Slashdot Mirror
Legacy-Free PCs
JeffM2001 writes "InformationWeek is running a story by Fred Langa which gives an overview of the ways to create a true-Legacy-free computer. Finally we can have a PC not based on twenty year old technology." Update: 04/07 17:34 GMT by T : Pages past the first one of this article seem just to loop; here's the printable version, which has the whole article in one go.
Down with PS/2! Down with RS232! Down with ECP+EPP! Down with floppy disks! Down with ATA/PI! Down with DB15/Analog!
Let's hear it for flash media formats, DVI, USB, SATA, and Firewire!
I'd prefer that my next motherboard contain only modern I/O ports. I wish that more vendors offered them, but they don't. The ones that do, do so at exorbitant prices.
am I the only one having issues getting to the other pages of the article? For some reason, no matter which page I click I never move off page 1. Tried 2 different browsers, page is b0rked.
ANYWAY, I fail to see why legacy is such a bad thing. Just because it's 20 years old doesn't mean it needs to go away. Using this guy's philosophy, Ethernet is 30 years old, and obviously that's a bottleneck compared to newer technologies like token ring and Turbo Arcnet. UNIX is over 30 years old, and obviously it's a bottleneck compared to the young NT kernel.
Just because the PC's core is 20 years old, I'm not sure why we suddenly need to drop everything and change it.
There is no reasonable defense against an idiot with an agenda
:wq
The term "legacy system" is now used to describe any piece of technology which actually works as opposed to "modern system" which describes things that might work.
Our houses, cars, TVs, ovens, toasters, etc... nearly everything we use on a day-to-day basis... contain "legacy" technology.
Our medical profession uses techniques that are centuries old. Why? Because they work.
Merely because something is old does not mean it's bad. My old external modem still works and is as fast as any USB modem. How am I harmed by using this "legacy" technology for faxing? How is my computer slower?!
There are times when old technology should be replaced by new technology. But, merely because it's old does not mean it's bad. We shouldn't be upgrading simply for the sake of it.
What used to be called "time tested" is now called "legacy". We live in a disposable culture where if it's a couple years old, it's worthless. No wonder our music sucks. It took the Beatles, the Stones, and the Who years before they make their best works.
If someone says he and his monkey have nothing to hide, they almost certainly do.
...is not the keyboard ports or RS-232 or floppy drives or BIOS or any of the other things he mentions in his article.
I want a saner interrupt system. We're still using the same 16 interrupts they introduced with the PC-AT, with a little bit of PnP gloss over them. And most systems seem to have certain IRQs reserved away for their respective devices, so you can't use them-- don't have a floppy drive? Well, it'd be nice to let the PnP stuff use that IRQ for something else, but on many systems, you can't. And in a world where ever processor has a math-coprocessor _built in_, what's the point of reserving IRQ 13 for it? (Yes, the current design of Pentiums and Athlons require it. But _why_?)
Building a completely legacy free PC is pretty unlikely at this juncture, because the underlying architecture simply hasn't changed...
-JDF
I don't know in what vein you meant that, but it is in general its a worthless thought.
A legacy free OS is about as useful as a legacy free automobile. There is this thing called evolution which is how tools, machines, and software develop. Because of evolution you can easily look at a modern tool and compare its lineage to an old tool.
For example just because you can compare a modern laser cutter with a sharp rock some one used a five thousand years ago doesn't mean the new technology it worthless or even the same because it serves the same function.
Linux(the OS based on 30 year old tech) is NOT 30 year old tech. That's a stupid arguement to make. Fundementals don't change and throwing away 30 years of knowledge would be foolish.
Careful with this, folks.
During the last months, whenever news about Palladium or any other DRM system that required hardware support appeared, a common answer was: "so what? As long as we have our legacy motherboards, HDs, etc., we'll be fine. They can't force us to buy new DRM-enabled hardware".
Well, now they can.
Imagine that Microsoft decides that their next version of Windows requires hardware support from this new EFI standard that Intel is pushing. And imagine that EFI carries with it a DRM system.
And what if you are using Linux? Oh yes, it will certainly boot in a new EFI PC. As long as the developers sign a NDA.
Basically, the entertainment industry has an interest in seeing all the PCs obsolete and replaced with DRM-enabled hardware, and this "revolution" is their golden chance. Not that replacing obsolete technology isn't a bad thing, but I'd be very wary of anything "they" try to sell us under the cover of being "innovative, cheaper, efficient, modern"... (have you read the first page of the article? It sounds like a hype piece from Intel itself).
There is nothing wrong with 20 year old technology if there isn't anythig better (or standarized) to replace it.
You are still using the wheel arent you?
16,777,216 comments ought to be enough for any forum!
You're full of it. I've got a Sun Ultra 10 at home that uses OpenProm instead of a BIOS. Works great with not only Solaris, but also Linux, FreeBSD and NetBSD. Having *real* firmware instead of the off the shelf hack that IBM did way back when is not an immediate path to TCPA architectures. Having TCPA hardware (which by itself isn't so bad) is a more realistic path.
Javascript + Nintendo DSi = DSiCade
No version of slackware will install if your usb keyboard and mouse are plugged into a usb hub. Actually, NO version of linux will install this way. Windows2000 is far superior to Linux in USB support, as it installs no problem.
A slashdotter who didn't build his own computer is like a Jedi who didn't build his own lightsaber.
"The installed base--that is, the mass of existing, older, in-use hardware--acts like a giant speed brake on the computer industry because businesses and users are loath to give up older equipment that's still functional, even if newer designs would perform better or faster."
Just like this says, this is about the computer industry - not about the users, the businesses that rely on computers or the businesses that develop software. It's about those who sell new systems.
Hell, what commodity industry wouldn't like to see the current technology stack thrown out the window every 20 years ? The perhaps largest change we see in consumer technology today is the current TV systems being replaced with HDTV. That too is driven by the industry, but has only become possible with the emergence of cheap DVD technologies and crappification of cinema theaters that makes the home experience better than the cinema experience. Consumers now feel that HDTV will give them a meaningful upgrade.
I doubt that very few home users feel that the 20 year old legacy is a problem. In fact, most users realize that there is little need to upgrade the core of the computer any longer, since performance for their basics needs isn't improved with new hardware (gamers excluded).
Oh, I can't help quoting you because everything that you said rings true
just because a technology is old and/or looks similar to how it used to a couple decades ago doesn't make it invalid. That was exactly my point. My company still uses serial port devices. Just because its an old standard, its not invaild. The serial devices we use still do what they where intended to do, why replace them? I don't mean that we shouldn't have forward progress, I just mean to imply that not all standard changes are progress.
OK. Ignore the x86 architecture bit completely, and just look at the add-ons you're addressing - ISA slots, floppy drives, RS232 ports, parallel ports, PS/2 keyboard/mouse adapters.
For instance, RS232 ports: What exactly is wrong with an RS232 port? Why is it "worse" than a USB port? There's no difficulty in actually using an RS232 port - UARTs are cheap, they're brain-dead easy to interface to, and they support rather modern interface methods (DMA, etc.). They are, however, low speed - but of course, for low speed operations that's all you need. You will never need high-speed data transfer to your keyboard or mouse - they're inherently low data transfer devices, since humans are slow.
Same goes for ISA slots and parallel ports. They don't hold back the state of the art. They're add-ons. If you don't use them, they don't do anything. It's just a memory space that doesn't get accessed. If you're complaining about their implementation on current PCs (the fact that they sit in I/O space, take up IRQs, etc.) then you're complaining about the BIOS, not the peripherals. I really didn't see the point of replacing the PS/2 keyboard and mouse. They're just serial devices - they interface via the same method that UARTs, etc. get addressed, which is ridiculously easy to interface to.
There is nothing fundamentally wrong with legacy components. Interfacing to a UART is trivial. Much more trivial than with USB, in fact. There's no reason a "clean" design of a PC couldn't have a serial port, ISA slot, ATA hard drives, and everything else.
Even the x86 architecture thing is 'not that bad'. Take the x87 architecture - everyone complains about the FXCH instruction, because it IS stupid, but on the P3 and Athlon (but not the P4 - one reason the P4's FP sucks) that instruction's 'free' - it takes 0 clock cycles to process. There's some overhead involved with it, but it's not clear to me that the small gain from fixing the overhead loss would offset the large loss of not being compatible with large portions of x87 software. And it's not clear to me that the overhead couldn't be compensated for in some other way, as well.
That having been said it's a lot easier to slap RS232 on a device than it is USB... but that's just a question of time before the USB chips become as cheap/easy as UART's...
When is that going to be? Let's try never.
I designed and built a UART in my first digital design class. The point isn't that cheap==easy, the point is that the protocol is so much simpler for RS232 than it is for USB.
Creating a USB device requires all sorts of stuff to comply with the standard. They don't put anti-lock brakes on wheelbarrows and never will, no matter how cheap they get. It doesn't make any sense.
Mad Software: Rantings on Developing So
I see a deeper issue with this apparent obsession with "legacy-free," and it has NOTHING to do with "holding back the state-of-the-art."
First, consider this; All the peripherals mentioned -- ISA slots (which, admittedly, I wouldn't mind seeing go away), serial ports, parallel ports, keyboard-and-mouse ports -- are all dirt cheap, and dead easy to implement. The technology to do so has been around for decades. It is proven, it's stable, and it's all (as others have pointed out) add-ons. Having add-ons does NOTHING that I can see to inhibit the "evolution" of the core microprocessor and support logic.
UNLESS, that is, you're Microsoft or Hollywood. Consider all the noise in recent years about digital copyrights, copy protection, ad nauseum. Consider the vast array of add-ons Out There that let consumers burn CDs, DVDs, make tape backups, etc., adding to Jack Valenti and Hilary Rosen's ongoing nightmares. Consider further that Microsoft is one of several companies in a partnership that dictates PC hardware standards.
Now, how do you wrest control away from the computer consumer, in a slow and insidious fashion, so they won't even guess what's happening until its too late? In other words, how do you turn those pesky general-purpose PCs into something that will still do everything Joe or Jane SixPack will want it to, but that exerts all kinds of copy controls and limitations when you hook one of those annoying CD or DVD burners to it?
Why, that's easy. Disguise the removal of those annoyingly versatile, general-purpose, and (most importantly) difficult-to-copy-control features like serial, parallel, SCSI, and others as moving towards "legacy-free" systems!
What's more, let's remake the operating system so that add-on peripherals have to be blessed by Microsoft in order to even run with Windows, today and more than ever in the future! Sure! Just let Uncle Steve, Uncle Bill, and the RIAA/MPAA take care of EVERYTHING, and you won't ever have to worry about violating copyrights, or learning ANYthing more about computers than where the "On" switch is, ever again. Trust us, we know what's best for You!
Consider that, in the not-too-distant future, might we see a "PC" that has NO expansion slots? Just Redmond and Hollywood-approved "ports?"
Yes, I probably am letting my paranoid side run rampant again. However, as I said in another post; If the consumer crowd wants to let themselves be led around by the nose, fine. That's their privilege. All I ask is don't force this "Legacy-Free!" crap down the throats of those of us who don't need it, don't want it, and can't possibly make use of it for our applications.
Bruce Lane, KC7GR,
Blue Feather Technologies
What I never understood was this:
-Industry is so keen on getting us to abandon our PS/2 keyboard and Centronics parallel ports.
-In theory, fewer connectors and less space taken on the chipset components should be cheaper
-Yet, the only "legacy-free" parts I see are either in OEM systems (and generally not for individual consumption), or sold to enthusiasts as wow special at a high price.
I can get decent full-legacy Athlon mainboards at USD 50-70. Why should I pay twice as much if not more for a legacy-free board, and actually get LESS?
Aside: if you're freeing all that mainboard space, can't you find something better to do with it than 144 USB ports? The whole point of USB is that you can use hubs and daisy-chaining so one or two ports should be enough.
It's just like a fascist dictatorship, without the punctual rail service!
@ uso
XP/2000/NT have 'emulation' of a DOS command prompt. They are not legacy, but instead are just being 'clever' (some may argue this).
Under your definition there is no OS that isn't legacy. MacOS X run on Unix, all Windows variants have a command prompt of some means......, BeOS...., oS2... Linux, BSD, Solaris.......
If I built a totally new OS with no command prompt, yet someone created a terminal emulation program or DOS command prompt emulation program, and I included it in my OS, would that make my OS legacy. Think about it before answering.