Having pulled a bunch of Cat5 through my house, I think the most important thing that you can do before the walls go up is to make sure that there's a fairly large plastic conduit/wire run that gets you from floor to floor. If your place is multi-level, that's by far the hardest part to do after the sheetrock is up. In addition to pulling the conduit, normally there would also be a number of nylon pull strings put in so you can pull that fat RG-6 coax later.
A lot of home construction codes are put in place to prevent your house from burning up. A pipe like this is the perfect way to spread fire from floor to floor if it's not done properly, so you want to make sure that that's the case. If you electrician has done data wiring before, you'll probably find that the cat5 that runs between floors is supposed to be "plenum" cabling which has a different sheath (more fire resistant and doesn't let off toxic smoke if it does burn). Yes, of course, plenum cabling costs a lot more. If you put a wire run in like this, I believe that the wires in the pipe are plenum and terminate in punch down blocks at the top and bottom of the tube. You can switch to regular cat5 there.
A big pipe like this running laterally to the demarc point (your telephone box) may be useful too. These days, the cable TV box is in the same area, so it's useful for TV drops too.
If you go to Home Depot, they now sell a big fat orange multi-conductor cable (it's about 1" in diameter) for new construction. There's a couple of cat5 sets in there, a couple of rg6 coax and plastic fiber optic as well. I think it's about $0.80/ft.
I have some rack mount servers and network gear. A lot of this stuff can be incredibly loud (strangely, the loudest thing I have is a Cisco 2924 switch - much louder than a Sun Netra and Compaq Proliant DL380). So you may want to think about identifying a server room. That room will need power, access to the big pipe, and may require additional climate control. You might not want or need it, but again, it's easier to do this now than later.
Finally, building codes may make it challenging or expensive in a home, but you may want to consider having him pull a separate AC circuit into your server room. If you have a lot of computers in there, then a separate feed (perhaps as big as 30A) will be really nice for a large UPS.
Well, ATI's disk recorder came out after TiVO, but it's not like the idea of a hard drive video recorder was TiVO's. Been around "forever" (more than a decade) in pro gear, and has been the standard for video editing for at least 8 years now.
TiVO's "thing" was that they (and ReplayTV) were the first consumer-targeted appliances that had this capability.
1000, 3000, 10000 Google servers... They definitely have a large installation, but these aren't *desktop* installs which was the point of the question.
> Its not that they are from Sony, that is an
> effect of aperature grille monitors which need
> the wires to keep the vertical wires straight.
For whatever the duration of the patent, "aperture grille monitor" == "Sony Trinitron" display. That was at least 17 years, perhaps 20 and only expired maybe 8 years or so ago. I'm sure sony had licensed producers but the bulk of non-shadow mask displays carried the Sony brand.
The original topic wasn't about dropping his software, OS, hardware and spending $$$ for a quiet system - he just wanted to know if spending $45 instead of $10 on a super-quiet power supply or case fan would actually make his machine quieter and whether anybody made anything like this.
I'm not so sure about the "type 3 edges" analysis though... When the images are prepped for online display, they will typically be resized to the display size which has a blurring effect because of interpolation, then resharpened using Unsharp Mask to synthetically recreate/restore detail. I think the artifacts that are getting pointed out here are typical of a slightly overzealous UM, not uncommon at all when the focus of the final image isn't right here.
I am DEFINITELY sure that this stuff hardly matters - the live onscreen graphics may not match these screenshot pixel for pixel, and I don't think this really matters to anybody shopping for an XBox.
The company has entered into an employment arrangement from hiring that probably laid it out to the employee - what they work on without regard to whose "time" it is while employed belongs to the company, who can decide to pass on it.
There's nothing that stops the employee from working on anything they want to, but they knew their IP relationship up front, and shouldn't expect the company to back down. This is particularly in a case like this - by virtue of what the employee does as their job, they probably see the opportunity and techniques to realize a solution.
All that said, if the project is distinct from the employee's explicit job responsbilties, it behooves him to disclose this to his management right now. A moral company (and that's not a ridiculous statement) can recognize that work that could be valuable to them might have been done outside of the bounds of work responsbilities and decide to compensate the employee for that work. Doing so benefits the company in that they can more quickly respond to a customer AND they can keep an employee who's clearly demonstrating that he's tuned into the technical and business environment.
If the company doesn't want to recognize the workers contribution, they you're still better of with disclosure now - the company is not going to let the employee take the work anyway, at least not without a fight that the worker is very unlikely to prevail in.
The worker could try to play dumb on the new project, leave the company and try to approach the client directly , but this is potentially disasterous. If you sold your package to somebody who might have been a company client but you got to them first, nobody would be the wiser. If the company loses the contract of a client that had come to them for specific expertise, and lost the contract to an ex-employee who happened to know their requirements and happened to have been working on such a product in their personal time, there's no limit to the amount of doo-doo that will rain down. And if any of the techniques, knowledge, or god forbid, source code of the employer is discovered in the private work, this whole thing will turn into a very serious matter.
If the employee leaves the company, and finds that it's hard to land "clean" customers with his own package, then he'll probably see why the company had him sign the employee IP agreement in the first place. There's a lot of aspects beyond the technical in providing solutions - sales, marketing, and support are just a small part of it all.
> I beg do differ. What I learn during work or off
> work is my experience, granted, I might have
> learned it at work, but it is I who can use it,
> not someone else.
It's your experience, but generally in the US, the intellectual property rights derived from what you learn belong to your employer, not you. Furthermore, it's up to the company as to whether you or others can apply this IP and the forms that it would be applied in.
Under the most restrictive IP agreements that people are asked to sign when taking jobs, everything you think of while you are employed belongs implicitly to the company. So if you are writing java code and think of the Heimlich Maneuver (the one for choking people), then the company may own rights to it. In practice, I think it's rare for a "real" company to press for ownership in a case where it's far afield like this.
The problem comes when you are writing java server code and come up with some significant invention related to java client-side code. It's not directly linked to your job responsibilities, but exposure to your work code might have been a factor. If you wanted to pursue a patent on this invention, generally you would need to offer it to the company first and could only take the patent in your name if they released rights to you. It's been my experience in the past that most companies will give you that release as well, although this may trigger other problems if you decide to commercialize your invention, as that would be a violation of your promise to work only for your employer.
This is true for authoring books as well (copyright here now, instead of patent). If you work on mail systems and want to write a book about them, you'll probably find your company will want to review what you wrote to make sure nothing proprietary is being divulged, but after that, I'd be suprised if they demand a cut of the book rights (not that that's ever much money!).
The company has to be a little careful about the release. If you invented something while an employee, they are also liable for patent/IP violations. If you worked for IBM and wrote Napster in your own time, then they released it to you and you attracted a firestorm of legal action, you can bet that the parties that want to sue you will blow through that release and try to hit your sugar-daddy employer.
This is why the company asked for the document that started this thread. If you were working on a Napster before you came and claimed it on your inventions list, then you pretty much can keep working on it at home. If the company does a reasonable job protecting their proprietary stuff from getting into the next release, then it's all pretty clean. The company needs to police what you started making while you were at the company, but that's what your normal employment agreement is about.
There are lots of famous examples of home projects that were passed on. Probalby the most famous is the original Apple computer - built as a hobby box by Steve Wozniak, he offered it to his employer, Hewlett-Packard. They didn't want it, gave it to him, and he ran with it.
> If they think I'm going to hand over code that
> I developed for use in their projects without
> compensation to me, regardless of license -
> then forget it.
If you developed code prior to being an employee, you have every right to be compensated for it, and I don't think any reputable company would ask you to donate it if you press the point in the slightest.
That is, in fact, why you are asked to identify prior inventions before you are offered employment. When you identify your IP, you are drawing a clear demarcation of what you worked on before you were an employee.
Once you become an employee, the situation is different. In most cases, anything you work on after becoming an employee belongs to the company. If your personal projects are completely unrelated to your job duties, then you can request that the company release it's rights to that work (usually granted, although there may be restrictions). If your personal project is related to your job work, then you probably won't get that release. If this situation is not acceptable to you, then don't become an employee there.
You might not like it, but think about it. If your job is to work on a mail list manager and you have access to proprietary techniques that your employer has developed in handling mail lists, then it's easy to see how your boss might not appreciate you applying those techniques to a personal project. If your job were to *develop* those proprietary techniques, then you may feel frustrated that you can't apply them freely, but that's part of the gig.
This is particularly messy when you have developed high proficiency in a particular area, either on your own or under somebody else's employ. That area is now the best thing to hire you for, but the one with the most complications.
dasunt's comment about how the typical small logo is changed is dead on.
Each BIOS (typically AMI or Award) knows to look in a bit of the battery-backed up RAM for a compressed image. The size and complexity of the image is usually dependent on the specifics of the BIOS (e.g., fancy PnP capabilities = smaller space for picture).
Each BIOS has it's own format for the compressed image. Typically, there's a integrator's utility floppy that tweaks all the board and BIOS setup. You make your image, copy it on the disk and pick the tool that uploads it. That floppy isn't usually a Windows app. Might use DOS to boot up, but it's usually a text menu app. So, theoretically, you could write a Linux version, but it's a lot easier to just use the app!
Geez, I can understand how this might be an interesting thing to do, but I certainly hope that I'm never in a position where I really see this logo a lot.
Funny thing though - the original iMac is probably the loudest, most annoying desktop I've ever heard. You can't hide the box anywhere since it's in one piece. And it has a terrible cheap fan. My wife has one, and you'd never leave it sleeping because the ungodly fan doesn't shut off (well, maybe it would eventually).
I did some contract work on her iMac while waiting for my new G4 tower. The G4 is much louder than my Compaq Prosignia, but not unreasonable. But the iMac really drove me nuts.
Intuitively, taking off the case seems like a good idea. In fact, for most computers that's worse.
Your computer has hot and cool spots. This is particularly true of those tightly packed expansion cards. Properly designed cases stir up the internal airflow so that the heat doesn't stay concentrated. Otherwise you develop a still pocket of hot air adjacent to the hot chips that makes them even hotter - often to the point that they components can debond leading to premature failure.
Usually, a big manufacturer will monitor operational component temps using a bunch of thermocouples to prove their design. And you'll see localized temps leap skyward when you take the case off and lose the controlled circulation.
This often shows up setups like video processors where there are a number of very large, very hot boards. Without the proper cooling, you see degradation of image quality and performance before failure.
An iMac beats this problem by clever design. Like the original Mac, the heat vent is on the top, so natural convection can cause airflow through the case - in the bottom and out the top. Inside, the processor uses a lot less power and there aren't a bunch of expansion card so it's easier to keep the entire board cooled.
David Fung Intuitively, taking off the case seems like a good idea. In fact, for most computers that's worse.
Your computer has hot and cool spots. This is particularly true of those tightly packed expansion cards. Properly designed cases stir up the internal airflow so that the heat doesn't stay concentrated. Otherwise you develop a still pocket of hot air adjacent to the hot chips that makes them even hotter - often to the point that they components can debond leading to premature failure.
Usually, a big manufacturer will monitor operational component temps using a bunch of thermocouples to prove their design. And you'll see localized temps leap skyward when you take the case off and lose the controlled circulation.
This often shows up setups like video processors where there are a number of very large, very hot boards. Without the proper cooling, you see degradation of image quality and performance before failure.
An iMac beats this problem by clever design. Like the original Mac, the heat vent is on the top, so natural convection can cause airflow through the case - in the bottom and out the top. Inside, the processor uses a lot less power and there aren't a bunch of expansion card so it's easier to keep the entire board cooled.
dfung
Re:This is the wrong question
on
Linux Failover?
·
· Score: 1
The previous responses are appropriate, but I'll add this as well...
Because dual-port NICs generally have a pretty specialized function in an enterprise, there's going to be higher expectations in terms of reliability and uptime. It'll also be expected to have support for enterprise-grade applications which may mean special drivers, special qualification, etc. And finally, there's an expectation that it'll have awesome support.
All that stuff might cost a lot, and definitely is something that the manufacturers don't have any qualms about charging heavily for.
And if they really do it, then I don't think anybody is mad.
There's nothing that stops a company from making a dual port card that is just two $10 enet cards on the same board, but they probably won't be able to deliver on the special drivers, support, etc.
Your problem is a tough one that people do face every day.
Unfortunately, the business reality of venture capital will put you at a serious disadvantage. If you were going to talk to a big company about an invention of this sort, you would probably ask them to sign a non-disclosure form that explicitly covers the areas of your invention that you consider proprietary. Sometimes they will sign, often big companies will not.
Because VCs are seeing multiple pitches per day, day after day, common practice is that they will NEVER sign an NDA. It seems unreasonable, but it's not if you think about it. You may think your idea is totally novel and you're the only one working on it. You're probably seriously wrong in that respect, and signing the NDA could lead to legal problems down the line if the were to partner with somebody else who you thought was infringing on your intellectual property.
All that said, it doesn't mean that you shouldn't make every effort to protect your IP. That means seeking patent and copyright protection as appropriate (this is as little as making sure you type a copyright notice in your source code, labels, etc. all the way up to applying for patents). You would also want to make sure that you have reasonably documented your work to date (what? You're not keeping a developmental notebook?) in the event that you do get ripped off later.
After that, it's pretty much trust. If you don't trust the person you're asking a million bucks for, then you probably shouldn't talk about what you're doing and you probably should stop asking for money. None of these rules are written in concrete - if Jim Clark has an idea and wants to pitch it tommorrow, then he probably can get that NDA signed!
After you get past the initial pitch, you'll probably find that NDA-type protection will be something that you get. Of course, 95% of the pitches never get to the discussion level.
I do have some background in this. I was VP of Engineering at a company that was venture funded and was involved in the later stages of dealing with a board of directors that were mostly VC folks through our IPO. My wife works at an A-tier VC firm and deals with the supplicant list every day. If it's any consolation, her firm treats all the business plans and other materials with an extremely high level of confidence, so you shouldn't really worry too much about your product description falling into your competitor's hands.
At the end of the day, your IP and your ability to execute on it will make the decision. I wouldn't be suprised to hear of people who had their ideas pirated out from under them by private investors, or unethical VCs but if you're really out ahead of the rest of the world on a concept and have the ability to execute, that's better than secrecy.
This is definitely one of those "yeah, but..." answers.
There's no doubt that most cards will be able to support your displays. The hardware part will probably be fairly simple. The problem will be that you will likely require some Software(tm). And we all know what that means...
From a hardware standpoint, you need to examine the sync requirement of the display. They come three ways - separate Horizontal (H) and Vertical (V) syncs on two lines, composite sync where both H and V are on the same line, and sync-on-green where the H and V signals are combined with the green video signal. I'll be pretty obvious when you see the pinout.
For the hobbyist, you can easily cheat. If you have sync-on-green coming out of the card and you need composite sync, then connect the green video line to both the sync input on the monitor and the green in. There are many entertaining permutations of this, but you get the idea. This will compromise the performance somewhat, but generally you can make this work by adjusting some of the magic trimpots in the display (the syncs and video levels come in at different voltages, which are being adjusted by the trimpots).
But the tougher problem will be that the monitor may be designed for a fixed sync timing. That would not be changeable except via drastic measures. A multisync monitor sort of watches the syncs coming in and tries to make sense of them, but many older displays are designed to work at one setup. If this is the case, then you need to tweak the sync timing of the video card, which will be an adventure in software hacking.
Pretty much all video cards are highly configurable in software with regard to sync timing. If you've used a control panel that shifted the raster around of the screen by software, it's simply changing the sync timing by shifting counter values. For lots of reasons, ranging from arcane hardware reasons to FCC compliance, the interface for changing timing is usually not public. But it is widely available to developers (some of whom most certainly are part of this community) finding the technical docs aren't that hard. I haven't made any attempt to look up the generic S3 driver in a Linux distribution, but there's probably a data table there that might do what you want it to do when poked with the appropriate number.
It's challenging finding the right setup parameters. You basically are setting a group of countdown timers to generate the sync waveform, but video systems have very high bandwidth requirements so there are often issues with how you configure the video RAM so that you don't get conflicts when timings overlap.
I guess I should start working on a sync tweaker! Issues like the ram array config, dot pitch of your display etc make it difficult for you to generally crank up an extra 100 pixels of width on the screen, but it's certainly possible.
Slashdot Mirror
Having pulled a bunch of Cat5 through my house, I think the most important thing that you can do before the walls go up is to make sure that there's a fairly large plastic conduit/wire run that gets you from floor to floor. If your place is multi-level, that's by far the hardest part to do after the sheetrock is up. In addition to pulling the conduit, normally there would also be a number of nylon pull strings put in so you can pull that fat RG-6 coax later.
A lot of home construction codes are put in place to prevent your house from burning up. A pipe like this is the perfect way to spread fire from floor to floor if it's not done properly, so you want to make sure that that's the case. If you electrician has done data wiring before, you'll probably find that the cat5 that runs between floors is supposed to be "plenum" cabling which has a different sheath (more fire resistant and doesn't let off toxic smoke if it does burn). Yes, of course, plenum cabling costs a lot more. If you put a wire run in like this, I believe that the wires in the pipe are plenum and terminate in punch down blocks at the top and bottom of the tube. You can switch to regular cat5 there.
A big pipe like this running laterally to the demarc point (your telephone box) may be useful too. These days, the cable TV box is in the same area, so it's useful for TV drops too.
If you go to Home Depot, they now sell a big fat orange multi-conductor cable (it's about 1" in diameter) for new construction. There's a couple of cat5 sets in there, a couple of rg6 coax and plastic fiber optic as well. I think it's about $0.80/ft.
I have some rack mount servers and network gear. A lot of this stuff can be incredibly loud (strangely, the loudest thing I have is a Cisco 2924 switch - much louder than a Sun Netra and Compaq Proliant DL380). So you may want to think about identifying a server room. That room will need power, access to the big pipe, and may require additional climate control. You might not want or need it, but again, it's easier to do this now than later.
Finally, building codes may make it challenging or expensive in a home, but you may want to consider having him pull a separate AC circuit into your server room. If you have a lot of computers in there, then a separate feed (perhaps as big as 30A) will be really nice for a large UPS.
Well, ATI's disk recorder came out after TiVO, but it's not like the idea of a hard drive video recorder was TiVO's. Been around "forever" (more than a decade) in pro gear, and has been the standard for video editing for at least 8 years now.
TiVO's "thing" was that they (and ReplayTV) were the first consumer-targeted appliances that had this capability.
1000, 3000, 10000 Google servers... They definitely have a large installation, but these aren't *desktop* installs which was the point of the question.
I don't work at IBM, but with all the commercial software they ship on Linux it would be odd if they don't win this contest...
:-)
Heck, I do believe I read somewhere that they had 2500 guys running Linux and their cubes were all crammed inside one big 'ol S/390 mainframe.
> Its not that they are from Sony, that is an
> effect of aperature grille monitors which need
> the wires to keep the vertical wires straight.
For whatever the duration of the patent, "aperture grille monitor" == "Sony Trinitron" display. That was at least 17 years, perhaps 20 and only expired maybe 8 years or so ago. I'm sure sony had licensed producers but the bulk of non-shadow mask displays carried the Sony brand.
DF
haroldK is dead on in his post.
The original topic wasn't about dropping his software, OS, hardware and spending $$$ for a quiet system - he just wanted to know if spending $45 instead of $10 on a super-quiet power supply or case fan would actually make his machine quieter and whether anybody made anything like this.
DF
I think the analysis of the lens flare is valid.
I'm not so sure about the "type 3 edges" analysis though... When the images are prepped for online display, they will typically be resized to the display size which has a blurring effect because of interpolation, then resharpened using Unsharp Mask to synthetically recreate/restore detail. I think the artifacts that are getting pointed out here are typical of a slightly overzealous UM, not uncommon at all when the focus of the final image isn't right here.
I am DEFINITELY sure that this stuff hardly matters - the live onscreen graphics may not match these screenshot pixel for pixel, and I don't think this really matters to anybody shopping for an XBox.
DF
Excellent post by russh347.
The company has entered into an employment arrangement from hiring that probably laid it out to the employee - what they work on without regard to whose "time" it is while employed belongs to the company, who can decide to pass on it.
There's nothing that stops the employee from working on anything they want to, but they knew their IP relationship up front, and shouldn't expect the company to back down. This is particularly in a case like this - by virtue of what the employee does as their job, they probably see the opportunity and techniques to realize a solution.
All that said, if the project is distinct from the employee's explicit job responsbilties, it behooves him to disclose this to his management right now. A moral company (and that's not a ridiculous statement) can recognize that work that could be valuable to them might have been done outside of the bounds of work responsbilities and decide to compensate the employee for that work. Doing so benefits the company in that they can more quickly respond to a customer AND they can keep an employee who's clearly demonstrating that he's tuned into the technical and business environment.
If the company doesn't want to recognize the workers contribution, they you're still better of with disclosure now - the company is not going to let the employee take the work anyway, at least not without a fight that the worker is very unlikely to prevail in.
The worker could try to play dumb on the new project, leave the company and try to approach the client directly , but this is potentially disasterous. If you sold your package to somebody who might have been a company client but you got to them first, nobody would be the wiser. If the company loses the contract of a client that had come to them for specific expertise, and lost the contract to an ex-employee who happened to know their requirements and happened to have been working on such a product in their personal time, there's no limit to the amount of doo-doo that will rain down. And if any of the techniques, knowledge, or god forbid, source code of the employer is discovered in the private work, this whole thing will turn into a very serious matter.
If the employee leaves the company, and finds that it's hard to land "clean" customers with his own package, then he'll probably see why the company had him sign the employee IP agreement in the first place. There's a lot of aspects beyond the technical in providing solutions - sales, marketing, and support are just a small part of it all.
> That wasn't very well written at all, was it?
Actually, your response is really well written!
> I beg do differ. What I learn during work or off
> work is my experience, granted, I might have
> learned it at work, but it is I who can use it,
> not someone else.
It's your experience, but generally in the US, the intellectual property rights derived from what you learn belong to your employer, not you. Furthermore, it's up to the company as to whether you or others can apply this IP and the forms that it would be applied in.
Under the most restrictive IP agreements that people are asked to sign when taking jobs, everything you think of while you are employed belongs implicitly to the company. So if you are writing java code and think of the Heimlich Maneuver (the one for choking people), then the company may own rights to it. In practice, I think it's rare for a "real" company to press for ownership in a case where it's far afield like this.
The problem comes when you are writing java server code and come up with some significant invention related to java client-side code. It's not directly linked to your job responsibilities, but exposure to your work code might have been a factor. If you wanted to pursue a patent on this invention, generally you would need to offer it to the company first and could only take the patent in your name if they released rights to you. It's been my experience in the past that most companies will give you that release as well, although this may trigger other problems if you decide to commercialize your invention, as that would be a violation of your promise to work only for your employer.
This is true for authoring books as well (copyright here now, instead of patent). If you work on mail systems and want to write a book about them, you'll probably find your company will want to review what you wrote to make sure nothing proprietary is being divulged, but after that, I'd be suprised if they demand a cut of the book rights (not that that's ever much money!).
The company has to be a little careful about the release. If you invented something while an employee, they are also liable for patent/IP violations. If you worked for IBM and wrote Napster in your own time, then they released it to you and you attracted a firestorm of legal action, you can bet that the parties that want to sue you will blow through that release and try to hit your sugar-daddy employer.
This is why the company asked for the document that started this thread. If you were working on a Napster before you came and claimed it on your inventions list, then you pretty much can keep working on it at home. If the company does a reasonable job protecting their proprietary stuff from getting into the next release, then it's all pretty clean. The company needs to police what you started making while you were at the company, but that's what your normal employment agreement is about.
There are lots of famous examples of home projects that were passed on. Probalby the most famous is the original Apple computer - built as a hobby box by Steve Wozniak, he offered it to his employer, Hewlett-Packard. They didn't want it, gave it to him, and he ran with it.
David Fung
> If they think I'm going to hand over code that
> I developed for use in their projects without
> compensation to me, regardless of license -
> then forget it.
If you developed code prior to being an employee, you have every right to be compensated for it, and I don't think any reputable company would ask you to donate it if you press the point in the slightest.
That is, in fact, why you are asked to identify prior inventions before you are offered employment. When you identify your IP, you are drawing a clear demarcation of what you worked on before you were an employee.
Once you become an employee, the situation is different. In most cases, anything you work on after becoming an employee belongs to the company. If your personal projects are completely unrelated to your job duties, then you can request that the company release it's rights to that work (usually granted, although there may be restrictions). If your personal project is related to your job work, then you probably won't get that release. If this situation is not acceptable to you, then don't become an employee there.
You might not like it, but think about it. If your job is to work on a mail list manager and you have access to proprietary techniques that your employer has developed in handling mail lists, then it's easy to see how your boss might not appreciate you applying those techniques to a personal project. If your job were to *develop* those proprietary techniques, then you may feel frustrated that you can't apply them freely, but that's part of the gig.
This is particularly messy when you have developed high proficiency in a particular area, either on your own or under somebody else's employ. That area is now the best thing to hire you for, but the one with the most complications.
David Fung
dasunt's comment about how the typical small logo is changed is dead on.
Each BIOS (typically AMI or Award) knows to look in a bit of the battery-backed up RAM for a compressed image. The size and complexity of the image is usually dependent on the specifics of the BIOS (e.g., fancy PnP capabilities = smaller space for picture).
Each BIOS has it's own format for the compressed image. Typically, there's a integrator's utility floppy that tweaks all the board and BIOS setup. You make your image, copy it on the disk and pick the tool that uploads it. That floppy isn't usually a Windows app. Might use DOS to boot up, but it's usually a text menu app. So, theoretically, you could write a Linux version, but it's a lot easier to just use the app!
Geez, I can understand how this might be an interesting thing to do, but I certainly hope that I'm never in a position where I really see this logo a lot.
David Fung
The new iMacDV is awesome in it's silence.
Funny thing though - the original iMac is probably the loudest, most annoying desktop I've ever heard. You can't hide the box anywhere since it's in one piece. And it has a terrible cheap fan. My wife has one, and you'd never leave it sleeping because the ungodly fan doesn't shut off (well, maybe it would eventually).
I did some contract work on her iMac while waiting for my new G4 tower. The G4 is much louder than my Compaq Prosignia, but not unreasonable. But the iMac really drove me nuts.
dfung
Your computer has hot and cool spots. This is particularly true of those tightly packed expansion cards. Properly designed cases stir up the internal airflow so that the heat doesn't stay concentrated. Otherwise you develop a still pocket of hot air adjacent to the hot chips that makes them even hotter - often to the point that they components can debond leading to premature failure.
Usually, a big manufacturer will monitor operational component temps using a bunch of thermocouples to prove their design. And you'll see localized temps leap skyward when you take the case off and lose the controlled circulation.
This often shows up setups like video processors where there are a number of very large, very hot boards. Without the proper cooling, you see degradation of image quality and performance before failure.
An iMac beats this problem by clever design. Like the original Mac, the heat vent is on the top, so natural convection can cause airflow through the case - in the bottom and out the top. Inside, the processor uses a lot less power and there aren't a bunch of expansion card so it's easier to keep the entire board cooled.
David Fung Intuitively, taking off the case seems like a good idea. In fact, for most computers that's worse.
Your computer has hot and cool spots. This is particularly true of those tightly packed expansion cards. Properly designed cases stir up the internal airflow so that the heat doesn't stay concentrated. Otherwise you develop a still pocket of hot air adjacent to the hot chips that makes them even hotter - often to the point that they components can debond leading to premature failure.
Usually, a big manufacturer will monitor operational component temps using a bunch of thermocouples to prove their design. And you'll see localized temps leap skyward when you take the case off and lose the controlled circulation.
This often shows up setups like video processors where there are a number of very large, very hot boards. Without the proper cooling, you see degradation of image quality and performance before failure.
An iMac beats this problem by clever design. Like the original Mac, the heat vent is on the top, so natural convection can cause airflow through the case - in the bottom and out the top. Inside, the processor uses a lot less power and there aren't a bunch of expansion card so it's easier to keep the entire board cooled.
dfung
The previous responses are appropriate, but I'll add this as well...
Because dual-port NICs generally have a pretty specialized function in an enterprise, there's going to be higher expectations in terms of reliability and uptime. It'll also be expected to have support for enterprise-grade applications which may mean special drivers, special qualification, etc. And finally, there's an expectation that it'll have awesome support.
All that stuff might cost a lot, and definitely is something that the manufacturers don't have any qualms about charging heavily for.
And if they really do it, then I don't think anybody is mad.
There's nothing that stops a company from making a dual port card that is just two $10 enet cards on the same board, but they probably won't be able to deliver on the special drivers, support, etc.
David Fung
Your problem is a tough one that people do face every day.
Unfortunately, the business reality of venture capital will put you at a serious disadvantage. If you were going to talk to a big company about an invention of this sort, you would probably ask them to sign a non-disclosure form that explicitly covers the areas of your invention that you consider proprietary. Sometimes they will sign, often big companies will not.
Because VCs are seeing multiple pitches per day, day after day, common practice is that they will NEVER sign an NDA. It seems unreasonable, but it's not if you think about it. You may think your idea is totally novel and you're the only one working on it. You're probably seriously wrong in that respect, and signing the NDA could lead to legal problems down the line if the were to partner with somebody else who you thought was infringing on your intellectual property.
All that said, it doesn't mean that you shouldn't make every effort to protect your IP. That means seeking patent and copyright protection as appropriate (this is as little as making sure you type a copyright notice in your source code, labels, etc. all the way up to applying for patents). You would also want to make sure that you have reasonably documented your work to date (what? You're not keeping a developmental notebook?) in the event that you do get ripped off later.
After that, it's pretty much trust. If you don't trust the person you're asking a million bucks for, then you probably shouldn't talk about what you're doing and you probably should stop asking for money. None of these rules are written in concrete - if Jim Clark has an idea and wants to pitch it tommorrow, then he probably can get that NDA signed!
After you get past the initial pitch, you'll probably find that NDA-type protection will be something that you get. Of course, 95% of the pitches never get to the discussion level.
I do have some background in this. I was VP of Engineering at a company that was venture funded and was involved in the later stages of dealing with a board of directors that were mostly VC folks through our IPO. My wife works at an A-tier VC firm and deals with the supplicant list every day. If it's any consolation, her firm treats all the business plans and other materials with an extremely high level of confidence, so you shouldn't really worry too much about your product description falling into your competitor's hands.
At the end of the day, your IP and your ability to execute on it will make the decision. I wouldn't be suprised to hear of people who had their ideas pirated out from under them by private investors, or unethical VCs but if you're really out ahead of the rest of the world on a concept and have the ability to execute, that's better than secrecy.
Good luck!
This is definitely one of those "yeah, but..." answers.
There's no doubt that most cards will be able to support your displays. The hardware part will probably be fairly simple. The problem will be that you will likely require some Software(tm). And we all know what that means...
From a hardware standpoint, you need to examine the sync requirement of the display. They come three ways - separate Horizontal (H) and Vertical (V) syncs on two lines, composite sync where both H and V are on the same line, and sync-on-green where the H and V signals are combined with the green video signal. I'll be pretty obvious when you see the pinout.
For the hobbyist, you can easily cheat. If you have sync-on-green coming out of the card and you need composite sync, then connect the green video line to both the sync input on the monitor and the green in. There are many entertaining permutations of this, but you get the idea. This will compromise the performance somewhat, but generally you can make this work by adjusting some of the magic trimpots in the display (the syncs and video levels come in at different voltages, which are being adjusted by the trimpots).
But the tougher problem will be that the monitor may be designed for a fixed sync timing. That would not be changeable except via drastic measures. A multisync monitor sort of watches the syncs coming in and tries to make sense of them, but many older displays are designed to work at one setup. If this is the case, then you need to tweak the sync timing of the video card, which will be an adventure in software hacking.
Pretty much all video cards are highly configurable in software with regard to sync timing. If you've used a control panel that shifted the raster around of the screen by software, it's simply changing the sync timing by shifting counter values. For lots of reasons, ranging from arcane hardware reasons to FCC compliance, the interface for changing timing is usually not public. But it is widely available to developers (some of whom most certainly are part of this community) finding the technical docs aren't that hard. I haven't made any attempt to look up the generic S3 driver in a Linux distribution, but there's probably a data table there that might do what you want it to do when poked with the appropriate number.
It's challenging finding the right setup parameters. You basically are setting a group of countdown timers to generate the sync waveform, but video systems have very high bandwidth requirements so there are often issues with how you configure the video RAM so that you don't get conflicts when timings overlap.
I guess I should start working on a sync tweaker! Issues like the ram array config, dot pitch of your display etc make it difficult for you to generally crank up an extra 100 pixels of width on the screen, but it's certainly possible.
Enough already!
David Fung