Though the raw bandwidth of FireWire is great, it has greater latency than SATA or PATA or SCSI. It's an external standard, and has support for things like multiple devices per channel and hotplug. This ads overhead to the data transmitted. Firewire is inherently superior to USB for storage, but not to the other internal standards. SCSI and SATA2 also support ordered command queueing, which ads to performance on multi-user systems.
Plus, with drives today, SATA's "mere" 150MB/s will never be saturated.
No, that's a poor metaphor. It's like saying that all your books from one publisher should be attached together. Or that your pen should be bolted onto your notebook. Or your palette should be bolted onto your easel. Our tools we use should not be bolted onto the medium.
Think of it this way... eventually what would be ideal is to make a generic document editing interface... it shows views of the document, and nothing else. Then you open up GIMP to give you tools with which to edit that document. And you should also be able to open up other tools and work on the same document -- equation editors, various exporters, "code palletes" for making it do stuff.
Perhaps my post contained a bit of unnecessary hubris. And yes, HCI is a young discipline (I wouldn't really call it a science per se; it isn't developed enough and relies way too much on trial and error), but many of the common principals are valid. But there are performance metrics to back up my claims about MDI in particular. I'm of the school of thought, as were my professors, that for the sake of progress if a better interface is developed it should be used if and only if it is significantly better than the previous interface.
Your example of QWERTY keyboards is the classic example. Numerous tests have been done, and the data I've seen has lead to the fact that superior standards such as Dvorak only provide a measurable (>3 WPM) benefit with typists that type over 100WPM, which is the vast minority.
The lack of MDI in such an application isn't even a switch; the GIMP was built that way. And it currently has numerous advantages over the MDI model. Thus, it should be kept. I believe it could be polished into being a more learnable interface than Photoshop as well. And learnability is more useful, long-term, than familiarity. The next generation of computer users would benefit from learnability, while they won't from metaphors that are obsolete.
The GIMP uses a different model... rather than the monolithic application model, it uses the document model. Only the stuff you need to manipulate the document at a given time needs to rank high in Z-order.
Now, that doesn't mean there's no room for improvement. I could see something like a checkbox in the config for "raise all tool palletes on document focus", that would raise all the tool palletes when the image being manipulated gains focus. This would be genuinely useful. Or better yet, make this model a standard type of development model in GTK and Qt, and add to compliant desktops the default behavior. This would kick some serious ass.
The MDI model is thouroughly pointless, and overall a poor model. In today's world, particularly in desktop publishing/art/design/etc., people use a plethora of applications at once. I'm a coder, but I work in a design studio, primarily tying Flash apps and other things to an Oracle database. The people here often have Flash, Photoshop, various explorer windows, and a few other applications open at once. Both Flash and Photoshop are MDI. It's such a pain to switch between them; they take up the whole damn screen.
I can do a lot of quick image manipulation (I know lots more about compression, color indexing, etc. than the art guys do) with GIMP, an explorer window, my AIM client, and a terminal open on the screen all at once. It all fits. I don't have to play the ALT-tab game and wait to see what items I'm using.
The only reason that people on Win32 platforms like MDI is because they are used to it. It suck balls. There's NO REASON to restrict the movements of a window to a parent window. None. My specialty in school (I have a BS in Computer Science) was Human-Computer Interaction -- I know these things. It makes a different window model within a window model -- two completely different ways of interacting with windows. It's not only unnecessary, but it slows users down (they need the mental time to process "oh, this is an internal window"), it confuses users new to the system, and offers no benefits, other than allowing a control-freak UI designer to dictate that no one will use any other application in conjunction with theirs. It's a stupid and arbitrary sandbox.
Are you aware of a splitter (KVM switch, two port, preferably cheap) that uses DVI instead of VGA?
Do they exist? Yes. Can you get one cheap? Hell no!
Belkin makes a range of such devices, with USB inputs for keyboard/mouse. The 2-port model MSRP's for $245, and each of the cables (you'll need to buy two) are $80. So that's a whopping $405 + S&H for this solution. You can probably shave $100 off the top by buying from 3rd-party retailer, but that's still a lot of cash.
Please note, I have zero experience with this sort of device. In theory, it should be better than the analog solution, with no attenuation problems.
Across the board, over the last two years, 2D quality has largely gone up, on average. Yes, comparing a cheaply manufactured nVIDIA card (like an older GF2MX) to a really high-end older card (like a Matrox) and you'll get huge difference. Compare a more recent ATI-branded Radeon 9600 or a GeForce FX 5X00 from a good company (Gainward, Leadtek, PNY, etc.) to a Matrox Parhelia, and there's not a whole hell of a lot of difference. Both nVIDIA and ATI have upped the minimum standards the card manufacturers can use for quality of the caps on thier low-pass filters.
Besides, 2D output quality is going to be largely unimportant when we all finally switch to DVI.
Yeah... they're mapping and graphing is a beautiful use of modern information visualization techniques. Honestly, after spending 10 minutes with her, I could see my mom using this kind of thing in a way that previous mail clients stumped her. Maybe I can hack an XUL plugin for Thunderbird that will do this...
Umm... this happenned because of email. This happenned because they got a trojan via Outlook.
Also, as a developer who worked for 6 months at a company without net access... it sucks. No access to online developer resources. We wrote enterprise-level backup software that ran on NT, Linux, UNIX (Solaris, Irix, HP-UX, *BSD, SCO, AIX, and a bunch of others I can't remember), Novell, OS/2, and a host of others I can't remember. Getting info from various online publications was a chore. Whenever we downloaded a file on our lone internet terminal split by 40 developers, we had to burn it to CD to take it to our desks. We NEVER ran Windows Update because we couldn't, and the sysadmin didn't have time to do it during his down time, so we had to live with bugs.
Not giving programmers Internet access is one of the dumbest things a company can do.
Re:Maybe it time to start working on HURD
on
Back To SCO
·
· Score: 1
If they win against Linux, what makes you think BSD isn't next? If BSD falls, there goes MacOS X. They'll probably try to claim ownership of running a POSIX-style OS on a microkernel, crushing HURD next.
I don't think SCO will win, but if they do, either 1) IBM will squish them with their patent portfolio, or 2) they'll go after everyone else. It's stupid to switch your OS just based on the potential threat of one stupid company.
Frankly, I don't know if either the PS2 or GC could handle Doom III. The PS2 doesn't have as much horsepower, and the GC's graphics core is rather fixed-function. Frankly, even on the XBox, the graphics will be diminished, as the XBox's GF3 core is a far cry from the GeForceFX and the Radeon 9700/9800 cores, which are more or less required to have decent performance with all the options on. Doom III has a ridiculous list of features, including stencil shadows, which generally triple the polycount of a given scene.
I personally could care less if ANY of the consoles get Doom III. I'm a big console fan, but some genres (namely FPS's and RTS's) are just better with PC-style input devices. By the same token, I'd probably never buy Soul Calibur or DOA or Zelda for the PC, if they ever came out.
I think the chance of Itanium being used on the desktop is nil. Itanium is too expensive, and isn't that fast, nor does it have the really attractive property of other Intel CPU's: volume. No one's buying Itanium, thus Itanium isn't being manufactured in volume, thus Itanium isn't cheap. And Apple doesn't have the sales figures to bring Itanium's volume up.
Apple would be much better off going with AMD's Hammer, or IBM's upcoming PowerPC 970 chips, or even a P4/Athlon (not likely... I can see Jobs craving 64-bitness). I'd personnally choose Hammer because AMD is going to produce, and probably sell in volumes similar to the Athlon. Although, the 970 looks mighty tasty...
They're probably doing this with RIAA's permission
on
Instant Concert CDs?
·
· Score: 1
Seriously... if a big company like ClearChannel is doing this, the RIAA has been consulted and is involved.
In all honsety, I think this is a great idea, assuming it can be done well. First, it will discourage artists that can't really sing/play instruments but rely on heavy post-production to make them sound good. Second, it's a good service to the fans. Third, if set up right, it could be a reasonable revenue stream for the artists -- and an incentive. Think about it... they're going to want to make every performance fantastic if they want the fans to buy the CD at the end of the concert.
Actually, those hundreds of millions of people with analog TV sets are largely unaware that this is happenning beneath them. They are all going to find out less than one year before it happens. I'm personally going to wait till the last minute, as newer and cooler digital TV's are coming out every 6 months, and they are beginning to approach affordability.
Also, if any of them are cable subscribers, they will be largely unnaffected. It's just those who recieve airwave-broadcast TV that will be affected.
I didn't specify something... the "can take you only a few hours at a time to do" was specifically about small applications, like custom database frontends, simple web sites, and even simple custom POS systems. I'm not trying to make the HCI field look like cake work; it's not. But there are lots of small jobs, and lots of developers that are only looking to avoid huge gaping errors, not minor little inconsistencies that wil slow down the user by 500 milliseconds. I've done simple jobs in less than 10 hours, and I've spent months at a time working 40+ hour weeks on much bigger jobs.
Umm, I think the poster wanted cash to pay for stuff like medical bills, baby clothing, diapers, furniture, etc. Doesn't sound like he has extra cash to spend on a plasma TV.
You could do consulting, but not for development. You could do something that takes lots of time, but not lots of hours, such as HCI (Human-computer Interaction) analasys. Consult for owners of websites and/or small application developers to perform various levels of HCI evaluation, such as analysis and testing. Some clients may want just a detail of obvious (to an HCI designer) problem areas, others will want results obtainable by user testing, and some of them will want designs.
The first suggestion, simple evaluations, can take you only a few hours at a time to do, and then another couple hours writing a report of suggestions to the client. User testing can be a bit more hairy, but the rule of thumb is for each round of user testing, you'll find all the snags with 4 test subjects, and tests should be 30 minutes or less per subject. Then you use about the same amount of time analyzing the results as the above option. The design-work can take many, many hours of designing, prototyping, and testing, but not every client wants immediate results. Some actually feel better if it takes you a couple months; they feel as if you'd been spending all that time musing over it and tweaking it.
I could give you a place to post it on my web server, if you'd like. It's a pretty nifty project.
Question... how do I decode base64?
I may wind up re-implementing the graphics part in SDL so it'll run on pretty much anything... think about it; roaming across a Windows PC, a SPARC, a Linux box, and a PlayStation2.... that would kick some major ass!
Slashdot Mirror
From the abit website you linked to:
* Note: Compatible with SATA controller of Silicon Image on motherboard only!
And that refers to using it on ATAPI devices. It's fine for hard drives on ANY SATA controller, but only on abit mobos with the SI controller.
Sorry... post to wrong thread.
Though the raw bandwidth of FireWire is great, it has greater latency than SATA or PATA or SCSI. It's an external standard, and has support for things like multiple devices per channel and hotplug. This ads overhead to the data transmitted. Firewire is inherently superior to USB for storage, but not to the other internal standards. SCSI and SATA2 also support ordered command queueing, which ads to performance on multi-user systems.
Plus, with drives today, SATA's "mere" 150MB/s will never be saturated.
It'd be even better if it used laptop-sized drives.
2.5" laptop hard drives suck. They're slow, and they have a way higher rate of failure than regular 3.5" hard drives.
From the abit website you linked to:
* Note: Compatible with SATA controller of Silicon Image on motherboard only!
And that refers to using it on ATAPI devices. It's fine for hard drives on ANY SATA controller, but only on abit mobos with the SI controller.
No, that's a poor metaphor. It's like saying that all your books from one publisher should be attached together. Or that your pen should be bolted onto your notebook. Or your palette should be bolted onto your easel. Our tools we use should not be bolted onto the medium.
Think of it this way... eventually what would be ideal is to make a generic document editing interface... it shows views of the document, and nothing else. Then you open up GIMP to give you tools with which to edit that document. And you should also be able to open up other tools and work on the same document -- equation editors, various exporters, "code palletes" for making it do stuff.
Perhaps my post contained a bit of unnecessary hubris. And yes, HCI is a young discipline (I wouldn't really call it a science per se; it isn't developed enough and relies way too much on trial and error), but many of the common principals are valid. But there are performance metrics to back up my claims about MDI in particular. I'm of the school of thought, as were my professors, that for the sake of progress if a better interface is developed it should be used if and only if it is significantly better than the previous interface.
Your example of QWERTY keyboards is the classic example. Numerous tests have been done, and the data I've seen has lead to the fact that superior standards such as Dvorak only provide a measurable (>3 WPM) benefit with typists that type over 100WPM, which is the vast minority.
The lack of MDI in such an application isn't even a switch; the GIMP was built that way. And it currently has numerous advantages over the MDI model. Thus, it should be kept. I believe it could be polished into being a more learnable interface than Photoshop as well. And learnability is more useful, long-term, than familiarity. The next generation of computer users would benefit from learnability, while they won't from metaphors that are obsolete.
The GIMP uses a different model... rather than the monolithic application model, it uses the document model. Only the stuff you need to manipulate the document at a given time needs to rank high in Z-order.
Now, that doesn't mean there's no room for improvement. I could see something like a checkbox in the config for "raise all tool palletes on document focus", that would raise all the tool palletes when the image being manipulated gains focus. This would be genuinely useful. Or better yet, make this model a standard type of development model in GTK and Qt, and add to compliant desktops the default behavior. This would kick some serious ass.
The MDI model is thouroughly pointless, and overall a poor model. In today's world, particularly in desktop publishing/art/design/etc., people use a plethora of applications at once. I'm a coder, but I work in a design studio, primarily tying Flash apps and other things to an Oracle database. The people here often have Flash, Photoshop, various explorer windows, and a few other applications open at once. Both Flash and Photoshop are MDI. It's such a pain to switch between them; they take up the whole damn screen.
I can do a lot of quick image manipulation (I know lots more about compression, color indexing, etc. than the art guys do) with GIMP, an explorer window, my AIM client, and a terminal open on the screen all at once. It all fits. I don't have to play the ALT-tab game and wait to see what items I'm using.
The only reason that people on Win32 platforms like MDI is because they are used to it. It suck balls. There's NO REASON to restrict the movements of a window to a parent window. None. My specialty in school (I have a BS in Computer Science) was Human-Computer Interaction -- I know these things. It makes a different window model within a window model -- two completely different ways of interacting with windows. It's not only unnecessary, but it slows users down (they need the mental time to process "oh, this is an internal window"), it confuses users new to the system, and offers no benefits, other than allowing a control-freak UI designer to dictate that no one will use any other application in conjunction with theirs. It's a stupid and arbitrary sandbox.
Are you aware of a splitter (KVM switch, two port, preferably cheap) that uses DVI instead of VGA?
Do they exist? Yes. Can you get one cheap? Hell no!
Belkin makes a range of such devices, with USB inputs for keyboard/mouse. The 2-port model MSRP's for $245, and each of the cables (you'll need to buy two) are $80. So that's a whopping $405 + S&H for this solution. You can probably shave $100 off the top by buying from 3rd-party retailer, but that's still a lot of cash.
Please note, I have zero experience with this sort of device. In theory, it should be better than the analog solution, with no attenuation problems.
Across the board, over the last two years, 2D quality has largely gone up, on average. Yes, comparing a cheaply manufactured nVIDIA card (like an older GF2MX) to a really high-end older card (like a Matrox) and you'll get huge difference. Compare a more recent ATI-branded Radeon 9600 or a GeForce FX 5X00 from a good company (Gainward, Leadtek, PNY, etc.) to a Matrox Parhelia, and there's not a whole hell of a lot of difference. Both nVIDIA and ATI have upped the minimum standards the card manufacturers can use for quality of the caps on thier low-pass filters.
Besides, 2D output quality is going to be largely unimportant when we all finally switch to DVI.
Yeah... they're mapping and graphing is a beautiful use of modern information visualization techniques. Honestly, after spending 10 minutes with her, I could see my mom using this kind of thing in a way that previous mail clients stumped her. Maybe I can hack an XUL plugin for Thunderbird that will do this...
Umm... this happenned because of email. This happenned because they got a trojan via Outlook.
Also, as a developer who worked for 6 months at a company without net access... it sucks. No access to online developer resources. We wrote enterprise-level backup software that ran on NT, Linux, UNIX (Solaris, Irix, HP-UX, *BSD, SCO, AIX, and a bunch of others I can't remember), Novell, OS/2, and a host of others I can't remember. Getting info from various online publications was a chore. Whenever we downloaded a file on our lone internet terminal split by 40 developers, we had to burn it to CD to take it to our desks. We NEVER ran Windows Update because we couldn't, and the sysadmin didn't have time to do it during his down time, so we had to live with bugs.
Not giving programmers Internet access is one of the dumbest things a company can do.
If they win against Linux, what makes you think BSD isn't next? If BSD falls, there goes MacOS X. They'll probably try to claim ownership of running a POSIX-style OS on a microkernel, crushing HURD next.
I don't think SCO will win, but if they do, either 1) IBM will squish them with their patent portfolio, or 2) they'll go after everyone else. It's stupid to switch your OS just based on the potential threat of one stupid company.
Sorry, I'm racist against elves. Maybe I should play DAOC and start the genocide!
Frankly, I don't know if either the PS2 or GC could handle Doom III. The PS2 doesn't have as much horsepower, and the GC's graphics core is rather fixed-function. Frankly, even on the XBox, the graphics will be diminished, as the XBox's GF3 core is a far cry from the GeForceFX and the Radeon 9700/9800 cores, which are more or less required to have decent performance with all the options on. Doom III has a ridiculous list of features, including stencil shadows, which generally triple the polycount of a given scene.
I personally could care less if ANY of the consoles get Doom III. I'm a big console fan, but some genres (namely FPS's and RTS's) are just better with PC-style input devices. By the same token, I'd probably never buy Soul Calibur or DOA or Zelda for the PC, if they ever came out.
There's different degrees of expensive. A 900mhz Itanium starts at around $3000 -- which alone is more than all but the most expensive Mac's.
I think the chance of Itanium being used on the desktop is nil. Itanium is too expensive, and isn't that fast, nor does it have the really attractive property of other Intel CPU's: volume. No one's buying Itanium, thus Itanium isn't being manufactured in volume, thus Itanium isn't cheap. And Apple doesn't have the sales figures to bring Itanium's volume up.
Apple would be much better off going with AMD's Hammer, or IBM's upcoming PowerPC 970 chips, or even a P4/Athlon (not likely... I can see Jobs craving 64-bitness). I'd personnally choose Hammer because AMD is going to produce, and probably sell in volumes similar to the Athlon. Although, the 970 looks mighty tasty...
Seriously... if a big company like ClearChannel is doing this, the RIAA has been consulted and is involved.
In all honsety, I think this is a great idea, assuming it can be done well. First, it will discourage artists that can't really sing/play instruments but rely on heavy post-production to make them sound good. Second, it's a good service to the fans. Third, if set up right, it could be a reasonable revenue stream for the artists -- and an incentive. Think about it... they're going to want to make every performance fantastic if they want the fans to buy the CD at the end of the concert.
If you wanna pirate, that's your call - but don't call it anything else.
I'll keep that in mind... if I want to commit robbery at sea, I'll make sure to refer to it as piracy.
Actually, those hundreds of millions of people with analog TV sets are largely unaware that this is happenning beneath them. They are all going to find out less than one year before it happens. I'm personally going to wait till the last minute, as newer and cooler digital TV's are coming out every 6 months, and they are beginning to approach affordability.
Also, if any of them are cable subscribers, they will be largely unnaffected. It's just those who recieve airwave-broadcast TV that will be affected.
I didn't specify something... the "can take you only a few hours at a time to do" was specifically about small applications, like custom database frontends, simple web sites, and even simple custom POS systems. I'm not trying to make the HCI field look like cake work; it's not. But there are lots of small jobs, and lots of developers that are only looking to avoid huge gaping errors, not minor little inconsistencies that wil slow down the user by 500 milliseconds. I've done simple jobs in less than 10 hours, and I've spent months at a time working 40+ hour weeks on much bigger jobs.
Umm, I think the poster wanted cash to pay for stuff like medical bills, baby clothing, diapers, furniture, etc. Doesn't sound like he has extra cash to spend on a plasma TV.
You could do consulting, but not for development. You could do something that takes lots of time, but not lots of hours, such as HCI (Human-computer Interaction) analasys. Consult for owners of websites and/or small application developers to perform various levels of HCI evaluation, such as analysis and testing. Some clients may want just a detail of obvious (to an HCI designer) problem areas, others will want results obtainable by user testing, and some of them will want designs.
The first suggestion, simple evaluations, can take you only a few hours at a time to do, and then another couple hours writing a report of suggestions to the client. User testing can be a bit more hairy, but the rule of thumb is for each round of user testing, you'll find all the snags with 4 test subjects, and tests should be 30 minutes or less per subject. Then you use about the same amount of time analyzing the results as the above option. The design-work can take many, many hours of designing, prototyping, and testing, but not every client wants immediate results. Some actually feel better if it takes you a couple months; they feel as if you'd been spending all that time musing over it and tweaking it.
I could give you a place to post it on my web server, if you'd like. It's a pretty nifty project.
Question... how do I decode base64?
I may wind up re-implementing the graphics part in SDL so it'll run on pretty much anything... think about it; roaming across a Windows PC, a SPARC, a Linux box, and a PlayStation2.... that would kick some major ass!