Wait -- I'm not sure if that's the same thing. Radius did sell 24-bit 2D-accelerated video cards for top dollar. But there was nothing specific to Photoshop other than the fact that the QuickDraw bit blasting functions were accelerated. (However the PC versions of these cards were sometimes called "Windows Accelerators" in the old days.) Now all videocards do this.
They also sold "Photoshop accelerators" -- these were co-processor boards which used specially written PS filters to offload the work from the main CPU. (googling finds one of these here -- "4 AT&T 66mhz DSPs")
Non demo-limited shareware is the Mac 'gift economy' as it stands. It's not like they don't have the technical means to time-limit or spray ads in your face or install spyware, and it's well known that only a tiny percentage of shareware users end up paying.
The Windows ecology is far worse. Not only is quality shareware few and far between, it's all spammed and there's hardly any freeware worth using. (Big exception is PowerArchiver and Open Source projects.) I often find myself digging ancient DOS programs out of Simtel just to find a freeware tool that does this or that.
Good story. Most people probably didn't know this part:
Thus, for several years, IBM made money from every PC ever made, whether it had IBM on the label or not.
It's my understanding is that IBM had to licence those patents to you under Reasonable and Non-discriminatory terms (RAND) due to anti-trust restrictions. When the restrictions were lifted, we got closed tech like MCA with much steeper fees.
I'm curious if you negotiated with them, or if they just gave you a pricesheet. (The sums I've heard bandied about were something like $5 per PC. AFAIK, IBM is still getting royalties for VGA controllers and some other bits.)
Does this make it any less of a PC? I think not, but this is just my opinion.
Any box that is wired to require a vendor tax and approval on software is not an open system and not a Personal Computer by definition. Personal Media and Game Terminal maybe.
I thing that it's only a a matter of a year or two before you could BYOB settop game box with decent 3D and TV out using standard parts for close to $300. (YES the PC market is that bad!).
A $300 PC AT-compatible gamebox might not run with the fastest setups, but for some reason I have a feeling that most upcoming PC games will be targetting a lowest common demoninator of... umm.. a GF3 and PIII-733.
Today, MPE has web services, ethernet support, and all the other modern trappings...
A couple years ago, I saw a service order for a 3000. The TCP/IP networking was called "ARPANET PAK" and the NIC being installed was a 10Base2 (thinnet). I guess that's modern enough:)
I think Compaq bought DEC without having the slightest clue about the massive revenues available in the minicomputer space. They probably thought they could engineer a quick death for VMS until they got in and actually took a look at the numbers.
The problem is that the organizational overhead to sustain that revenue is humongus -- loads of hardware and software engineers, thousands of support people, etc. In a recessionary economy, Wall Street starts to turn against such businesses.
And that's what HP is up to. I have no doubt that the 3000 was still profitable, but they are dumping it to get lean-n-mean. Same with the 9000s (hello generic IA-64 boxes) and HPUX as soon as they can figure out how.
My understanding was that there was 2 PC projects going on at IBM -- the cheap off-the-shelf time-to-market one (8088), and the more expensive custom IBM job (68000) [which would have ran a UNIX?].
The histories seem to indicate that time-to-market was the big factor (not that compatibility and supply wasn't also important). I don't know how much more expensive it would have been to go with 68K, but Sun and other people who used it were pretty pricey.
It's very cool that you've seen one of the legends of microcomputer histories. If things had worked out differently, maybe your company would be Slashdot's Evil Enemy instead of Microsoft.
Actually, I think the consumer OEM Windows licence is something like $50 in volume. (Your UID is low enough that you might remember the Windows Refund hullabaloo that netted eMachines owners something like $29.)
It's only a matter of time until decent hardware is cheap enough - I'm just afraid that the OEMs would be scared that MS would cut their air supply for going up against the XBox. There's also the issue of customizing the desktop for TV displays and so on (which would be prohibited by MS's OEM terms).
TI was being nasty relative to their competitors, most of whom had open systems.
But, I suspect it's only a matter of a year or two until you see real $300 "Wintendo" boxes based on the IBM AT architecture that hook up to your TV. That is, unless Microsoft plays some nasty OEM tricks.
Ugg - the non-standard plug problem is the worst thing about SCSI and probably contributed greatly to killing it's adoption. (Next to pricey, fragile external cables) But for consumer (internal) usage, neither problem would have been that bad.
However, for the most part, you can stick any SCSI device on any controller, and given the correct cabling and termination (see above), it will work, even if it slows your chain down.
I dunno, from my point of view as someone who generally sticks with SCSI, it seems that the ATA people have had sheetloads of BIOS incompatibities (it seems like this years drive's are always too big for last year's BIOSes), motherboard incompatibilities (remember hardcoded 'types'? Geometry translation?), and plug-and-play failures (cable select, slave jumpers, IRQ incompatibilies).
Better than MFM stopped being relevant years ago. I figure that the people running the IDE spec just figured that they would eventually be replacd by SCSI, so the never got off their ass to fix the problems and instead barely got around to extend it by a couple bits to support the latest disks. The result is something that's certain less supportable and less installation friendly than SCSI. It's a shame that there was never a forced migration to so that the better tech could get the economies of scale.
The AC's point is that IBM supports Linux because it allows them to sell closed-source, expensive middleware like WebSphere and DB2 to a larger marketscope than the current customer base.
IBM could give a crap about Linux's merits as an OS. It's a marketing platform for the products that make real money for them.
If this middleware stuff become commoditized (eventually), IBM will be a lot less interested in Linux.
Actually, the way I read that is "CIO tried to implement application X on Linux and failed. Meanwhile they shipped application Y on Windows. We have now a shot of getting app X on Windows."
As you point out, 50% of all IT projects end up as clusterfucks. It's pretty certain that Microsoft didn't straighten out that mess in 1 month.
I don't think he's talking about low-end stuff. He's not bitching and moaning about Samba printservers -- he's talking about big UNIX applications that probably have strategic value that are not being migrated to Windows.
And he's practically admitted defeat -- he knows that customers trying rid themselves of expensive midrange stuff by doing a straight migration will find Linux the cheapest bet.* What he's going for is find places where these apps are being upgraded, rewritten, or replaced and make sure that Microsoft has their salesbots in there before the UNIX guys go forward with their solution.
All in all, I doubt he cares about 'legacy' apps. His real worry is that Linux migrations are an opportunity for Java and other cross-platform middleware to come into the picture.
(* Even Microsoft's migration of Hotmail made heavy use of Interix [UNIX] as a compatibility layer instead of rewriting the software to be Windows native.)
Actually the Audio Home Recording act gives you the right to copy CDs for personal use, so long as you are using RIAA-taxed equipment and media. Seen those consumer CD copiers?
Well, as I indicated, I'm pretty conservative, so I'll wait and see. I've heard "These new brand new boards fix all the stability problems" time and time again, and I'd have a little trouble trusting a SiS on facevalue. (Another important factor for me is vidcap hardware compatiblity.)
Cheaper according to who? PriceWatch? If AMD was able to sell out their entire production run to (say) Compaq, I don't think you'd see the price advantage on the commodity market - they'd be just as expensive as Intel. They aren't lowballing their product to DIY folks because they love you.
(Kind of like when PIII-1Ghz were going for something ridiculous like $900 on PW, but you could go and buy a 1Ghz system from Dell for $1500)
My comments are far less FUD oriented than the person I replied to, by your definition.
(Full disclosure: I run an older IBM 2xPIII BX system with no intention to upgrade for a year or two. If I was in the BYOB market today, I'd buy AMD.)
First of all -- I was talking about motherboard chipsets, not CPUs. The CPU has no value until you plug it into something.
Second, Dell is shipping PIIIs for the exact reasons I mentioned -- known stability and standard RAM is more important than performance for their customers. Intel has to 'prove' their new chipsets to this market just as AMD does, and Intel has a much better trackrecord of doing so.
Sure it's fun to think about an Andy Grove/Micheal Dell goatsex conspiracy, but just maaaybeee Dell buys Intel because it's a better value for them and their customers.
(Nice karma whore post with the links! See you at 5:Insightful.)
AMD is leading the market and producing technology that is faster, more reliable, and cheaper.
Everything I've heard about AMD mobos is that they are *less* reliable than the Intel ones. For corporate customers (cf Dell), that's far more important than the 10-20% speedboost that AMD or better-than-SDRAM memory technology gives you.
You have to get out of the gameboy thinking that performance matters uber alles. Any ol' 2001 CPU, even Celerons, are fast enough for the vast majority of users, even those using client-side Java. That puts the "value" somewhere other than the Quake FPS benchmarks over at Biff's Hardware.
Furthermore, AMD might be cheaper on "Pricewatch", but that's not where Dell buys their CPUs. I would suspect that with the whole package (CPU/Mobo/RAM/Discounts/Kickbacks), Intel isn't a whole lot more expensive than AMD for a big OEM. You see those chips cheap on PriceWatch because the vendor has excess stock.
PayPal had a great marketing scheme to start -- they paid you something like $5 to sign up. I believe they also gave a referrer bonus of a dollar or two.
Thus eBay sellers started hawking PayPal in their auction pages. Compared to CyberCash* who tried dotcom-style advertising blitzes, it was a genius marketing plan.
* I don't know if I got the name right here, but there was some Internet money place that bought cheezy softporn-looking ads on half the city buses in San Francisco.
It doesn't look like Kidall committed suicide (http://www.maxframe.com/DR.HTM), although that's apparently a common urban legend.
He owned the market and lost it. Even so, I think the with excessive amount that stupid ol' Novell paid him for DR-DOS was enough to make his heirs happy.
Actually, I think the appeals court threw out the "DLL Mixing" charge, which means (legally) that IE is a 100K executable and the icon and the rest of it is Windows.
Doesn't really matter though (except for the fact that it could affect any Windows app, not just IE) -- they still have to fix the bug.
Yeah, but they still lose more money on an unsold XBox than one that they sold.
If those things are sitting on the shelves gathering dust, MS will drop the price to $200, and that will put them further in the hole than any concerted effort to put them out of business by spending $300.
(Anyone know what the revenue per game is like? I figure $10, which means that MS will need to sell every XBox owner 20 games to break even! Not that they really need to break even. The people I know with N64 and PS units certainly don't have 20 or more games. I know about the razors and blades bit with the console market but this seems a little crazy.)
You can create an LGPL library on top of a GPL library.
Doing that essentially re-licences the LPGL code as GPL (KDE as "modified work as a whole" would include QT, which is GPL and therefore the rest of it is GPL) -- That is unless you use the Operating System exception, and I think most Linux distros ship with Qt now.
Now if you wanted to run a commercial K* app on top of that, on the face of it the licence would prevent you! (Kapital is a derived work of KDE which is a derived work of QT which is GPL.) That is, unless KDE has a Linux-like "API exemption".
Now if you are using the commercial or 'free' Qt, then KDE stays LGPL and you are cool with running Kaptial.
Slashdot Mirror
Wait -- I'm not sure if that's the same thing. Radius did sell 24-bit 2D-accelerated video cards for top dollar. But there was nothing specific to Photoshop other than the fact that the QuickDraw bit blasting functions were accelerated. (However the PC versions of these cards were sometimes called "Windows Accelerators" in the old days.) Now all videocards do this.
They also sold "Photoshop accelerators" -- these were co-processor boards which used specially written PS filters to offload the work from the main CPU. (googling finds one of these here -- "4 AT&T 66mhz DSPs")
Non demo-limited shareware is the Mac 'gift economy' as it stands. It's not like they don't have the technical means to time-limit or spray ads in your face or install spyware, and it's well known that only a tiny percentage of shareware users end up paying.
The Windows ecology is far worse. Not only is quality shareware few and far between, it's all spammed and there's hardly any freeware worth using. (Big exception is PowerArchiver and Open Source projects.) I often find myself digging ancient DOS programs out of Simtel just to find a freeware tool that does this or that.
Good story. Most people probably didn't know this part:
Thus, for several years, IBM made money from every PC ever made, whether it had IBM on the label or not.
It's my understanding is that IBM had to licence those patents to you under Reasonable and Non-discriminatory terms (RAND) due to anti-trust restrictions. When the restrictions were lifted, we got closed tech like MCA with much steeper fees.
I'm curious if you negotiated with them, or if they just gave you a pricesheet. (The sums I've heard bandied about were something like $5 per PC. AFAIK, IBM is still getting royalties for VGA controllers and some other bits.)
Does this make it any less of a PC? I think not, but this is just my opinion.
... umm .. a GF3 and PIII-733.
Any box that is wired to require a vendor tax and approval on software is not an open system and not a Personal Computer by definition. Personal Media and Game Terminal maybe.
I thing that it's only a a matter of a year or two before you could BYOB settop game box with decent 3D and TV out using standard parts for close to $300. (YES the PC market is that bad!).
A $300 PC AT-compatible gamebox might not run with the fastest setups, but for some reason I have a feeling that most upcoming PC games will be targetting a lowest common demoninator of
Today, MPE has web services, ethernet support, and all the other modern trappings...
:)
A couple years ago, I saw a service order for a 3000. The TCP/IP networking was called "ARPANET PAK" and the NIC being installed was a 10Base2 (thinnet). I guess that's modern enough
I think Compaq bought DEC without having the slightest clue about the massive revenues available in the minicomputer space. They probably thought they could engineer a quick death for VMS until they got in and actually took a look at the numbers.
The problem is that the organizational overhead to sustain that revenue is humongus -- loads of hardware and software engineers, thousands of support people, etc. In a recessionary economy, Wall Street starts to turn against such businesses.
And that's what HP is up to. I have no doubt that the 3000 was still profitable, but they are dumping it to get lean-n-mean. Same with the 9000s (hello generic IA-64 boxes) and HPUX as soon as they can figure out how.
My understanding was that there was 2 PC projects going on at IBM -- the cheap off-the-shelf time-to-market one (8088), and the more expensive custom IBM job (68000) [which would have ran a UNIX?].
The histories seem to indicate that time-to-market was the big factor (not that compatibility and supply wasn't also important). I don't know how much more expensive it would have been to go with 68K, but Sun and other people who used it were pretty pricey.
It's very cool that you've seen one of the legends of microcomputer histories. If things had worked out differently, maybe your company would be Slashdot's Evil Enemy instead of Microsoft.
Actually, I think the consumer OEM Windows licence is something like $50 in volume. (Your UID is low enough that you might remember the Windows Refund hullabaloo that netted eMachines owners something like $29.)
It's only a matter of time until decent hardware is cheap enough - I'm just afraid that the OEMs would be scared that MS would cut their air supply for going up against the XBox. There's also the issue of customizing the desktop for TV displays and so on (which would be prohibited by MS's OEM terms).
TI was being nasty relative to their competitors, most of whom had open systems.
But, I suspect it's only a matter of a year or two until you see real $300 "Wintendo" boxes based on the IBM AT architecture that hook up to your TV. That is, unless Microsoft plays some nasty OEM tricks.
Ugg - the non-standard plug problem is the worst thing about SCSI and probably contributed greatly to killing it's adoption. (Next to pricey, fragile external cables) But for consumer (internal) usage, neither problem would have been that bad.
However, for the most part, you can stick any SCSI device on any controller, and given the correct cabling and termination (see above), it will work, even if it slows your chain down.
I dunno, from my point of view as someone who generally sticks with SCSI, it seems that the ATA people have had sheetloads of BIOS incompatibities (it seems like this years drive's are always too big for last year's BIOSes), motherboard incompatibilities (remember hardcoded 'types'? Geometry translation?), and plug-and-play failures (cable select, slave jumpers, IRQ incompatibilies).
Better than MFM stopped being relevant years ago. I figure that the people running the IDE spec just figured that they would eventually be replacd by SCSI, so the never got off their ass to fix the problems and instead barely got around to extend it by a couple bits to support the latest disks. The result is something that's certain less supportable and less installation friendly than SCSI. It's a shame that there was never a forced migration to so that the better tech could get the economies of scale.
The AC's point is that IBM supports Linux because it allows them to sell closed-source, expensive middleware like WebSphere and DB2 to a larger marketscope than the current customer base.
IBM could give a crap about Linux's merits as an OS. It's a marketing platform for the products that make real money for them.
If this middleware stuff become commoditized (eventually), IBM will be a lot less interested in Linux.
Actually, the way I read that is "CIO tried to implement application X on Linux and failed. Meanwhile they shipped application Y on Windows. We have now a shot of getting app X on Windows."
As you point out, 50% of all IT projects end up as clusterfucks. It's pretty certain that Microsoft didn't straighten out that mess in 1 month.
I don't think he's talking about low-end stuff. He's not bitching and moaning about Samba printservers -- he's talking about big UNIX applications that probably have strategic value that are not being migrated to Windows.
And he's practically admitted defeat -- he knows that customers trying rid themselves of expensive midrange stuff by doing a straight migration will find Linux the cheapest bet.* What he's going for is find places where these apps are being upgraded, rewritten, or replaced and make sure that Microsoft has their salesbots in there before the UNIX guys go forward with their solution.
All in all, I doubt he cares about 'legacy' apps. His real worry is that Linux migrations are an opportunity for Java and other cross-platform middleware to come into the picture.
(* Even Microsoft's migration of Hotmail made heavy use of Interix [UNIX] as a compatibility layer instead of rewriting the software to be Windows native.)
Actually the Audio Home Recording act gives you the right to copy CDs for personal use, so long as you are using RIAA-taxed equipment and media. Seen those consumer CD copiers?
Well, as I indicated, I'm pretty conservative, so I'll wait and see. I've heard "These new brand new boards fix all the stability problems" time and time again, and I'd have a little trouble trusting a SiS on facevalue. (Another important factor for me is vidcap hardware compatiblity.)
And since AMD chips are cheaper to start with
Cheaper according to who? PriceWatch? If AMD was able to sell out their entire production run to (say) Compaq, I don't think you'd see the price advantage on the commodity market - they'd be just as expensive as Intel. They aren't lowballing their product to DIY folks because they love you.
(Kind of like when PIII-1Ghz were going for something ridiculous like $900 on PW, but you could go and buy a 1Ghz system from Dell for $1500)
My comments are far less FUD oriented than the person I replied to, by your definition.
(Full disclosure: I run an older IBM 2xPIII BX system with no intention to upgrade for a year or two. If I was in the BYOB market today, I'd buy AMD.)
First of all -- I was talking about motherboard chipsets, not CPUs. The CPU has no value until you plug it into something.
Second, Dell is shipping PIIIs for the exact reasons I mentioned -- known stability and standard RAM is more important than performance for their customers. Intel has to 'prove' their new chipsets to this market just as AMD does, and Intel has a much better trackrecord of doing so.
Sure it's fun to think about an Andy Grove/Micheal Dell goatsex conspiracy, but just maaaybeee Dell buys Intel because it's a better value for them and their customers.
(Nice karma whore post with the links! See you at 5:Insightful.)
AMD is leading the market and producing technology that is faster, more reliable, and cheaper.
Everything I've heard about AMD mobos is that they are *less* reliable than the Intel ones. For corporate customers (cf Dell), that's far more important than the 10-20% speedboost that AMD or better-than-SDRAM memory technology gives you.
You have to get out of the gameboy thinking that performance matters uber alles. Any ol' 2001 CPU, even Celerons, are fast enough for the vast majority of users, even those using client-side Java. That puts the "value" somewhere other than the Quake FPS benchmarks over at Biff's Hardware.
Furthermore, AMD might be cheaper on "Pricewatch", but that's not where Dell buys their CPUs. I would suspect that with the whole package (CPU/Mobo/RAM/Discounts/Kickbacks), Intel isn't a whole lot more expensive than AMD for a big OEM. You see those chips cheap on PriceWatch because the vendor has excess stock.
PayPal had a great marketing scheme to start -- they paid you something like $5 to sign up. I believe they also gave a referrer bonus of a dollar or two.
Thus eBay sellers started hawking PayPal in their auction pages. Compared to CyberCash* who tried dotcom-style advertising blitzes, it was a genius marketing plan.
* I don't know if I got the name right here, but there was some Internet money place that bought cheezy softporn-looking ads on half the city buses in San Francisco.
It doesn't look like Kidall committed suicide (http://www.maxframe.com/DR.HTM), although that's apparently a common urban legend.
He owned the market and lost it. Even so, I think the with excessive amount that stupid ol' Novell paid him for DR-DOS was enough to make his heirs happy.
Actually, I think the appeals court threw out the "DLL Mixing" charge, which means (legally) that IE is a 100K executable and the icon and the rest of it is Windows.
Doesn't really matter though (except for the fact that it could affect any Windows app, not just IE) -- they still have to fix the bug.
Yeah, but they still lose more money on an unsold XBox than one that they sold.
If those things are sitting on the shelves gathering dust, MS will drop the price to $200, and that will put them further in the hole than any concerted effort to put them out of business by spending $300.
(Anyone know what the revenue per game is like? I figure $10, which means that MS will need to sell every XBox owner 20 games to break even! Not that they really need to break even. The people I know with N64 and PS units certainly don't have 20 or more games. I know about the razors and blades bit with the console market but this seems a little crazy.)
You can create an LGPL library on top of a GPL library.
Doing that essentially re-licences the LPGL code as GPL (KDE as "modified work as a whole" would include QT, which is GPL and therefore the rest of it is GPL) -- That is unless you use the Operating System exception, and I think most Linux distros ship with Qt now.
Now if you wanted to run a commercial K* app on top of that, on the face of it the licence would prevent you! (Kapital is a derived work of KDE which is a derived work of QT which is GPL.) That is, unless KDE has a Linux-like "API exemption".
Now if you are using the commercial or 'free' Qt, then KDE stays LGPL and you are cool with running Kaptial.
All very very confusing, eh? It's by design.
Best reference I could find :) Prove me wrong.