There is a basic problem asking a consulting group to send source code. Generally they don't actually "own" much. And code they may have may really belong to their other clients. I have some code that is in the BSD printer subsystem of Dec's Unix offering, but I am not sure DEC shipped the source it. OSF members could look for it, but the entry fee is a bit steep. Many long term consultants don't actually write code they keep. Although if you can find a card reader I could locate lots of code from school nad shortly after if you can find an IBM 96 column reader too. Would you want someone to send your sorce code to your potential competitor's as a sample of their work? Not likely. If pressed we could do a mutual NDA and my company could reelase some internal project code samples, but it won't happen easily without some expectation already established that this is just a form of final vetting to assure the last checkbox item to get the contract. Our latest DRM offering has a mix of our licensed IP and the cutomers IP in the product we wrote for them as an example. So even though we wrote the code and it contains our video processing IP there is IP that belongs to the client and so we could not release that as a sample without breaking our contract obligations. I agree with 1 and 2 and have a positive personal bias for #3 but that is not a real qualifier to determine a groups skills.
Blatantly connectted to me. I own 1/3 of it. We have 3 core people and lots of known contractors we work with regularly. All three core people are the former staff Scientists at xSides Corporation and are into all phases of development. They have extensive embedded and set top box experience, OS dev experience, applications framework development, image processing and video processing and well, the list is pretty long. Take a look at the site Cryptocybernetics.com for more of the flavor of who we are. And, oh yes, we have IP in the security and DRM area. We also have the start of an office in Ukraine (not open just yet) so we can have a not-in-US presence for some work. Pick any of the email contacts and ask about CryptoICS our consulting operation. We don't mention it on the site at the moment (we are bringing a site dedicated to our consulting up but it is not reachable). Our client list is pretty favorable though.
If you are the government it is OKAY to violate patents or uphold the interests of those who do. The Feds should "belly up to the bar" and put up some cold hard cash to NTP to allow the system to continue and require (if feasible) all non-government users shutdown, and only, and this is a big one, only those government officials blackberries should be left functional if their use is in the interests of national security. Let the rest of the Fed-o-crats suffer like the rest of us. If the government does not respect the rights of intellectual property holders what does this say to the citizens and ip creators in this country about how to behave or what they can expect.
Digital made a switch to a vectored kernal for later RSX-11 M Plus releases and it was "A Good Thing". It leads to more stability not less. Linux could do the same. It would not be all that hard to do, and although there is one additional layer of indirection, with current design patterns that would only affect the initial loading of a driver. Other options are to create a self linking driver format where the drivers are JIT linked when loaded. The kernal would be responsible for loading the drivers own jump table based on what it could provide. A minimal set of required routines and the driver could then see if it had enough information / routines / support and unload itself if it was not in an appropriate environment, or could provide support that some compiled kernals have and others omit if it was not present. Old technology made new again through lack of use!
The chaos that would surely happen if the UN were to control the IP adn DNS roots and allocations would be interesting. And not in a good way. Think of a UN resolution to require UN membership to get an allocation... Or abide by this treaty of face being cut off.
The US is a pretty graceful steward of these services (that our tax dollars paid to develop) and we have a state interest in their smooth administration. So much so that registrations that used to cost $100 for two years are now under $7 (usd) per year. And I can register a domain through my favorite Teutonic registrar. You know the clown one. And while I pay a high price to get IP allocations from our local authority (non-profit my butt! Any justified allocation should cost the same then, and no charge for changes to the record. It is not different than domain registrations.), the controlling authority also allocates IP space as requested (and justified) for the other regional concerns. Without a profit motive. I send more money overseas for registrations to help profits for the "clown" (they are great by the way, we resell them here at our ISP) than most any other single line item in some months, save our backbone connection charges.
And any country that feels the US has too much control, hire me as a consultant. I'll show you how to set up your national infrastructure so the US control can be ripped away in moments to form your own isolated nationalistic infrastructure. And recommend technical regulations to support that infrastructure so that access to such critical items is virtually forced (and how to actually force it if they like). Of course if the non-US/US-haters fragment the Internet we'll be back to the days of gateways like gateway.dec.com or other neccessary evils of the past. Don't get me started on the use of '::' to separate nodenames to force routing!
By the way, if DECnet Phase IV and V had won over IP this argument would be moot. No addressspace issues either.
He also says "There are Open Source zealots who believe that unless an application is part of a stack which includes 100% Open Source services and components, that it can not claim to be Open Source. [...] But does this "stack" argument actually make any sense?"
Only if the BIOS for your computer is also "open source". Or maybe only if the hardware is also open source. That leaves a pretty small world for the open source zealots.
Its the iTunes stupid. (no/. reader is being called stupid BTW)
It is the iTMS and iTunes ease of use. And agressive marketing. If the Palm Life drive had a virtual cell phone keypad, GSM ability and the iTunes app, I'd buy one now. Right now. I have a large iTMS collection so want / need iPod-lite at a minimum iTunes ability (which I can get through JHymn and a genric player app). But the real iTunes app is just nice to me. It an opinion thing of course, but apparently shared by quite a few people. And, an 80 GB drive in a LifeDrive would be nice. Build in the equivalent of the camera connector and a portable iPhoto app and Palm could have a great product. Of course Palm has now aligned with the dark empire forces.:) It is all just opinion. But the opinions that spend win. I have purchased 4 iPods at this point (originally I had an SSI America MP3 brick of a player). But a Lifedrive sized device with a SIM card slot 4 band cell ability and just require a wired or bluetooth headset. Add the existing LifeDrive PDA type apps, the QVGA screen, although real WVGA would be nicer and not affect the form factor too much, just make the controls soft onscreen controls. Add some better integration inside to make room for a 1.8 inch drive instead of the CF sized one and it would not make the device appreciably bigger than a current lifedrive. Use a small NAND flashdrive for the app and OS / boot device for speedy access and turn-on time. Voice call function for the cell phone would mean for the cell phone voice calls the brick sits in your pocket almost all the time. Just use the headset to initiate the calls. It is possible to build this now, technically. It may not be possible to get all the players needed to come to the table and license the technologies needed at a reasonable rate. I am pretty sure I could bring such a device to market (and I am available for consulting!) for less than the combined price of say a Lifedrive and premium (like a Razr or V635 or similar Nokia or other) cell phone. And save the cost of an iPod. Even paying a hefty fee to Apple per shipped unit for license of the iPod engineering and design. Any one want to hire my company to get it done???:)
People won't pay to develop good code. Period. There is no demand for perfection. I was part of a three man team that wrote a prototype media viewer for early release movie content. We provided the backend encrypter application and ancillary libraries under license. Our "proof of concept" was finished in 8 weeks and was so successful that we had our code in live in air airline flight tests with real customers. Awesome work. Very stable. The encrypter application we wrote had a few issues with some non-standard compliant streams from a common encoder suite, which we didn't write. We charged T&M to fly an engineer to the content providers location (they don't like to ship unencrypted media for obvious reasons, and we wanted to do side by side compares for the streams to debug the problem). They already signed off on the PROTOTYPE delivery. They balked at us billing for the post delivery work and turned us down for the follow up contract. The fee schedule for post delivery work was a part of the original contracts. Needless to say they aren't in that business anymore and sold the division in question.
So far our original code for the prototype remains with no bugs outstanding. They just can't encrypt all the possible movie bitstreams they'd like. The same team put an applications development framework together in 3 months of very long days with very few bugs against it. A lot of those are documentation related. The relevant company no longer is selling anything but intellectual property now, but the last code shipped was very solid. about 70,000 lines of code and about 15 total real bugs.
My point is that no matter how good the engineers, there is a cost in time for projects you are passionate about, and a cost in real dollars for real engineers with appropriate architecture and development skills. Most companies don't want ot spend the money. If customers would pay more money for products then companies would pay more to develop them. Customers drive the price structure, and that drives the expenses a company will invest to develop the product. Add to that that a "bandage" patch is cheaper in current dollars that a major rewrite and you end up with large commercial products that are bandages on bandages on bandages. Refactor the code and it will almost always improve if the requirements and specifications changed since inception. Few companies will start a side development group to redesign and rewrite an existing product.
The articles author can just start a grass roots movement to drive the marketplace to only purchase warranteed bug free software. I don't predict success.
the close window control is supposed to close a window, and not shut down the application. you may be used to windows, where closing the last open window also shuts down the app, but many ux peeps will tell you this is not a good assumption to make:
Yet iPhoto closes when you close the window... It is annoying that Apple chose to make single dialog apps behave different in this respect.
The music player choice means it is always with you. If you have a waterproof "sport" case for it even better. A reason to get that new 80 Gig iPod (wait for a few days) as a corporate expense. It is your backup drive.:) Just put a virtual encrypted disk image like PGP Disk had and off you go. Drag and drop your life.
As long as the industry first varies the ASCAP fees charged for music in the same direction as they expect Apple to charge for different tunes. More popular, charge more. I don't think they will ever very the ASCAP fees charged on the basis of the popularity of a song.
ASCAP represents tens of thousands of copyright owners and millions of songs and an ASCAP license will give you the right to perform them all.
So one license fee based on the business model of the business performing and that grants license to all ASCAP material regardless of popularity of songs, artist, etc. One of their organizations business models. They really are greedy and want ot renegotiate the deals already in place that looked so good when Napster was their nightmare and iTunes their salvation.
It would be nice if they spent the extra 10 cents to make the black models back case plated with nickle then blackened. Ive's laser etch would stand out even more, it would be very nice indeed. Black earphones would be good too. For the complaints about the laser etch, the serial number is etched by the laser, the additional logo etch is essentially a freebie as is personalization from a cost perspective. The etch on the click wheel provides important tactical feedback, not just something to look at.
I owned some prior art. The SSI MP3 player (about the size of a house brick, almost as heavy.:)
I bought it because of/. reviews I think. I owned it before the filing dates for the patent. One assumes this means at least the filing date is after a public disclosure of the same invention. If it was marketed at least a year prior then Creative loses all claim to the invention/patent, I think. Not a patent attorny, but deal with them often enough.
The new stuff is r-134. The chips are semiconductor chips that when current is applied exhibit the peltier effect. One side gets warm, the other cooler. Essentially a solid state heat pump. No compressor, no liquid refrigerant needed. Instead just blow air over the device and its "cold sink" (same essentially as the expansion side air handler for a liquid refrigerant system in principle). So fewer moving parts. Especially the blasted compressor clutch assembly which in some cases makes it cheaper to replace the whole compressor with a rebuilt one than separate the clutch from it. The clutch causes the pulley to spin freely and not drive the compressor when cold is not demanded by the air temp controls, hopefully thermostats, but n ot always in cars.
in your disk drive analogy, it would be like coming up with a cheap flash drive that beat the specs for lifetime and cost to those spinning magnets you mentioned. It makes it last longer by eliminating wearing spinning parts that rub against each other roatating and moving up and down and up and down... Not sure how the parent was maked "insightful" but there is no "lacks relevant technology background". Seriously though it is fascinating technology and along with heat pipes makes overclocking much less of a mess.
There are already individual keys that are graphic displays of a size appropriate for a keyboard. So the innovation is arrangement? Use of OLED instead of LED or LCD? I wanted to do this years and years ago. Long enough that had I patented it, it would have expired (I wanted to change between qwerty, dvorak and things like chinese or japanese characters at the time, although glyphs for tasks were in my write up as well). For a full keyboard it just isn't practical.
Of course the main reason it is not practical is the cats would just it as a heating pad because of the amount of energy it would concume. ANd when your display went to sleep would you also 'sleep' the keyboard?
I wish them luck as I really want one of these. (currently typing on the Apple cyrillic keyboard where the character sets are both in the same shade of gray for some reason.)
And all its kin around the world. I Thomas Edison had to do all the current paperwork and pay the current fees (even adjusted for inflation) he'd not have a tenth the patents he had. I can just imagine the 4 month delay when the Patent Office asks Mr. Edision to clarify the difference between his claim of a wire support for a filiment made of carbonized cotten from the previous claims of a wire inserted through the glass for the "florescent" light previously patented. And having to pay the continuing support fees to protect his patents would have bankrupted (more than it was / faster) Menlo Park. Also he is looking at innovation per billion people. More people mean more people will do simultaneous inventions like Fulton's steamboat and whathisnames and Newton's Calculus and whatshisnames (well I do know the names but you get the point).
Maybe the rate of innovation has exceeded the rate that is easily understood by one man.:)
It is still not an OS. What you need is a layer over the OS for UI for blind, etc. users. The OS is the OS. The UI is the UI. Only a few OSes have been so structured towards visual displays that the primary concerns were video related and core to the OS (or even lower!). Really what you want is nothing more than, for example, something that uses mouse gestures, a data glove, etc. and provides audio or a braille bar for the feedback. The OS just thinks those are peachy keeno devices. They still aren't part of the OS and neither is the application installer. You just need to write the stuff above the OS. The closest you come to OS work is the device drivers for the required interface devices accepting input and providing feedback. I don't look for this to be a hot seller though as the limited market raises the price to a premium level. Especially if someone sells investors on the need for a whole new OS when Linux, BSD, Mac OS (I know that's redundant), or even Windows suffices. Even of them can be pretty easily configured to use some different UI rather than the one(s) shipped with them. Thogh Windows requires the most work, it is still not all that hard.
If the cable Internet companies are not "Common Carriers" do they lose the common carrier exemptions in the laws? Can we now go after them for the damages caused by the machines they host on their "information services"?
Even more complicated would be using a single NIC to connect two operating systems to the same network. Unless someone came up with a clever solution, each OS would need its own IP address. However, routers and switches outside the computer would become immensely confused when a single NIC and a single MAC address belong to two IP addresses, since most routers/switches only have a one-to-one correlation between MAC addresses and IP addresses.
Since when! Even my little 5 port dumb switch can handle 1024 MACs, and unless it is a very expensive switch the concept of a switch even being concerned about IP addresses is a little off. Consider that current off the shelf infrastructure easily handles things like Cobalt RaQs where you can have dozens to hundreds of IP addresses for the single MAC address. No problems because of it. Routers are only concerned with the MAC an IP came from so they route the packets back the correct path, they don't really care much (for basic protocol handling) about the actual original MAC. MACs don't exist once you transit off ethernet anyway in most cases, so IP over WAN routes to local ports (in normal cases) based on where the traffic came from. There are also alternative protocols of course for load balancing and least cost routing as well.
The easy way is to virtualize all the hardware in a hardware abstraction layer. NT and later (muddied a bit in the graphics area, but that is DirectX at work) already have this on windows it would be easy to target one to a single virtual machine instead of the generic chipsets or in addtion too them. In OS-X86 just do this at the interface to BSD (OSX is BSD over MACH still, right? Not teasing legitimate question!) So in summation, the network stack of ech OS would communicate to virtual virtual hardware and the software implemented virtual hardware talks to the real interface. It is really not a complex deal.
More likely however would be a micro kernal (like MACH) that supports SMP running the OSes as privledged tasks. And ideal would be to sandbox them as much as possible so a blue screen of death would not kill the OS-X86 task. Reinvent some old Apollo (now HP BTW) technology where the display manager allowed emulation of multiple varieties of Unix interfaces and systems simultaneously. The industry already has virtualized the display technology in enough ways to be useful. You do have some unique issues with UI though, as Mac OS has a common top of the screen menu bar and users might be confused when it winks in an out as you select a Mac OS hosted window then switch between that and a Windows window. But solutions exist for that too. Of course you could just write a Windows NT style HAL and run Windows as a task under Mac OSX and be done with it. Conceptually it is not that hard, it just takes time and good design.
While you may like the concept of thinking of multi-core chips as SMP on a chip, in a lot of cases teh designs don't allow the OS to set affinity for a particular core, the hardware on the chip decides which core executes a particular instruction stream. So it is better to use hardware virtualization if available of support it in software. Otherwise you may be tied to a specific chip family, or chip, or maybe even stepping of a particular chip.
iMacs, and iBooks are not allowed to have secondary displays except in mirror mode. Apple's official stance. What does it take to enable the external video port on many models that use a dual display capable chip so that more than just "mirror mode" is possible? Just a quick trip to OpenFirmware and change the settings on a few bits. Could Apple prevent this. Sure, a quick trip to system profiler says I am on an iMac G5 and my serial number is right there. (Heck by Pismo powerbook had my sales order number since it was a built to order machine.) All they'd have to do was check this at boot time and disable the dual display ability. Maybe even a "spank me" by assumming excessive power consuption and heat generation and just turning the fans on "high". Annoying but non lethal. But no, they don't do a thing because they don't like to annoy their customers. And their customers that do these things know it is unsupported. So if I ever do this (wink wink nudge nudge... know what I mean, know what I mean) I'll be sure to remove it before servicing and hope that open firmware doesn't keep a log somewhere of these "bad" choices a naive user might make.
Remember iMacs and iBooks don't support dual displays, only mirror mode. Riiiigggghhhhttt!!!!
Considering Design Patterns arguably originated with Charles Alexanders work in architecture (for house, communities, etc.) I find the use of a badly built house to describe badly written software appropriate. Without specifically mentioning design patterns the author derides the use of doors that don't behave like normal doors (although I think European outside doors open outward, my Ukraine apartment door did), and other features that don't share the normal characteristics of the items they are supposed to be. Exactly the areas were design patterns solve problems.
Pretty cool commentary to push people toward design patterns if only they would have pushed people toward design patterns in that commentary!
Slashdot Mirror
I ate lunch with Donald Trump, I knew Donald Trump, Donald Trump was a client of mine. Mr. Gates, you are no Donald Trump.
There is a basic problem asking a consulting group to send source code. Generally they don't actually "own" much. And code they may have may really belong to their other clients. I have some code that is in the BSD printer subsystem of Dec's Unix offering, but I am not sure DEC shipped the source it. OSF members could look for it, but the entry fee is a bit steep. Many long term consultants don't actually write code they keep. Although if you can find a card reader I could locate lots of code from school nad shortly after if you can find an IBM 96 column reader too. Would you want someone to send your sorce code to your potential competitor's as a sample of their work? Not likely. If pressed we could do a mutual NDA and my company could reelase some internal project code samples, but it won't happen easily without some expectation already established that this is just a form of final vetting to assure the last checkbox item to get the contract. Our latest DRM offering has a mix of our licensed IP and the cutomers IP in the product we wrote for them as an example. So even though we wrote the code and it contains our video processing IP there is IP that belongs to the client and so we could not release that as a sample without breaking our contract obligations. I agree with 1 and 2 and have a positive personal bias for #3 but that is not a real qualifier to determine a groups skills.
Blatantly connectted to me. I own 1/3 of it. We have 3 core people and lots of known contractors we work with regularly. All three core people are the former staff Scientists at xSides Corporation and are into all phases of development. They have extensive embedded and set top box experience, OS dev experience, applications framework development, image processing and video processing and well, the list is pretty long. Take a look at the site Cryptocybernetics.com for more of the flavor of who we are. And, oh yes, we have IP in the security and DRM area. We also have the start of an office in Ukraine (not open just yet) so we can have a not-in-US presence for some work. Pick any of the email contacts and ask about CryptoICS our consulting operation. We don't mention it on the site at the moment (we are bringing a site dedicated to our consulting up but it is not reachable). Our client list is pretty favorable though.
If you are the government it is OKAY to violate patents or uphold the interests of those who do. The Feds should "belly up to the bar" and put up some cold hard cash to NTP to allow the system to continue and require (if feasible) all non-government users shutdown, and only, and this is a big one, only those government officials blackberries should be left functional if their use is in the interests of national security. Let the rest of the Fed-o-crats suffer like the rest of us. If the government does not respect the rights of intellectual property holders what does this say to the citizens and ip creators in this country about how to behave or what they can expect.
Digital made a switch to a vectored kernal for later RSX-11 M Plus releases and it was "A Good Thing". It leads to more stability not less. Linux could do the same. It would not be all that hard to do, and although there is one additional layer of indirection, with current design patterns that would only affect the initial loading of a driver. Other options are to create a self linking driver format where the drivers are JIT linked when loaded. The kernal would be responsible for loading the drivers own jump table based on what it could provide. A minimal set of required routines and the driver could then see if it had enough information / routines / support and unload itself if it was not in an appropriate environment, or could provide support that some compiled kernals have and others omit if it was not present. Old technology made new again through lack of use!
The chaos that would surely happen if the UN were to control the IP adn DNS roots and allocations would be interesting. And not in a good way. Think of a UN resolution to require UN membership to get an allocation... Or abide by this treaty of face being cut off.
The US is a pretty graceful steward of these services (that our tax dollars paid to develop) and we have a state interest in their smooth administration. So much so that registrations that used to cost $100 for two years are now under $7 (usd) per year. And I can register a domain through my favorite Teutonic registrar. You know the clown one. And while I pay a high price to get IP allocations from our local authority (non-profit my butt! Any justified allocation should cost the same then, and no charge for changes to the record. It is not different than domain registrations.), the controlling authority also allocates IP space as requested (and justified) for the other regional concerns. Without a profit motive. I send more money overseas for registrations to help profits for the "clown" (they are great by the way, we resell them here at our ISP) than most any other single line item in some months, save our backbone connection charges.
And any country that feels the US has too much control, hire me as a consultant. I'll show you how to set up your national infrastructure so the US control can be ripped away in moments to form your own isolated nationalistic infrastructure. And recommend technical regulations to support that infrastructure so that access to such critical items is virtually forced (and how to actually force it if they like). Of course if the non-US/US-haters fragment the Internet we'll be back to the days of gateways like gateway.dec.com or other neccessary evils of the past. Don't get me started on the use of '::' to separate nodenames to force routing!
By the way, if DECnet Phase IV and V had won over IP this argument would be moot. No addressspace issues either.
The RIAA closed down the "open source" alternatives to the iTMS (iTunes Music Store)...
Only if the BIOS for your computer is also "open source". Or maybe only if the hardware is also open source. That leaves a pretty small world for the open source zealots.
"/.ers" ???
Its the iTunes stupid. (no /. reader is being called stupid BTW)
:) It is all just opinion. But the opinions that spend win. I have purchased 4 iPods at this point (originally I had an SSI America MP3 brick of a player). But a Lifedrive sized device with a SIM card slot 4 band cell ability and just require a wired or bluetooth headset. Add the existing LifeDrive PDA type apps, the QVGA screen, although real WVGA would be nicer and not affect the form factor too much, just make the controls soft onscreen controls. Add some better integration inside to make room for a 1.8 inch drive instead of the CF sized one and it would not make the device appreciably bigger than a current lifedrive. Use a small NAND flashdrive for the app and OS / boot device for speedy access and turn-on time. Voice call function for the cell phone would mean for the cell phone voice calls the brick sits in your pocket almost all the time. Just use the headset to initiate the calls. It is possible to build this now, technically. It may not be possible to get all the players needed to come to the table and license the technologies needed at a reasonable rate. I am pretty sure I could bring such a device to market (and I am available for consulting!) for less than the combined price of say a Lifedrive and premium (like a Razr or V635 or similar Nokia or other) cell phone. And save the cost of an iPod. Even paying a hefty fee to Apple per shipped unit for license of the iPod engineering and design. Any one want to hire my company to get it done??? :)
It is the iTMS and iTunes ease of use. And agressive marketing. If the Palm Life drive had a virtual cell phone keypad, GSM ability and the iTunes app, I'd buy one now. Right now. I have a large iTMS collection so want / need iPod-lite at a minimum iTunes ability (which I can get through JHymn and a genric player app). But the real iTunes app is just nice to me. It an opinion thing of course, but apparently shared by quite a few people. And, an 80 GB drive in a LifeDrive would be nice. Build in the equivalent of the camera connector and a portable iPhoto app and Palm could have a great product. Of course Palm has now aligned with the dark empire forces.
People won't pay to develop good code. Period. There is no demand for perfection. I was part of a three man team that wrote a prototype media viewer for early release movie content. We provided the backend encrypter application and ancillary libraries under license. Our "proof of concept" was finished in 8 weeks and was so successful that we had our code in live in air airline flight tests with real customers. Awesome work. Very stable. The encrypter application we wrote had a few issues with some non-standard compliant streams from a common encoder suite, which we didn't write. We charged T&M to fly an engineer to the content providers location (they don't like to ship unencrypted media for obvious reasons, and we wanted to do side by side compares for the streams to debug the problem). They already signed off on the PROTOTYPE delivery. They balked at us billing for the post delivery work and turned us down for the follow up contract. The fee schedule for post delivery work was a part of the original contracts. Needless to say they aren't in that business anymore and sold the division in question.
So far our original code for the prototype remains with no bugs outstanding. They just can't encrypt all the possible movie bitstreams they'd like. The same team put an applications development framework together in 3 months of very long days with very few bugs against it. A lot of those are documentation related. The relevant company no longer is selling anything but intellectual property now, but the last code shipped was very solid. about 70,000 lines of code and about 15 total real bugs.
My point is that no matter how good the engineers, there is a cost in time for projects you are passionate about, and a cost in real dollars for real engineers with appropriate architecture and development skills. Most companies don't want ot spend the money. If customers would pay more money for products then companies would pay more to develop them. Customers drive the price structure, and that drives the expenses a company will invest to develop the product. Add to that that a "bandage" patch is cheaper in current dollars that a major rewrite and you end up with large commercial products that are bandages on bandages on bandages. Refactor the code and it will almost always improve if the requirements and specifications changed since inception. Few companies will start a side development group to redesign and rewrite an existing product.
The articles author can just start a grass roots movement to drive the marketplace to only purchase warranteed bug free software. I don't predict success.
the close window control is supposed to close a window, and not shut down the application. you may be used to windows, where closing the last open window also shuts down the app, but many ux peeps will tell you this is not a good assumption to make:
... It is annoying that Apple chose to make single dialog apps behave different in this respect.
Yet iPhoto closes when you close the window
The music player choice means it is always with you. If you have a waterproof "sport" case for it even better. A reason to get that new 80 Gig iPod (wait for a few days) as a corporate expense. It is your backup drive. :) Just put a virtual encrypted disk image like PGP Disk had and off you go. Drag and drop your life.
So one license fee based on the business model of the business performing and that grants license to all ASCAP material regardless of popularity of songs, artist, etc. One of their organizations business models. They really are greedy and want ot renegotiate the deals already in place that looked so good when Napster was their nightmare and iTunes their salvation.
It would be nice if they spent the extra 10 cents to make the black models back case plated with nickle then blackened. Ive's laser etch would stand out even more, it would be very nice indeed. Black earphones would be good too. For the complaints about the laser etch, the serial number is etched by the laser, the additional logo etch is essentially a freebie as is personalization from a cost perspective. The etch on the click wheel provides important tactical feedback, not just something to look at.
I owned some prior art. The SSI MP3 player (about the size of a house brick, almost as heavy. :)
/. reviews I think. I owned it before the filing dates for the patent. One assumes this means at least the filing date is after a public disclosure of the same invention. If it was marketed at least a year prior then Creative loses all claim to the invention/patent, I think. Not a patent attorny, but deal with them often enough.
I bought it because of
The new stuff is r-134.
... Not sure how the parent was maked "insightful" but there is no "lacks relevant technology background". Seriously though it is fascinating technology and along with heat pipes makes overclocking much less of a mess.
The chips are semiconductor chips that when current is applied exhibit the peltier effect. One side gets warm, the other cooler. Essentially a solid state heat pump. No compressor, no liquid refrigerant needed. Instead just blow air over the device and its "cold sink" (same essentially as the expansion side air handler for a liquid refrigerant system in principle). So fewer moving parts. Especially the blasted compressor clutch assembly which in some cases makes it cheaper to replace the whole compressor with a rebuilt one than separate the clutch from it. The clutch causes the pulley to spin freely and not drive the compressor when cold is not demanded by the air temp controls, hopefully thermostats, but n ot always in cars.
in your disk drive analogy, it would be like coming up with a cheap flash drive that beat the specs for lifetime and cost to those spinning magnets you mentioned. It makes it last longer by eliminating wearing spinning parts that rub against each other roatating and moving up and down and up and down
There are already individual keys that are graphic displays of a size appropriate for a keyboard. So the innovation is arrangement? Use of OLED instead of LED or LCD? I wanted to do this years and years ago. Long enough that had I patented it, it would have expired (I wanted to change between qwerty, dvorak and things like chinese or japanese characters at the time, although glyphs for tasks were in my write up as well). For a full keyboard it just isn't practical.
Of course the main reason it is not practical is the cats would just it as a heating pad because of the amount of energy it would concume. ANd when your display went to sleep would you also 'sleep' the keyboard?
I wish them luck as I really want one of these. (currently typing on the Apple cyrillic keyboard where the character sets are both in the same shade of gray for some reason.)
And all its kin around the world. I Thomas Edison had to do all the current paperwork and pay the current fees (even adjusted for inflation) he'd not have a tenth the patents he had. I can just imagine the 4 month delay when the Patent Office asks Mr. Edision to clarify the difference between his claim of a wire support for a filiment made of carbonized cotten from the previous claims of a wire inserted through the glass for the "florescent" light previously patented. And having to pay the continuing support fees to protect his patents would have bankrupted (more than it was / faster) Menlo Park. Also he is looking at innovation per billion people. More people mean more people will do simultaneous inventions like Fulton's steamboat and whathisnames and Newton's Calculus and whatshisnames (well I do know the names but you get the point).
:)
Maybe the rate of innovation has exceeded the rate that is easily understood by one man.
It is still not an OS. What you need is a layer over the OS for UI for blind, etc. users. The OS is the OS. The UI is the UI. Only a few OSes have been so structured towards visual displays that the primary concerns were video related and core to the OS (or even lower!). Really what you want is nothing more than, for example, something that uses mouse gestures, a data glove, etc. and provides audio or a braille bar for the feedback. The OS just thinks those are peachy keeno devices. They still aren't part of the OS and neither is the application installer. You just need to write the stuff above the OS. The closest you come to OS work is the device drivers for the required interface devices accepting input and providing feedback. I don't look for this to be a hot seller though as the limited market raises the price to a premium level. Especially if someone sells investors on the need for a whole new OS when Linux, BSD, Mac OS (I know that's redundant), or even Windows suffices. Even of them can be pretty easily configured to use some different UI rather than the one(s) shipped with them. Thogh Windows requires the most work, it is still not all that hard.
If the cable Internet companies are not "Common Carriers" do they lose the common carrier exemptions in the laws? Can we now go after them for the damages caused by the machines they host on their "information services"?
Even more complicated would be using a single NIC to connect two operating systems to the same network. Unless someone came up with a clever solution, each OS would need its own IP address. However, routers and switches outside the computer would become immensely confused when a single NIC and a single MAC address belong to two IP addresses, since most routers/switches only have a one-to-one correlation between MAC addresses and IP addresses.
Since when! Even my little 5 port dumb switch can handle 1024 MACs, and unless it is a very expensive switch the concept of a switch even being concerned about IP addresses is a little off. Consider that current off the shelf infrastructure easily handles things like Cobalt RaQs where you can have dozens to hundreds of IP addresses for the single MAC address. No problems because of it. Routers are only concerned with the MAC an IP came from so they route the packets back the correct path, they don't really care much (for basic protocol handling) about the actual original MAC. MACs don't exist once you transit off ethernet anyway in most cases, so IP over WAN routes to local ports (in normal cases) based on where the traffic came from. There are also alternative protocols of course for load balancing and least cost routing as well.
The easy way is to virtualize all the hardware in a hardware abstraction layer. NT and later (muddied a bit in the graphics area, but that is DirectX at work) already have this on windows it would be easy to target one to a single virtual machine instead of the generic chipsets or in addtion too them. In OS-X86 just do this at the interface to BSD (OSX is BSD over MACH still, right? Not teasing legitimate question!) So in summation, the network stack of ech OS would communicate to virtual virtual hardware and the software implemented virtual hardware talks to the real interface. It is really not a complex deal.
More likely however would be a micro kernal (like MACH) that supports SMP running the OSes as privledged tasks. And ideal would be to sandbox them as much as possible so a blue screen of death would not kill the OS-X86 task. Reinvent some old Apollo (now HP BTW) technology where the display manager allowed emulation of multiple varieties of Unix interfaces and systems simultaneously. The industry already has virtualized the display technology in enough ways to be useful. You do have some unique issues with UI though, as Mac OS has a common top of the screen menu bar and users might be confused when it winks in an out as you select a Mac OS hosted window then switch between that and a Windows window. But solutions exist for that too. Of course you could just write a Windows NT style HAL and run Windows as a task under Mac OSX and be done with it. Conceptually it is not that hard, it just takes time and good design.
While you may like the concept of thinking of multi-core chips as SMP on a chip, in a lot of cases teh designs don't allow the OS to set affinity for a particular core, the hardware on the chip decides which core executes a particular instruction stream. So it is better to use hardware virtualization if available of support it in software. Otherwise you may be tied to a specific chip family, or chip, or maybe even stepping of a particular chip.
iMacs, and iBooks are not allowed to have secondary displays except in mirror mode. Apple's official stance. What does it take to enable the external video port on many models that use a dual display capable chip so that more than just "mirror mode" is possible? Just a quick trip to OpenFirmware and change the settings on a few bits. Could Apple prevent this. Sure, a quick trip to system profiler says I am on an iMac G5 and my serial number is right there. (Heck by Pismo powerbook had my sales order number since it was a built to order machine.) All they'd have to do was check this at boot time and disable the dual display ability. Maybe even a "spank me" by assumming excessive power consuption and heat generation and just turning the fans on "high". Annoying but non lethal. But no, they don't do a thing because they don't like to annoy their customers. And their customers that do these things know it is unsupported. So if I ever do this (wink wink nudge nudge ... know what I mean, know what I mean) I'll be sure to remove it before servicing and hope that open firmware doesn't keep a log somewhere of these "bad" choices a naive user might make.
Remember iMacs and iBooks don't support dual displays, only mirror mode. Riiiigggghhhhttt!!!!
Considering Design Patterns arguably originated with Charles Alexanders work in architecture (for house, communities, etc.) I find the use of a badly built house to describe badly written software appropriate. Without specifically mentioning design patterns the author derides the use of doors that don't behave like normal doors (although I think European outside doors open outward, my Ukraine apartment door did), and other features that don't share the normal characteristics of the items they are supposed to be. Exactly the areas were design patterns solve problems.
Pretty cool commentary to push people toward design patterns if only they would have pushed people toward design patterns in that commentary!
'nuff said folks.