price of the printed books precludes individuals establishing significant libraries - no such restriction necessarily hinders the digital realm... and, hence, the potential market size should be considered when the price is set.
In my opinion, there should be a much bigger and bolder effort with books than happened with music - to reflect the increased diversity of and minority interest in the printed word over recorded music.
I think there's a considerable problem in monetizing authoring - just as there is currently a problem in monetising recorded music. The problem, as I see it is a reluctance to embrace technology - and pricing models that assume a restricted audience.
I'd like a personal library - and, for now - at least, I'd like it in dead tree format (as well as electronic) - ideally... I'd like many thousands of books - covering a wide range of disciplines - I want copies to keep forever - and I'm willing to pay a sensible price to have a physical book - preferably bound in a consistent style for my library. I'm not content with a 'large bookshop with thousands of books' in cities - I want to be able to chose from any book ever published at any supermarket - and have it printed and bound while I wait... and the price should not exceed that of a cheap paperback. I see absolutely no reason why this should be impractical.
In this model, it's true, little revenue is generated by the sale of books - either in electronic or paper forms... with a product that is trivially copied - so should it be. The value that publishers should offer is the ability to find the book you want to read. Like with the world of recorded music, the value proposition is not in preventing the public access to material - but the opposite - i.e. the value is in filtering an impossibly large number of candidate texts (or recordings) down to sufficiently few that the customer can receive maximum benefit. This service would be extraordinarily valuable - and, what's more - because it would necessarily be personal and bespoke for every customer - it would be pointless to copy - hence eliminating piracy.
With books, like with recorded music, I'd happily pay a considerable annual sum to be recommended books and albums worthy of my time. I want a system that allows me to offer feedback and for that to guide future suggestions. Sure, this task is less easily established than mass-marketing a few select authors or musicians - but that's the whole point... that's what would make such a service valuable, while the current publishers (and laws) behave more like parasites hindering access to material - treating consumers as if they are indistinct and lacking any context for purchases.
For me, while I like listening to albums - buying them is insanely expensive if I'm going to hear each one once - then discard it. Similarly, if I want to do personal research in a topic, it is preposterously expensive and time-consuming to acquire each individual book (from, potentially a stack of thousands of possibly relevant texts) from a different supplier - just as it is inconvenient and impractical to engage in research only in public libraries.
I am not willing to give up the opportunity to hold onto a text for as long as I feel it might be important to me - nor am I willing to restrict myself to a small number of expensive titles - nor to purchase every copy for my private library of Alexandria. I would be willing to pay a flat-fee for access to a library of every work - where authors are remunerated proportionally to my rating their work (for relevance and quality) in my feedback. I'd also be willing to pay handsomely for a personal editor service (AI or otherwise) if it made easy the challenge of finding quality texts on subjects of interest.
Until these facilities arise, I'm forced to make the best of what is available. If that includes pdf copies of books - so be it. Surely any copyright infringement is, at least, morally mitigated by the absence of a comparable facility for a reasonable price?
What is the best compass technology available today? A magnetised needle is so, erm 20th century - I'd expect a solid state mechanism to identify orientation by now... it's an application I'd like my mobile phone to have.
2. It's very likely nothing to do with our brains, and a lot to do with more rigorous testing.
I don't buy the 'more rigorous testing' argument - I think that pre-supposes that testing was not performed diligently in the past. I think the most likely explanation is that the diagnoses were always flawed. Depression, mentioned in the blurb, for example has physical symptoms, but no known physical cause. My hunch is that many of the ailments we have are caused by factors outside the control of drugs, and it is the extent to which taking regular medication alters behaviour that makes a difference. For example, medication that can't be taken with alcohol presents a positive side-effect for heavy drinkers if taken diligently. Any regular activity has the same positive effects as observing a ritual.
Perhaps a larger proportion of ailments today are not the result of an illness? I'd find that easy to believe.
I'm sure that would come as a great surprise to anyone who ever implemented a virtual memory system.
You should assume the word "Random" is in bold-type, then the claim makes more sense.
Virtual memory systems only work effectively to the extent that data access has 'locality of reference' - which is often, but not always, found in practice.
To my mind, the real promise of solid-state is the random access. Since the earliest DP, software has had to take into account the sequential nature of access to durable storage - disk based storage never did have a uniform access time for blocks - and this has influenced everything from file-system design to memory architectures. In an SSD environment, it becomes possible to accurately model performance at higher levels of logical abstraction - and, in my view, better systems should emerge as a consequence... assuming, of course, the world at large doesn't do something crazy - such as always access flash through a FAT file-system... LOL!
Thanks for your reply. I'm actually considering writing some software that needs non-volatile storage of blocks of data - where constant-time random-access (for read, and especially write) is very desirable. I am free (at this stage) to chose an arbitrary block size (or collection of supported block sizes) if it will make any difference. I'm expecting that it will make a difference (if I get the block-size right) since RAM-IO has a bandwidth significantly higher than Fash-IO... making flash IO the bottleneck that counts.
The blocking approach of my application requires an entire block to be written (to a new location) - even if it is only 1 byte different to a previous version. A naive approach to optimisation might be to use a large block size to maximise throughput in the context of large changes... but, I strongly expect, this will only yield a benefit up-to a block size that matches the device itself.
Another reason for wanting to match block sizes relates to reliability. As far as I'm concerned, if a device exhibits a single failure, I'd prefer it to fail to read a single application-level block rather than two blocks... hence demanding an application block size as a multiple of the device block size...
A complication with respect to 'suck-it-and-see', for me, is that I'm running with USB devices (for now) making performance characteristics less obvious given a shared bus - and the possibility that IO bandwidth might be constrained by this interface rather than Flash itself.
I'd definitely like to find more information (especially about cheap flash storage) - since I don't intend to use any conventional file-system when I use flash to store my blocked data... a file system like FAT would be an overhead I don't need or want for my (specific) purposes.
Thanks for your exposition, I found it clear and interesting.
Specifically, I am very interested by the 512KB claim (I assumed, here, you meant KiB - i.e. 512*(2^10 bytes)... though that's not what interests me...)
I've read various claims from extremely non-authoritative sources about 'block sizes' for flash devices - and I've read about everything from 512 bytes up-to 64KiB - and 512KiB is the largest I've heard about.
It strikes me that, in the context of applications that store data on flash, where the application is free to define a storage format, that considerable performance optimisation is possible here. I'd love to find out:
1. Do flash device manufacturers (especially those selling SD cards or USB sticks, as well as SSDs) document their block-size? If so, can anyone point me at such a document?
2. Is there any 'standard' interface (on Windows or Unix) to reliably discover the block-size of an attached flash device?
3. Is there any method - other than empiric experiment (i.e. timing lots of writes) - to determine the block size of flash devices?
Everything you say makes sense, and I'm familiar with this technical background.
If I were to re-phrase my question, in the context of your reply, I'd be asking if equivalences have been shown between known-hard problems and the inverse to standard HMACs.
The particular issue that interests me is extremely similar to (type-1) preimage... but it is slightly easier to break. T1-preimage requires me to find m where H(m) is known. In my specific situation, we also know the length of m - it's the same as H(m) - or, equivalently, it is longer than H(m) but all except length(H(m)) bits are known to the attacker.
This might sound uninformed, but it simply makes me feel uneasy that I can't find any work (from a theoretical perspective) to show how hard a problem this is for specific HMACs... just a hypothesis and an open challenge. This might be good enough for the day-to-day, but is it OK if trying to put together a system that I want to remain robust for 50 years or longer?
I've been doing a fair bit of reading about secure hashes recently... as I'm interested in one specific property of hash functions that, as far as I can tell, is not discussed as widely as I'd expect... While collisions seem to draw a lot of attention, they are not of particular interest to me... since collisions (for me) affect only performance - not security.
The closest term for what I'm interested in is preimage resistance - but, as far as I can tell, this property is often disregarded as it is considered implausible to reconstruct a large input from a small output.
In a context in which I'm interested, however, the input to my hash function is itself a hash. I'm interested to know if I can rely on attackers being unable to compute the inverse of the hash function. Is there any work on formally establishing that any hash function is 'one way'?
I was very unimpressed with every USB stick I used until I bought a Titanium one. Though I had to pay about 5 times that price to get mine, I'm extremely happy with it as hardware.
I'm less impressed by the US software... and neither TruCrypt nor PortableApps are really make the mark as far as I'm concerned. 7zip archives with AES encryption works OK for transfer of data... but I'd really like to abandon the notion of a traditional file system... it simply fails to help me organise my data in a logical way in the context of portable drives.
I do not see the acquisition or reading of electronic copies to be unethical - either for in print or out-of-print books. In fact, one of the best pre-cursors to my opting to buy a physical book (my preferred format) is my acquisition of an electronic copy which I treat as a 'proof'.
I consider acquiring an electronic copy to be remarkably similar to borrowing a copy from a library - all be it at a lower cost to the state... and more convenient.
The entire approach to copyright is completely broken - a new business model is required. It is preposterous to be buying expensive books with zero evidence of applicability or quality; it is preposterous that access to published work today should be more restricted than in an era where lending libraries were used actively.
I recognise that authors might feel to have lost out should I read an unauthorised electronic copy - but not buy a dead-tree version... but this is a misunderstanding... If it weren't for a freely available electronic copy, I'd likely never have considered their work at all. A new business model is needed.
I'd love to see a commercial ebook library which would allow me access to an electronic copy of any published work (for a minimal fee)... and for the establishment of print-and-bind shops at which any books I want to buy can be printed and bound for my 'library'... and authors to be compensated at today's prices. I find it difficult to believe that we can't overcome the engineering problems which in years gone by left bulk printing and distribution far cheaper than bespoke production. It would be wonderful to walk into a small shop in a small town and be able to buy a bound copy of any book.
It's been a while since I was doing school exams, but I was aware even then (late 80s/early 90s) that the subjects I was learning were significantly less complex than the same subjects one or two decades earlier.
I think it is important to note, however, that this doesn't make the exams easier. By this, I mean, it is not better to sit less complex examinations. Where examinations are "easy" it rewards disproportionately those who can regurgitate quickly and accurately over those who comprehend. I think this measures the wrong ability... sure, good candidates might also excel at the parlour trick, but - I think - it creates unnecessary and counter-productive stress.
I've recently been fiddling with Twiki, and looking at structured wikis in general.
I've been looking at ways to build collaborated structured data... where the structure of the data (i.e. the sort of fields that need to be filled in) evolves over time - and where some form of editorial control can allow the reliability of data to improve over time... or, at least, allow several people to come to agree on the same false data.
I've not been getting very far very quickly - though Drupal is showing some promise. Have other Slashdotters tried to address this problem? What software did you find most useful?
To put it back on topic, if the TWiki team go off to build the structured collaborative tool I hoped TWiki would be from the outset... I'm in favour of whatever restrictions they see fit.
He's not inept, he's actually a pretty bright fella.
I was being forgiving.
Greenspan definitely understood the seriousness of the implications of dabbling with the Fed rate - and we mustn't forget that the influence of the Fed goes far beyond setting the Funds rate. The Fed acts as a backstop for commercial banks engaging in fractional reserve lending. If Greenspan was bright when he slashed to 1% and held... it proves him to be malevolent.
Slashdot Mirror
Care to share your credentials?
I'm 34 - and have also worked in software my entire life. I've a CS degree and I make regular professional use of material I learned.
I consider myself lucky never to have worked in the military, and I've managed to avoid working for banks so far, too. I prefer not to think of what I do as IT, either - that conjures visions of installing Microsoft products and fixing printers.... there's little need for an academic qualification in that field.
I'm amused that you assumed I was criticising your standing - rather than asking about the seniority of the qualified people to which you purjoratively referred. While I've no evidence to suggest that qualified people are more competent or pleasant, I've none to the contrary either. I have encountered circumstances where over-qualified people feel undervalued - and wonder if this might explain. My suspicion, now, is that, possibly subconsciously, you're anxious about your own credentials - and this colours your views on the 'educated' demographic.
I can only wish you all the best, and comment that - in my experience - acquiring a degree does not make someone less affable, competent or capable than they otherwise were.
Oddly, out of the "real asshole" crowd, many of them had a CS degree in addition to sucking at their jobs.
Out of interest, how old are you, and at what level of seniority are you talking?
My experience upgrading 9.04 to 9.10 Kubuntu:
I needed to make room to upgrade, because the 4 Gb SSD in the EEE was close to full.
Ditto... I spannered my 9.04 and then decided to swap the 16GB and 4GB partitions - since I'll likely only ever need a few megs for my own documents. It took several attempts - but, eventually it worked... (and I'm not sure what went wrong when the re-install failed... it might have been my fault)
That said, I'm fairly impressed. I love PAN support for 3G via a bluetooth phone in NetworkManager - but I am frustrated that VPN support seems intermittently buggy. I've had to start my VPN up-to a dozen times... it's not clear what's the problem - but there's an open ticket that seems to describe my issue.
Very impressed with the release in general...
I'm curious about "Ubuntu One"... on the surface it looks incredibly useful - but, conversely, I definitely do not want to put my personal documents (perhaps including banking details, say) into the cloud unencrypted.
I've done some careful thinking about a related situation recently, and I've come to a conclusion that might, at first, sound surprising. It is sure to be contested by many...
I think you're looking at the situation all wrong... it isn't something intrinsic about people or the public perception of them that defines popularity... quite the opposite. What is most important is how people see others.
One extreme has internalised that others deserve respect and assumes that they are likely intelligent and decent people - entitled to whatever views they hold. The other extreme assumes others are likely fundamentally stupid selfish people - a potential threat where it is best to put in the metaphorical boot first.
I think this dimension to personality best explains popular cliques. Far from having established the 'best' friends, it is more likely high anxiety and fear of exclusion that binds the 'popular' groups. The 'popular' are more likely to hold others in contempt - and assume others "stupid". The unpopular are more likely self-confident and and making the error that they over estimate others' intelligence relative to their own faculties.
I think it's all about what you think others (should/do) think of you that defines how you are perceived. If you show contempt for the unpopular, you're likely to be more popular.
Don't worry, he's only had two strikes - he's in the clear until he gets a third.
Really? Only last week I was looking at NetworkManager - and it didn't support this - even in the development version... based upon the information I could find.
What gives?
Is the problem with Visio one of compatibility with existing Visio files, or in creating Visio-style diagrams?
If it's the latter, I was recently extremely imprssed with Dia, and prefer it to Visio for my purposes. If it doesn't suit yours, I'd be interested to know what features are missing.
LOL, very Dawkinsian. The problem with the divine is that, by definition, it defies scientific investigation. Sure, you might be able to set up these experiments - but you won't be able to overcome my objection that correlation is not causation.
This relates to other ideas I've been pondering of late (of a more earthly variety...) relating to economic policy. Essentially, central bankers are faced with something of a 'divine' problem in the task they're set. There are 'levers' that can be pulled, and there are statistics that can be watched - but it is impossible to say, with any scientific certainty, that any intervention had the desired effect... since there is no control-experiment... no parallel universe in which a different decision can be made with which to compare. This is rather like the prayer experiments... where, if you insist on a scientific-sounding explanation, the effect of praying might be subtle and indirect - like the proverbial butterfly wings in chaos theory. Until you can provide a parallel universe with a different praying behaviour... it is impossible to confirm or refute the value of the activity to the extent that any thinking scientist should be happy to agree.
Divine questions are about establishing what should be - independent of what is or what was. Simply thinking about these things has an impact on our environment - even if that is restricted to a few neurons in a skull... I think it would be very foolish to rule out any wider influence... irrespective of whether existing scientific theories can explain a mechanism. Our theories approximate reality - not the other way around.
The study about whether or not prayer helps the sickly followed this mindset, and since the sick humans had no idea what the study was about, when they were told that people were praying for them they thought they were much sicker and actually recovered more slowly.
I've recently read about this - in anticipation of seeing Richard Dawkins speak next month. In spite of 'Christian origins' I've always considered evolution to be as rock-solid a theory as mankind has discovered... but, on reading this prominent Atheist's work, I find myself aligning more and more with the perspectives of those Dawkins dismisses as 'Creationists'. Dawkins' arguments are so unbelievably poor, it seems to me that - in effect, if not by intention - he serves as a fifth-columnist for religious cults in undermining the intellectual integrity I associate with science.
I strongly suspect that the benefit of prayer is predominantly for those who pray as opposed to those for whom prayers are directed. I definitely see benefits from 'meditation' - and expect that ritual observance is beneficial in its own right... allowing people to 'do something' when they feel helpless... it likely also has subconscious influences which are significant. Perhaps the act influences the praying to subtly alter their practical behaviour? Similarly, being prayed for by a bunch of strangers who clearly have no particular reason to care must feel depressing - the idea that "it's come to this" could easily be a negative influence - especially for those who aren't immersed in the culture.
I consider myself comfortable with software - and I'm fascinated by the prospect of synthetic DNA. A glitch is that I don't have a degree-level background in either chemistry or biology - and this makes it very hard for me to appreciate what sort of things are viable. I would definitely like to know more about this field...
Can you recommend a text-book that explains the relevant biochemistry for this from-the-ground-up?
More to the point, an electronic copy is a sale to a different demographic - especially in the context of textbooks... A novel will likely be read cover-to-cover and most novels are inexpensive to purchase - making them fair value. Conversely, a technical text may only be referenced by index and only a single paragraph ever read. The price of the printed books precludes individuals establishing significant libraries - no such restriction necessarily hinders the digital realm... and, hence, the potential market size should be considered when the price is set.
In my opinion, there should be a much bigger and bolder effort with books than happened with music - to reflect the increased diversity of and minority interest in the printed word over recorded music.
I think there's a considerable problem in monetizing authoring - just as there is currently a problem in monetising recorded music. The problem, as I see it is a reluctance to embrace technology - and pricing models that assume a restricted audience.
I'd like a personal library - and, for now - at least, I'd like it in dead tree format (as well as electronic) - ideally... I'd like many thousands of books - covering a wide range of disciplines - I want copies to keep forever - and I'm willing to pay a sensible price to have a physical book - preferably bound in a consistent style for my library. I'm not content with a 'large bookshop with thousands of books' in cities - I want to be able to chose from any book ever published at any supermarket - and have it printed and bound while I wait... and the price should not exceed that of a cheap paperback. I see absolutely no reason why this should be impractical.
In this model, it's true, little revenue is generated by the sale of books - either in electronic or paper forms... with a product that is trivially copied - so should it be. The value that publishers should offer is the ability to find the book you want to read. Like with the world of recorded music, the value proposition is not in preventing the public access to material - but the opposite - i.e. the value is in filtering an impossibly large number of candidate texts (or recordings) down to sufficiently few that the customer can receive maximum benefit. This service would be extraordinarily valuable - and, what's more - because it would necessarily be personal and bespoke for every customer - it would be pointless to copy - hence eliminating piracy.
With books, like with recorded music, I'd happily pay a considerable annual sum to be recommended books and albums worthy of my time. I want a system that allows me to offer feedback and for that to guide future suggestions. Sure, this task is less easily established than mass-marketing a few select authors or musicians - but that's the whole point... that's what would make such a service valuable, while the current publishers (and laws) behave more like parasites hindering access to material - treating consumers as if they are indistinct and lacking any context for purchases.
For me, while I like listening to albums - buying them is insanely expensive if I'm going to hear each one once - then discard it. Similarly, if I want to do personal research in a topic, it is preposterously expensive and time-consuming to acquire each individual book (from, potentially a stack of thousands of possibly relevant texts) from a different supplier - just as it is inconvenient and impractical to engage in research only in public libraries.
I am not willing to give up the opportunity to hold onto a text for as long as I feel it might be important to me - nor am I willing to restrict myself to a small number of expensive titles - nor to purchase every copy for my private library of Alexandria. I would be willing to pay a flat-fee for access to a library of every work - where authors are remunerated proportionally to my rating their work (for relevance and quality) in my feedback. I'd also be willing to pay handsomely for a personal editor service (AI or otherwise) if it made easy the challenge of finding quality texts on subjects of interest.
Until these facilities arise, I'm forced to make the best of what is available. If that includes pdf copies of books - so be it. Surely any copyright infringement is, at least, morally mitigated by the absence of a comparable facility for a reasonable price?
What is the best compass technology available today? A magnetised needle is so, erm 20th century - I'd expect a solid state mechanism to identify orientation by now... it's an application I'd like my mobile phone to have.
2. It's very likely nothing to do with our brains, and a lot to do with more rigorous testing.
I don't buy the 'more rigorous testing' argument - I think that pre-supposes that testing was not performed diligently in the past. I think the most likely explanation is that the diagnoses were always flawed. Depression, mentioned in the blurb, for example has physical symptoms, but no known physical cause. My hunch is that many of the ailments we have are caused by factors outside the control of drugs, and it is the extent to which taking regular medication alters behaviour that makes a difference. For example, medication that can't be taken with alcohol presents a positive side-effect for heavy drinkers if taken diligently. Any regular activity has the same positive effects as observing a ritual.
Perhaps a larger proportion of ailments today are not the result of an illness? I'd find that easy to believe.
Random I/O is essentially uncacheable.
I'm sure that would come as a great surprise to anyone who ever implemented a virtual memory system.
You should assume the word "Random" is in bold-type, then the claim makes more sense.
Virtual memory systems only work effectively to the extent that data access has 'locality of reference' - which is often, but not always, found in practice.
To my mind, the real promise of solid-state is the random access. Since the earliest DP, software has had to take into account the sequential nature of access to durable storage - disk based storage never did have a uniform access time for blocks - and this has influenced everything from file-system design to memory architectures. In an SSD environment, it becomes possible to accurately model performance at higher levels of logical abstraction - and, in my view, better systems should emerge as a consequence... assuming, of course, the world at large doesn't do something crazy - such as always access flash through a FAT file-system... LOL!
Thanks for your reply. I'm actually considering writing some software that needs non-volatile storage of blocks of data - where constant-time random-access (for read, and especially write) is very desirable. I am free (at this stage) to chose an arbitrary block size (or collection of supported block sizes) if it will make any difference. I'm expecting that it will make a difference (if I get the block-size right) since RAM-IO has a bandwidth significantly higher than Fash-IO... making flash IO the bottleneck that counts.
The blocking approach of my application requires an entire block to be written (to a new location) - even if it is only 1 byte different to a previous version. A naive approach to optimisation might be to use a large block size to maximise throughput in the context of large changes... but, I strongly expect, this will only yield a benefit up-to a block size that matches the device itself.
Another reason for wanting to match block sizes relates to reliability. As far as I'm concerned, if a device exhibits a single failure, I'd prefer it to fail to read a single application-level block rather than two blocks... hence demanding an application block size as a multiple of the device block size...
A complication with respect to 'suck-it-and-see', for me, is that I'm running with USB devices (for now) making performance characteristics less obvious given a shared bus - and the possibility that IO bandwidth might be constrained by this interface rather than Flash itself.
I'd definitely like to find more information (especially about cheap flash storage) - since I don't intend to use any conventional file-system when I use flash to store my blocked data... a file system like FAT would be an overhead I don't need or want for my (specific) purposes.
Thanks for your exposition, I found it clear and interesting.
Specifically, I am very interested by the 512KB claim (I assumed, here, you meant KiB - i.e. 512*(2^10 bytes)... though that's not what interests me...)
I've read various claims from extremely non-authoritative sources about 'block sizes' for flash devices - and I've read about everything from 512 bytes up-to 64KiB - and 512KiB is the largest I've heard about.
It strikes me that, in the context of applications that store data on flash, where the application is free to define a storage format, that considerable performance optimisation is possible here. I'd love to find out:
1. Do flash device manufacturers (especially those selling SD cards or USB sticks, as well as SSDs) document their block-size? If so, can anyone point me at such a document?
2. Is there any 'standard' interface (on Windows or Unix) to reliably discover the block-size of an attached flash device?
3. Is there any method - other than empiric experiment (i.e. timing lots of writes) - to determine the block size of flash devices?
Thanks for any hints...
Everything you say makes sense, and I'm familiar with this technical background.
If I were to re-phrase my question, in the context of your reply, I'd be asking if equivalences have been shown between known-hard problems and the inverse to standard HMACs.
The particular issue that interests me is extremely similar to (type-1) preimage... but it is slightly easier to break. T1-preimage requires me to find m where H(m) is known. In my specific situation, we also know the length of m - it's the same as H(m) - or, equivalently, it is longer than H(m) but all except length(H(m)) bits are known to the attacker.
This might sound uninformed, but it simply makes me feel uneasy that I can't find any work (from a theoretical perspective) to show how hard a problem this is for specific HMACs... just a hypothesis and an open challenge. This might be good enough for the day-to-day, but is it OK if trying to put together a system that I want to remain robust for 50 years or longer?
I've been doing a fair bit of reading about secure hashes recently... as I'm interested in one specific property of hash functions that, as far as I can tell, is not discussed as widely as I'd expect... While collisions seem to draw a lot of attention, they are not of particular interest to me... since collisions (for me) affect only performance - not security.
The closest term for what I'm interested in is preimage resistance - but, as far as I can tell, this property is often disregarded as it is considered implausible to reconstruct a large input from a small output.
In a context in which I'm interested, however, the input to my hash function is itself a hash. I'm interested to know if I can rely on attackers being unable to compute the inverse of the hash function. Is there any work on formally establishing that any hash function is 'one way'?
I was very unimpressed with every USB stick I used until I bought a Titanium one. Though I had to pay about 5 times that price to get mine, I'm extremely happy with it as hardware.
I'm less impressed by the US software... and neither TruCrypt nor PortableApps are really make the mark as far as I'm concerned. 7zip archives with AES encryption works OK for transfer of data... but I'd really like to abandon the notion of a traditional file system... it simply fails to help me organise my data in a logical way in the context of portable drives.
He should ask them to reconsider if he promises to sell beer and not stock gin.
I do not see the acquisition or reading of electronic copies to be unethical - either for in print or out-of-print books. In fact, one of the best pre-cursors to my opting to buy a physical book (my preferred format) is my acquisition of an electronic copy which I treat as a 'proof'.
I consider acquiring an electronic copy to be remarkably similar to borrowing a copy from a library - all be it at a lower cost to the state... and more convenient.
The entire approach to copyright is completely broken - a new business model is required. It is preposterous to be buying expensive books with zero evidence of applicability or quality; it is preposterous that access to published work today should be more restricted than in an era where lending libraries were used actively.
I recognise that authors might feel to have lost out should I read an unauthorised electronic copy - but not buy a dead-tree version... but this is a misunderstanding... If it weren't for a freely available electronic copy, I'd likely never have considered their work at all. A new business model is needed.
I'd love to see a commercial ebook library which would allow me access to an electronic copy of any published work (for a minimal fee)... and for the establishment of print-and-bind shops at which any books I want to buy can be printed and bound for my 'library'... and authors to be compensated at today's prices. I find it difficult to believe that we can't overcome the engineering problems which in years gone by left bulk printing and distribution far cheaper than bespoke production. It would be wonderful to walk into a small shop in a small town and be able to buy a bound copy of any book.
It's been a while since I was doing school exams, but I was aware even then (late 80s/early 90s) that the subjects I was learning were significantly less complex than the same subjects one or two decades earlier.
I think it is important to note, however, that this doesn't make the exams easier. By this, I mean, it is not better to sit less complex examinations. Where examinations are "easy" it rewards disproportionately those who can regurgitate quickly and accurately over those who comprehend. I think this measures the wrong ability... sure, good candidates might also excel at the parlour trick, but - I think - it creates unnecessary and counter-productive stress.
I've recently been fiddling with Twiki, and looking at structured wikis in general.
I've been looking at ways to build collaborated structured data... where the structure of the data (i.e. the sort of fields that need to be filled in) evolves over time - and where some form of editorial control can allow the reliability of data to improve over time... or, at least, allow several people to come to agree on the same false data.
I've not been getting very far very quickly - though Drupal is showing some promise. Have other Slashdotters tried to address this problem? What software did you find most useful?
To put it back on topic, if the TWiki team go off to build the structured collaborative tool I hoped TWiki would be from the outset... I'm in favour of whatever restrictions they see fit.
He's not inept, he's actually a pretty bright fella.
I was being forgiving.
Greenspan definitely understood the seriousness of the implications of dabbling with the Fed rate - and we mustn't forget that the influence of the Fed goes far beyond setting the Funds rate. The Fed acts as a backstop for commercial banks engaging in fractional reserve lending. If Greenspan was bright when he slashed to 1% and held... it proves him to be malevolent.