the new grad knew a hot emerging technology that a client wanted.
Did the senior engineer know the technology? Sure the technology may not have existed when they were hired, but for a senior engineer to deserve the title they learn or invent new technologies. If the senior engineer did know the technology, or could pick it up before the new hire becomes productive (i.e. learns company procedures and enough politics to operate effectively), then the new hire isn't worth the extra cash. If the senior engineer doesn't know the new technology and can't pick it up fast enough, then the new guy deserves more. At least until the senior engineer catches up.
NOTE: I don't mean to imply a senior engineer knows all new technologies, nor that they can alway guess what might be important to their company. Just that they can and do keep up to date on things, and sure sometimes a new library or language they missed or dismissed as "no better then this other thing I already know" becomes more important then they had guessed, but they ought to be able to pick it up. For example prior to the iPhone becoming popular it might have been reasonable to not know any Obj-C. Someone coming out of collage might know Obj-C, but a senior engineer with knowledge of C and any sort of smalltalk style OO language should be able to pick up Obj-C really fast. Likely fast enough that they are able to make real programs while the new guy is still learning the ins and outs of the bug tracking system, and who to listen to and who to ignore in meetings.
My personal pet theory is yes it really does matter. For a long time doubling every 18 months was the arbitrary goal, and Intel could say "if we don't hit X by Y then AMD will overtake us because they double every 18 months", and AMD could say "If we don't hit X by Y then Intel will smash us because they double every 18 months". So each poured whatever they needed into R&D to make it more or less happen. Sometimes one got more ahead then the other and got to roll in the hay for a while (or be unable to fill all the market demand).
If Moore's law wasn't sitting their prodding them to double or die every 18 months my guess is they would have sat there and gone "Will they REALLY invest $12.6 billion dollars on a new fab? That seems stupid, I bet they will try to get by with a 10% bump from a new microarchicture, so we should aim for 15%". We would have seen a LOT more severely lopsided product matchups, but overall I'm guessing slower growth.
Not that it matters so much now since they don't seem to be able to keep up with doubling processing capability (unless your problem set is very parallel...and even so Amdahl's Law will get them sooner or later -- add as many CPUs as you like, at some point there is a non-parallel part of your problem space and that part's performance will dominate, even with an infinite number of CPUs).
A dock connector gets you line level output which needs less fiddling to get acceptable results from. I'll also bet there are far more people with a iPod dock connector on their iPod/iPhone then there are folks that walk around with a 3.5mm male-male cable.
However I think a 3.5mm line-in would be a good thing to have in addition to a dock connector.
The threat meter is arguably much more "necessary". I need to know, in real time, during the fight, if I'm about to pull aggro.
The default UI can be told to make a sound and flash the edges of the screen when you get "close" to pulling aggro (90% or so I think). Not at all perfect, but will do in a pinch. The tank is more in need of an addon, as nothing tells them when someone else is getting close (as far as I know).
So instead of clicking on a player's name in the raid frame and then hitting 1 for a flash heal or 2 for a shield
I use to use grid+clique for that, but with a minor change you don't need addons:
Macro:/cast [@mouseover] flash heal
Macro:/cast [@mouseover] shield
Put flash on the 1, the shield on 2. Mouse over whoever needs that flash and press 1. No "click then press". It works nicely (you can also even set it up to use target or self if there is no mouseover).
I've discarded clique. I still use grid as it can be configured to display a LOT more information in a small space then the built in unit frames, and just as importantly it can be configured NOT to display information you don't care about.
Stock raid unit frames group by party (5 players) within the raid, put one healer per party and have them heal their group. Doable with stock unit frames. (You could also "cheat" and put groups of five in the same vent channel and have them yell out to their healer, that'll work for a whole lot of things)
I prefer grid, configured to stick a big center icon on for the debuf, but that is a preference, not a requirement.
I'm sure that you can raid "fine" at a low level with a vanilla client, and perhaps even certain high-level roles can be done with the stock interface. But in general, you absolutely need the conveniences that raiding mods provide for high-level raiding.
Sure tanks and healers have a harder time in raids without mods then dps, but you can raid ICC without mods as any of the roles.
[...]what concessions do the automakers get for this service? [...]
Well, I'm not sure I believe any of that, but I can see a reason automakers might be down of diesels if they had the chance... a diesel engine tends to last way longer then gas engines, and it is to the advantage of car companies.
That said, conspiracy theories are fun to think about, but the real reasons behind stuff tend to be more complex and less fun to BS about.
About 2 out of 3 gas stations around here (Bay Area CA) don't sell diesel. I'm assuming some combination of making the equipment cheaper, maybe environmental licensing and inspection cost savings and such.
The argument that you can take 30 books on an electronic reader just doesn't wash with me. This reading is for enjoyment, and I would rather savor my joy one book at a time. If I am not reading it, it is because the story has lost my interest and I no longer want that book with me no matter what the source.
Ok, how about some related arguments:
In a space/weight limited situation (like living in an RV) you might not be able to have more then a half dozen paper books, or at the very least you would have to decide what else not to own.
A regular house still has limited space. In an apartment it could be hard to have more then one or two bookcases. The basement of my house had about 10 or so.
That won't apply so much if you only ever read the same half dozen books, or if you never reread books. (personally the last trip I took it was nice to have a fair number of books to choose from but only have to budget space for them...however for me it the bigger value is always having a book with me, bank lines, restaurants and such -- I have read more books per month since starting on ebooks, and I've been a big reader all my life)
Just my preference though
Thats cool. Some folks like paperbacks and some like hardbacks. I happen to like ebooks.
It is morally wrong to exploit someone in that position financially (or otherwise). Claiming that you disclosed the risks and they signed the waiver doesn't make it ok. In a sense they do have a gun pointed at their head... whats a raft of fine print and a 2nd mortgage when your life is on the line.
Absolutely right! We ought not take payment from anyone who is sick! Experimental medicine? Feh! Who needs that crap! I say we should have stuck to bleeding people. However as long as we have developed some form of medicine we better not get any better at it. Also, let's be sure to treat sick people like they are babies, and other people always know better.
I mean, if I get a terminal illness I sure hope nobody is allowed to offer me any experimental cures. If it hasn't been FDA approved, and in use for at least a decade it would be far better to die then take that kind of risk!
Ok, so there were no significant "give backs" from MS, but on the other hand they didn't start out with some form of mutant TCP/IP that required the rest of the internet to bend over backwards to sort of interoperate with. So that is still a win in my book.
Debugger speed matters when you debug something large. I frequently have things where gdb takes close to a minute to load symbols.
Debugger speed also matters for things like watchpoints. Watchpoints in gdb are so slow I almost never use them (like not at all this year, and maybe once last year).
Last time I have fast watch points (which was 1991, I had a CPU emulation box that ran at the full 68k speed with up to 8 watchpoints, and could dump the prior 200 cycles of bus activity and internal register state) I used them a LOT. Like three times a week.
A faster debugger might also make integrating a debugger and a unit test environment simpler. "run this code, fail the test if function X is not called, or if function Y is called, or if function Z isn't called exactly 8 times..." As it is there are some tests that can't be written using the API, you need scaffolding, and that scaffolding can introduce or suppress it's own errors...
I think your conception of this is interesting from a pure economics standpoint, but it's a little off from how publishers are actually viewing the situation.
Fair enough, I'm not an author or a publisher, so I don't know how they really work. It is the model movies use. Theater release (in fact in 4 different runs), Cable TV release, DVD priced to rent, PPV release, DVD priced to own. Oh, and digital downloads are in the mix somewhere. Foreign and domestic releases as well. All timed to try to extract the maximum revenue, all the same or nearly the same actual product.
I don't believe that there's as smooth or continuous a scale of when readers want to read something as you suggest. People either want to read it right away, or they'll read it when they get around to it.
I'm not so sure. When I get to the end of a book I like, if it is part of a seres I'll look at the next book. If it is a hardback I make a choice about buying it now at that price vs. reading something else now and checking back later to see if it has hit paperback. If I like the book enough (or the ending was enough of a clifhanger) I'm more likely to pay the hardback price.
Since I have switched to exclusively ebooks (partly because they really are more convent, and partly because I'm planning to retire to an RV) I'm less likely to pay hardback prices then I use to, but it does happen. It would happen more if the choice wasn't $30 now vs. $5 later, but $10 now vs. $5 later.
You are right that people don't have a fine grained idea of a price they are willing to pay for a given book. However if you give people a price for a book they tend to be able to very easily decide if they are willing to pay that much. Generally people are better about making a "buy/no-buy at $X" choice then a "what would you pay for this?" choice.
Most people are aware that (many) things will be cheaper if they wait ("wait for a sale, those things always go on sale"... "new movie for $30, or I could wait two months and I bet it will be $10", "new high end laptop now, or next years mid range"), but they still buy stuff now when they feel like it. The interesting question is how transparent can you make the "price decay" part before people stop being happy when they get something for the price they want, and start being resentful when they pick their price (or decided to wait).
Yes if you drop prices week by week some people will wait until next week. You may need to make the process a little more random. Or you may find that it really works just because you pick up a lot more sales somewhere in the middle of the curve that otherwise waited for paperback prices.
Unfortunately as long as it is "might be better", "could get higher margins", and even "could get higher margins AND more total sales" publishers will rightly say "um, dude, we have a book publishing model that makes money why do you think we want to experiment with it?". They are totally correct, I find the problem space interesting, and would be pleased to come out with a better book market to buy stuff from, but at the end of my day my paycheck comes from outside the book industry, and they have a paycheck that comes from entirely within it. It would be insane to just tinker with stuff because it "might" get better since it "might" get worse, and it isn't too badly off just now (note that is books, not newspapers or magazines).
But nobody except for a few freaks (the "I'll never touch a paper book again!" crowd) is actually willing to pay hardback prices for an e-book.
There are some people that are extremely space constrained (fulltime RVers for example) who won't buy paper books, but I would't say are freaks. Well, technically they are more weight constrained then space constrained for books. I don't think many of them want to pay hardback prices for ebooks, but some will. The people that pay hardback prices for ebooks tend to be the people who were aware of why hardbacks really cost more and come out sooner. They are just a way for book publishers to rent seek (get the highest price people are willing to pay). That could actually be done even better with ebooks. Start the price at hardback level, and decay the price each week, hitting a somewhere under paperback price about a month after the paperback is out.
In a perfectly rational world it would work well, people would buy when the price hit what they are willing to pay to read it now...and since it has more steps then the current scheme many people that would pay a little more to read it "now not next month" but not "a whole lot more" can actually do that (this works better with a series of novels that is still being added to, or a author with a distinct style). In the real world however it does run the risk that it will just piss people off because they don't like rent seeking except when they are the seeker.
Layers and layers of "middle men" are not needed anymore.
Or more to the point, each layer of middle men need to add value. Between author and reader the author wants someone to deal with all the bother of collecting money, the reader wants someone to deal with filtering out all the obviously bad works, and either or both the author and reader wants someone to fix up the language usage and the other nice stuff the publisher's editor normally does. There are some other tasks in there that are useful as well.
There is no need for that to be spread across two middle-men (publisher + bookseller), but there is no special need for it to be one middle man vs. one per operation as long as having 12 companies touch the book doesn't make it cost more then only having 1 or 2 touch it.
With the current state of things the publisher filters out the worst of the bad writing (that is to say 99.998% of what they get), fixes up the english usage, remind the author when the sex of one of their characters changes midway through the book, when things get talked about in Ch4 but don't happen until Ch7, and all that kind of nice stuff. Then Amazon makes sure the book got nicely "kindle formatted" puts a price on it, handles some of the money, arranges network service for Kindle devices (iPod touch and iPhone users get to manage that on their own), handles long term storage of customer purchases.
All of those are valuable. If we stuff all that into one company that charges me less, I'm happy. If we stuff that all into one company that charges me the same amount, I'm still happy. If we break that out into 1 companies that charge me the same or less, well I'm still happy. If we charge more, I get cranky. If we do less stuff for the same money, I get cranky. If we charge less for less I probably still get cranky.
So I'm less happy by the proposal of a bizaar. Sure I probably pay less, but did the author hire a proofreader? How do I find the tiny number of authors that don't suck? It would work out well for authors who are already successful under the existing system. I know some of their past works weren't crap, so their new works have enough chance to be not crap that I can "afford" to spend the time to read a review or just buy the book. I know they have published stuff in the past, so they will know the value of a good editor and hire one. I might even pay less. It will totally suck for knowing what new authors I could buy from and not run into extremely poor writing, or a good story but no proofreading....
Solvable problems, but mostly by adding middlemen back in (for example, tags for "professional editing", tags for "slushpile review" and the like). They may not be middlemen in this model, but they still operate on the text, and then a tag gets affixed to it, and I filter out stuff lacking the tags. How they get payed might be the only difference (plus the "in theory" difference that people could choose to buy stuff that didn't pass a slushpile review, or wasn't edited by anyone....but I doubt enough people would do this to make dismantling the current system a good idea!). If we move the way something gets a slushpile review we could also break the value of that system. Currently a publisher does that, and if a publisher culls too many books they can lose good sellers. If they cull too loosely then they run the risk of people deciding "oh they publish crap, I won't risk reading that". If the "publisher middleman" goes away and an author has to pay for a slushpile review then they have incentive to pick registered slushpile reviewers that rubber stamp the most content. There may be ways to fix that, but struggling to get "as good as the existing system" while offering no real advantages doesn't sound exciting.
If you love your dog or cat, keep it on a leash outdoors. Being able to track it down when it's road kill, or frozen to death and chewed up by a snowblower, isn't being a good owner.
One isn't a substitute for the other. I don't let my dogs wonder, they go in the yard when I go in the yard to watch them. However over the years their have been some escapes. The fence gate has twice in five years managed to be open, or unlatched enough that a dog has opened it by pushing against it. Herding up six escaped dogs is tricky. My fence is 6.5 feet tall (legal limit for where I live). At least two of my dogs can jump over it, they just don't know it, so they only leap it when something extra interesting is on the other side. Then I have to shove five dogs inside, and go fetch the escapee. I had a (very minor) fire. I evac'ed all the dogs to kennels I keep in my van. One of the dogs popped the gate off the kennel (it is a cramped fit, so the kennel is a little distorted, and the gate doesn't fit perfectly...but she had no risk of being burned alive). I'm pretty sure I've even brought dogs in from the backyard, and had them run out a house door someone had open. Once or twice, they have hopped a baby gate and gone out a door I was bringing things in from (I normally kennel them if I'm bringing much in and out of the house though). When camping I've had dogs escape from tents twice too.
I know the moment my dog is missing (or you know, in a really short span of time). Long before they freeze anywhere. However the longer it takes me to figure out where they have decided to go exploring, the bigger the chance they become road kill. Or in the case of my small dogs, get eaten by something. Normally I recover them within a few minutes. A few times it has been a half hourish. Once it was about two hours (I was living in an area with huge yards with lots of trees, 4+ acres per house, hard to find a dog that escapes your field of view).
So yeah, a pet tracking GPS could be very useful to me. If they weren't so amazingly expensive (monthly plan!), so unlikely to be charged when I need them, and too big for at least 2 of my dogs anyway. I'm sure there are a lot of responsible pet owners who would find them handy. The same folks who think of it more like having them ID chipped "this will help me recover them if they go missing" would. Not everyone who wants a GPS for their dog plans on just opening the front door and letting the beastie entertain itself until dinner time.
Those 600 megabytes or so of cached memory can be instantly allocated for anything which suddenly needs it.
Does Linux have a seporate report for "cached dirty"? Most Unixes don't. So some (or even all) of it could be data that has been written, so it can't be reallocated until it gets flushed. I would guess it is very rarely "all", or even rarely "most", but that'll depend on your workload.
As far as I'm concerned, if the battery loses the ability to store the same amount of power as it did when you first bought it, then it has a problem with battery memory.
That isn't "battery memory" any more then "mail from my boss about me using the wrong TPS sheet is spam", or "my video game drops to 10FPS is crashed".
It is most definitely a "battery problem" (maybe even an unexpected one), but it isn't "battery memory". We use specific words to mean specific things so we can talk about problems and solutions rather then having a lot of words that just mean "DO NOT WANT". So if someone says a specific laptop has a battery memory problem and they really mean memory problem then you can avoid that by doing the inconvenient thing of making sure that whenever you use the battery you run it down to zero. If they really just mean "it only has 300 full charge cycles" then treating it like a battery memory problem is so totally the wrong thing -- their warning has caused you to destroy your battery far far faster.
(yes, some batteries have limited numbers of charge cycles, NiMH batteries do, but no memory effect, at least here on earth)
How many times have I dropped my Treo? Many, and it's still fine, the phone isn't as pretty as it once was but it's still fully functional.
Either you are very lucky, or I'm unlucky, or both. Many times my Treo has fallen out of my shirt pocket onto asphalt and wacked one of it's keys hard enough that the bit of tinfoil it has under each key gets dinted backwards. That leaves a treo that has no keyboard, and is apparently spending so much of it's CPU dealing with keyboard interrupts that it is pretty unresponsive to the stylus.
It is an easy fix, I (or more likely my wife) open the case and move a circle of foil from my "dead, for parts" 650 into the wounded 650 to 700. However the Treo is basically unusable until the fix is made. I'm running low on foil bits now, and getting the supply cost me dearly. Fortunately I broke myself of the habit of keeping the phone in a shirt pocket except while driving.
He does absolutely nothing what-so-ever to back those statements up. Why is a GPL a "non-starter" for "some" environments? What environments?
Easy enough, places that want to ship a product (assuming the interface libraries are GPL not LGPL). Or just places that want to ship a product with the database embedded inside it. They might be better served with Postgress, or sqlite. Or licensing Oracle or something (of corse if licensing a database is an option mysql can do that).
Which doesn't mean the article wasn't kind of weak, and half the points aren't lame...but this one at least made sense.
Re:It's really not clear from this review...
on
The DV Rebel's Guide
·
· Score: 1
I have read a few sections of the book and there is a lot in it that isn't After Effects specific (or even DV specific). Things like filming from the bed of a pickup truck to avoid the need to get a film license from many local governments (most places - apparently - only require a license if you put your tripod on the ground -- so you can do handheld and get that shaky "this ain't hollywood" feel, or you can put your tripod in the pickup and roll).
I haven't read many books on making a film though, so I don't know how good it is if you have already read other "make a film on a budget" AND you don't use After Effects.
only shows that are availible in my cable package but that I forget to DVR, am I a criminal? I am already paying to watch the shows, and archiving is fair use...if others DLing dont have the legal right, that is their problem!
If those shows are normal "over the air" broadcast, and they come to you with the original commercials, then whatever laws the cable companies used when they started up are on your side...unless those laws have been repealed or superseded.
However if the shows had the commercials stripped from them, or were from HBO or something, then there isn't as much on your side (I'm not saying it is illegal, I wouldn't know). Does anyone know of any bittorrent feeds of TV shows that still have commercials in them? They may be bigger files, and more awkward to watch, but hey, but far more likely to be legal, and definitely morally no different from what cable companies did 20~30 years ago.
All of my external drives are triple interface (FW400/FW800/USB), but I really prefer using 800. Speed is important for me when I'm moving around lots of data; it's not uncommon for me to have to move 40 gigs from one drive to another a couple of times a week. FW800 is also fast enough that I can keep my Virtual PC partitions on it and not really have any noticible slowdown from the internal drive.
Have you tried timing the mass data move with FW400? There are a lot of disks that can saturate FW400 on paper, but don't in practice (or do, but not by a whole lot). If you have I would love to know what the actual speed difference is (I recently bought a FW400 enclosure to do backups, and would like to know what I'm missing since I have FW800 on my laptop...not that I think I'll care for backups).
Apple claims they do this [release the fastest computer ever]every time they release a new Apple computer. And they're usually right, because there's zero competition on Apple computers.
Not really. The iBooks when announced were slower then the PowerBooks. The iMac is slower then the PowerMac (well, except for the intel iMac)...
[...shared_ptr and weak_ptr...] They work great and they give you the good parts of garbage collection without the bad parts.
Well a lot of the good points. But, it isn't pervasive like a real GCed language, so you may have to interface with code that doesn't use it (and has leaks or unhappy ways to use it). It also actually has a higher (but more predictable) cost.
It also has some advantages over GC the perceptibility isn't just for CPU load, the exact lifetime of objects is predictable, which is very useful for RAII. E.g. a shared_ptr to a lock object will destruct exactly when the last reference goes away releasing the lock...on a GC system the destructor fires "sometime" after the last reference goes away. Very useful with exceptions and other complex control flow.
Slashdot Mirror
Did the senior engineer know the technology? Sure the technology may not have existed when they were hired, but for a senior engineer to deserve the title they learn or invent new technologies. If the senior engineer did know the technology, or could pick it up before the new hire becomes productive (i.e. learns company procedures and enough politics to operate effectively), then the new hire isn't worth the extra cash. If the senior engineer doesn't know the new technology and can't pick it up fast enough, then the new guy deserves more. At least until the senior engineer catches up.
NOTE: I don't mean to imply a senior engineer knows all new technologies, nor that they can alway guess what might be important to their company. Just that they can and do keep up to date on things, and sure sometimes a new library or language they missed or dismissed as "no better then this other thing I already know" becomes more important then they had guessed, but they ought to be able to pick it up. For example prior to the iPhone becoming popular it might have been reasonable to not know any Obj-C. Someone coming out of collage might know Obj-C, but a senior engineer with knowledge of C and any sort of smalltalk style OO language should be able to pick up Obj-C really fast. Likely fast enough that they are able to make real programs while the new guy is still learning the ins and outs of the bug tracking system, and who to listen to and who to ignore in meetings.
My personal pet theory is yes it really does matter. For a long time doubling every 18 months was the arbitrary goal, and Intel could say "if we don't hit X by Y then AMD will overtake us because they double every 18 months", and AMD could say "If we don't hit X by Y then Intel will smash us because they double every 18 months". So each poured whatever they needed into R&D to make it more or less happen. Sometimes one got more ahead then the other and got to roll in the hay for a while (or be unable to fill all the market demand).
If Moore's law wasn't sitting their prodding them to double or die every 18 months my guess is they would have sat there and gone "Will they REALLY invest $12.6 billion dollars on a new fab? That seems stupid, I bet they will try to get by with a 10% bump from a new microarchicture, so we should aim for 15%". We would have seen a LOT more severely lopsided product matchups, but overall I'm guessing slower growth.
Not that it matters so much now since they don't seem to be able to keep up with doubling processing capability (unless your problem set is very parallel...and even so Amdahl's Law will get them sooner or later -- add as many CPUs as you like, at some point there is a non-parallel part of your problem space and that part's performance will dominate, even with an infinite number of CPUs).
A dock connector gets you line level output which needs less fiddling to get acceptable results from. I'll also bet there are far more people with a iPod dock connector on their iPod/iPhone then there are folks that walk around with a 3.5mm male-male cable.
However I think a 3.5mm line-in would be a good thing to have in addition to a dock connector.
The default UI can be told to make a sound and flash the edges of the screen when you get "close" to pulling aggro (90% or so I think). Not at all perfect, but will do in a pinch. The tank is more in need of an addon, as nothing tells them when someone else is getting close (as far as I know).
I use to use grid+clique for that, but with a minor change you don't need addons:
Macro: /cast [@mouseover] flash heal
Macro: /cast [@mouseover] shield
Put flash on the 1, the shield on 2. Mouse over whoever needs that flash and press 1. No "click then press". It works nicely (you can also even set it up to use target or self if there is no mouseover).
I've discarded clique. I still use grid as it can be configured to display a LOT more information in a small space then the built in unit frames, and just as importantly it can be configured NOT to display information you don't care about.
Stock raid unit frames group by party (5 players) within the raid, put one healer per party and have them heal their group. Doable with stock unit frames. (You could also "cheat" and put groups of five in the same vent channel and have them yell out to their healer, that'll work for a whole lot of things)
I prefer grid, configured to stick a big center icon on for the debuf, but that is a preference, not a requirement.
Sure tanks and healers have a harder time in raids without mods then dps, but you can raid ICC without mods as any of the roles.
[...]what concessions do the automakers get for this service? [...]
Well, I'm not sure I believe any of that, but I can see a reason automakers might be down of diesels if they had the chance... a diesel engine tends to last way longer then gas engines, and it is to the advantage of car companies.
That said, conspiracy theories are fun to think about, but the real reasons behind stuff tend to be more complex and less fun to BS about.
About 2 out of 3 gas stations around here (Bay Area CA) don't sell diesel. I'm assuming some combination of making the equipment cheaper, maybe environmental licensing and inspection cost savings and such.
Ok, how about some related arguments:
That won't apply so much if you only ever read the same half dozen books, or if you never reread books. (personally the last trip I took it was nice to have a fair number of books to choose from but only have to budget space for them...however for me it the bigger value is always having a book with me, bank lines, restaurants and such -- I have read more books per month since starting on ebooks, and I've been a big reader all my life)
Thats cool. Some folks like paperbacks and some like hardbacks. I happen to like ebooks.
Absolutely right! We ought not take payment from anyone who is sick! Experimental medicine? Feh! Who needs that crap! I say we should have stuck to bleeding people. However as long as we have developed some form of medicine we better not get any better at it. Also, let's be sure to treat sick people like they are babies, and other people always know better.
I mean, if I get a terminal illness I sure hope nobody is allowed to offer me any experimental cures. If it hasn't been FDA approved, and in use for at least a decade it would be far better to die then take that kind of risk!
Ok, so there were no significant "give backs" from MS, but on the other hand they didn't start out with some form of mutant TCP/IP that required the rest of the internet to bend over backwards to sort of interoperate with. So that is still a win in my book.
Debugger speed matters when you debug something large. I frequently have things where gdb takes close to a minute to load symbols.
Debugger speed also matters for things like watchpoints. Watchpoints in gdb are so slow I almost never use them (like not at all this year, and maybe once last year).
Last time I have fast watch points (which was 1991, I had a CPU emulation box that ran at the full 68k speed with up to 8 watchpoints, and could dump the prior 200 cycles of bus activity and internal register state) I used them a LOT. Like three times a week.
A faster debugger might also make integrating a debugger and a unit test environment simpler. "run this code, fail the test if function X is not called, or if function Y is called, or if function Z isn't called exactly 8 times..." As it is there are some tests that can't be written using the API, you need scaffolding, and that scaffolding can introduce or suppress it's own errors...
Fair enough, I'm not an author or a publisher, so I don't know how they really work. It is the model movies use. Theater release (in fact in 4 different runs), Cable TV release, DVD priced to rent, PPV release, DVD priced to own. Oh, and digital downloads are in the mix somewhere. Foreign and domestic releases as well. All timed to try to extract the maximum revenue, all the same or nearly the same actual product.
I'm not so sure. When I get to the end of a book I like, if it is part of a seres I'll look at the next book. If it is a hardback I make a choice about buying it now at that price vs. reading something else now and checking back later to see if it has hit paperback. If I like the book enough (or the ending was enough of a clifhanger) I'm more likely to pay the hardback price.
Since I have switched to exclusively ebooks (partly because they really are more convent, and partly because I'm planning to retire to an RV) I'm less likely to pay hardback prices then I use to, but it does happen. It would happen more if the choice wasn't $30 now vs. $5 later, but $10 now vs. $5 later.
You are right that people don't have a fine grained idea of a price they are willing to pay for a given book. However if you give people a price for a book they tend to be able to very easily decide if they are willing to pay that much. Generally people are better about making a "buy/no-buy at $X" choice then a "what would you pay for this?" choice.
Most people are aware that (many) things will be cheaper if they wait ("wait for a sale, those things always go on sale"... "new movie for $30, or I could wait two months and I bet it will be $10", "new high end laptop now, or next years mid range"), but they still buy stuff now when they feel like it. The interesting question is how transparent can you make the "price decay" part before people stop being happy when they get something for the price they want, and start being resentful when they pick their price (or decided to wait).
Yes if you drop prices week by week some people will wait until next week. You may need to make the process a little more random. Or you may find that it really works just because you pick up a lot more sales somewhere in the middle of the curve that otherwise waited for paperback prices.
Unfortunately as long as it is "might be better", "could get higher margins", and even "could get higher margins AND more total sales" publishers will rightly say "um, dude, we have a book publishing model that makes money why do you think we want to experiment with it?". They are totally correct, I find the problem space interesting, and would be pleased to come out with a better book market to buy stuff from, but at the end of my day my paycheck comes from outside the book industry, and they have a paycheck that comes from entirely within it. It would be insane to just tinker with stuff because it "might" get better since it "might" get worse, and it isn't too badly off just now (note that is books, not newspapers or magazines).
There are some people that are extremely space constrained (fulltime RVers for example) who won't buy paper books, but I would't say are freaks. Well, technically they are more weight constrained then space constrained for books. I don't think many of them want to pay hardback prices for ebooks, but some will. The people that pay hardback prices for ebooks tend to be the people who were aware of why hardbacks really cost more and come out sooner. They are just a way for book publishers to rent seek (get the highest price people are willing to pay). That could actually be done even better with ebooks. Start the price at hardback level, and decay the price each week, hitting a somewhere under paperback price about a month after the paperback is out.
In a perfectly rational world it would work well, people would buy when the price hit what they are willing to pay to read it now...and since it has more steps then the current scheme many people that would pay a little more to read it "now not next month" but not "a whole lot more" can actually do that (this works better with a series of novels that is still being added to, or a author with a distinct style). In the real world however it does run the risk that it will just piss people off because they don't like rent seeking except when they are the seeker.
Or more to the point, each layer of middle men need to add value. Between author and reader the author wants someone to deal with all the bother of collecting money, the reader wants someone to deal with filtering out all the obviously bad works, and either or both the author and reader wants someone to fix up the language usage and the other nice stuff the publisher's editor normally does. There are some other tasks in there that are useful as well.
There is no need for that to be spread across two middle-men (publisher + bookseller), but there is no special need for it to be one middle man vs. one per operation as long as having 12 companies touch the book doesn't make it cost more then only having 1 or 2 touch it.
With the current state of things the publisher filters out the worst of the bad writing (that is to say 99.998% of what they get), fixes up the english usage, remind the author when the sex of one of their characters changes midway through the book, when things get talked about in Ch4 but don't happen until Ch7, and all that kind of nice stuff. Then Amazon makes sure the book got nicely "kindle formatted" puts a price on it, handles some of the money, arranges network service for Kindle devices (iPod touch and iPhone users get to manage that on their own), handles long term storage of customer purchases.
All of those are valuable. If we stuff all that into one company that charges me less, I'm happy. If we stuff that all into one company that charges me the same amount, I'm still happy. If we break that out into 1 companies that charge me the same or less, well I'm still happy. If we charge more, I get cranky. If we do less stuff for the same money, I get cranky. If we charge less for less I probably still get cranky.
So I'm less happy by the proposal of a bizaar. Sure I probably pay less, but did the author hire a proofreader? How do I find the tiny number of authors that don't suck? It would work out well for authors who are already successful under the existing system. I know some of their past works weren't crap, so their new works have enough chance to be not crap that I can "afford" to spend the time to read a review or just buy the book. I know they have published stuff in the past, so they will know the value of a good editor and hire one. I might even pay less. It will totally suck for knowing what new authors I could buy from and not run into extremely poor writing, or a good story but no proofreading....
Solvable problems, but mostly by adding middlemen back in (for example, tags for "professional editing", tags for "slushpile review" and the like). They may not be middlemen in this model, but they still operate on the text, and then a tag gets affixed to it, and I filter out stuff lacking the tags. How they get payed might be the only difference (plus the "in theory" difference that people could choose to buy stuff that didn't pass a slushpile review, or wasn't edited by anyone....but I doubt enough people would do this to make dismantling the current system a good idea!). If we move the way something gets a slushpile review we could also break the value of that system. Currently a publisher does that, and if a publisher culls too many books they can lose good sellers. If they cull too loosely then they run the risk of people deciding "oh they publish crap, I won't risk reading that". If the "publisher middleman" goes away and an author has to pay for a slushpile review then they have incentive to pick registered slushpile reviewers that rubber stamp the most content. There may be ways to fix that, but struggling to get "as good as the existing system" while offering no real advantages doesn't sound exciting.
One isn't a substitute for the other. I don't let my dogs wonder, they go in the yard when I go in the yard to watch them. However over the years their have been some escapes. The fence gate has twice in five years managed to be open, or unlatched enough that a dog has opened it by pushing against it. Herding up six escaped dogs is tricky. My fence is 6.5 feet tall (legal limit for where I live). At least two of my dogs can jump over it, they just don't know it, so they only leap it when something extra interesting is on the other side. Then I have to shove five dogs inside, and go fetch the escapee. I had a (very minor) fire. I evac'ed all the dogs to kennels I keep in my van. One of the dogs popped the gate off the kennel (it is a cramped fit, so the kennel is a little distorted, and the gate doesn't fit perfectly...but she had no risk of being burned alive). I'm pretty sure I've even brought dogs in from the backyard, and had them run out a house door someone had open. Once or twice, they have hopped a baby gate and gone out a door I was bringing things in from (I normally kennel them if I'm bringing much in and out of the house though). When camping I've had dogs escape from tents twice too.
I know the moment my dog is missing (or you know, in a really short span of time). Long before they freeze anywhere. However the longer it takes me to figure out where they have decided to go exploring, the bigger the chance they become road kill. Or in the case of my small dogs, get eaten by something. Normally I recover them within a few minutes. A few times it has been a half hourish. Once it was about two hours (I was living in an area with huge yards with lots of trees, 4+ acres per house, hard to find a dog that escapes your field of view).
So yeah, a pet tracking GPS could be very useful to me. If they weren't so amazingly expensive (monthly plan!), so unlikely to be charged when I need them, and too big for at least 2 of my dogs anyway. I'm sure there are a lot of responsible pet owners who would find them handy. The same folks who think of it more like having them ID chipped "this will help me recover them if they go missing" would. Not everyone who wants a GPS for their dog plans on just opening the front door and letting the beastie entertain itself until dinner time.
Does Linux have a seporate report for "cached dirty"? Most Unixes don't. So some (or even all) of it could be data that has been written, so it can't be reallocated until it gets flushed. I would guess it is very rarely "all", or even rarely "most", but that'll depend on your workload.
That isn't "battery memory" any more then "mail from my boss about me using the wrong TPS sheet is spam", or "my video game drops to 10FPS is crashed".
It is most definitely a "battery problem" (maybe even an unexpected one), but it isn't "battery memory". We use specific words to mean specific things so we can talk about problems and solutions rather then having a lot of words that just mean "DO NOT WANT". So if someone says a specific laptop has a battery memory problem and they really mean memory problem then you can avoid that by doing the inconvenient thing of making sure that whenever you use the battery you run it down to zero. If they really just mean "it only has 300 full charge cycles" then treating it like a battery memory problem is so totally the wrong thing -- their warning has caused you to destroy your battery far far faster.
(yes, some batteries have limited numbers of charge cycles, NiMH batteries do, but no memory effect, at least here on earth)
Either you are very lucky, or I'm unlucky, or both. Many times my Treo has fallen out of my shirt pocket onto asphalt and wacked one of it's keys hard enough that the bit of tinfoil it has under each key gets dinted backwards. That leaves a treo that has no keyboard, and is apparently spending so much of it's CPU dealing with keyboard interrupts that it is pretty unresponsive to the stylus.
It is an easy fix, I (or more likely my wife) open the case and move a circle of foil from my "dead, for parts" 650 into the wounded 650 to 700. However the Treo is basically unusable until the fix is made. I'm running low on foil bits now, and getting the supply cost me dearly. Fortunately I broke myself of the habit of keeping the phone in a shirt pocket except while driving.
Easy enough, places that want to ship a product (assuming the interface libraries are GPL not LGPL). Or just places that want to ship a product with the database embedded inside it. They might be better served with Postgress, or sqlite. Or licensing Oracle or something (of corse if licensing a database is an option mysql can do that).
Which doesn't mean the article wasn't kind of weak, and half the points aren't lame...but this one at least made sense.
I have read a few sections of the book and there is a lot in it that isn't After Effects specific (or even DV specific). Things like filming from the bed of a pickup truck to avoid the need to get a film license from many local governments (most places - apparently - only require a license if you put your tripod on the ground -- so you can do handheld and get that shaky "this ain't hollywood" feel, or you can put your tripod in the pickup and roll). I haven't read many books on making a film though, so I don't know how good it is if you have already read other "make a film on a budget" AND you don't use After Effects.
If those shows are normal "over the air" broadcast, and they come to you with the original commercials, then whatever laws the cable companies used when they started up are on your side...unless those laws have been repealed or superseded.
However if the shows had the commercials stripped from them, or were from HBO or something, then there isn't as much on your side (I'm not saying it is illegal, I wouldn't know). Does anyone know of any bittorrent feeds of TV shows that still have commercials in them? They may be bigger files, and more awkward to watch, but hey, but far more likely to be legal, and definitely morally no different from what cable companies did 20~30 years ago.
Have you tried timing the mass data move with FW400? There are a lot of disks that can saturate FW400 on paper, but don't in practice (or do, but not by a whole lot). If you have I would love to know what the actual speed difference is (I recently bought a FW400 enclosure to do backups, and would like to know what I'm missing since I have FW800 on my laptop...not that I think I'll care for backups).
Not really. The iBooks when announced were slower then the PowerBooks. The iMac is slower then the PowerMac (well, except for the intel iMac)...
Well a lot of the good points. But, it isn't pervasive like a real GCed language, so you may have to interface with code that doesn't use it (and has leaks or unhappy ways to use it). It also actually has a higher (but more predictable) cost.
It also has some advantages over GC the perceptibility isn't just for CPU load, the exact lifetime of objects is predictable, which is very useful for RAII. E.g. a shared_ptr to a lock object will destruct exactly when the last reference goes away releasing the lock...on a GC system the destructor fires "sometime" after the last reference goes away. Very useful with exceptions and other complex control flow.