Anybody can write bad code. It is particularly easy to write bad C++ code and fall into the common traps, such as lots of hidden object instantiation issues, for example. Besides, I can always put empty loops all over the place and made hand-used calculators faster than a program:)
Funny how "have better software" is always the "logical" following of "free software" to some folks. Any dolt can put out crappy code just as easily (or more easy) than someone can put out good code in the GPL realm. The fact that it is "free" (in any sense of the word) has no bearing on whether or not it is "good".
This permits both the embrace-and-extend and the planned obsolescene (word 5 cant open word 6).
You get slammed for not having backwards compatibility BUT making sure you are backwards compatible from the start is a contributor to "bloat" in that the product keeps growing and growing, including code that may never be used.
Of course, backwards compatibility is probably better to have and just live with accusations of bloat.
Heh... does that mean he can no longer find the email client or one that he can use?:P
I assume that you are slickly saying that he doesn't email you with problems anymore.
Intuitive also means that there is logic behind how something behaves and that you can deduce it based on past experiences and logical layout. For example, a big red X on a button that does not delete, but instead turns everything red is not intuitive.
Theres also the consistency issues: If I right-click on this thing here to get the properties for it so I can meddle with the properties, right-click will work anywhere when I want to meddle with the properties.
No... it's probably the ones who have mortages and families to feed.
I have helped start a business that developed software for Windows, Linux, VxWorks, and a few other OSs most won't have heard of. The Linux development had to leach money from the other efforts to be funded/supported because the vast majority of the people who wanted to use our product wanted to use it for free. Well, free doesn't put food on the table for the developers while you develop a product (no product = no revenue).
Once you get the product out, you then can go into a support model for revenue. The problem was the same, the Linux groups didn't want to pay for support either. In the end, our Linux efforts drew money out of our other efforts to keep it funded/supported and it was hard to justify keeping it going.
Linux programming for us became a contract programming (when there were folks who actually wanted to pay for something) and/or in-house work for companies (same thing). In either case, there was always quite a bit of uncertainty in the Linux world because we really had no idea what to expect from quarter to quarter in this mode. Support contracts were nice but they weren't enough, typically, and not enough to really support further/extended development.
First, monopoly means single source. I can get Linux from many different sources
However, there is only one OS. Software released would be for that OS only, regardless if there was a Dell, HP, or Compaq splash screens on it.
Well, you thought wrong. The movement is about creating good software.
Hmmm... then what is this Free (as in speech) stuff that we see thrown about here so much? Free (as in speech) has nothing to do with 'quality', regardless if you argue the quality or lack thereof with the Free (as in speech) software.
Once it loses domination, it will die
I have no problem with survival of the fittest via capitalism. BTW: Capitalism includes free (as in beer) services and products as long as they aren't single-sources or government provided or required.
The problem with the model is paying programmers to write the software. I would bet that a great deal of the programmers who release code into the OSS world are actually programmers during the day. I guess folks will say that it will create more demand for programmers doing in-house work for companies but most companies like the fact that they do not have (nor do they need) in-house programming staff. This leads to the support model that is widely embraced in the OSS world. The problem with the support model is that it assumes you have already paid the programmers.
Actually, Netscape gave the product away to academic sites and similar organizations but required payment for personal/commercial use. Both options were available to you via their site. You can guess how many people actually paid for the software.
I also remember Netscape having a reported 85%+ share of the browser market at one time.
Actually, the "FUD" that I know means (F)ear, (U)ncertainty, and (D)oubt. Lowering Microsoft's revenue would inject even more FUD because the market leader (for whatever reason) will be seen to be less stable. Loss of stability is unnerving to people, thus generating FUD.
So, we'll be trading one monopoly for another and this is a good thing? Don't give me the BS about different distributions being choices. I thought this whole 'movement' was about choice? Why are you so adamant about killing off the Windows choice?
Branch prediction is sort of handled by the predicates. You execute both instruction streams until you say that one of the streams is invalid and negate that stream through the use of predicates.
One reason why C# and VB are interesting is because the same reason as Java. You download the byte-codes and compile them on the machine you have. Using this method, the differences between the EPIC architectures aren't supposed to be too bad.
As someone else said, the onus is put on the compiler writers to optimize for the processor. Yes, this is 'hard'. The fact that compilers may not be so great kind of reflects on the skillz of the compiler writers;p
(Exposing the implementation details to the compiler is one reason why I think that the whole Itanium concept is a bad idea in the long run.)
Yes, one of the biggest weaknesses of the IA64 processors is that it relies on good compilers. The obvious counter-argument to this is that really good compilers can make it do well.
Eliminating all the logic that detects/forwards data dependencies should help it reach higher clockspeeds, too. If you are one of those type people who claim the Athlon is a superior processor to the P4 simply because of the Athlon's higher IPC, then you should also be loving the Itanium line too. The Itanium is designed to sustain an IPC of as wide as there are execution units (of course, dependent upon the compiler and the parallelism naturally in the code being compiled).
Why do you think Sun started/pushed Java so much? They knew that they were getting behind but they wanted a way for developers to still develop software for their platform, even if it was a simple by-product of software development for another platform. It also makes a good selling point for Java - Write it once and run it anywhere. Loosely paraphrased: Write it for your favorite platform and we can run it, too.
Hmm... never heard of the Amiga being called a 16-bit machine.
The 68000 had 32-bit internal registers, 16-bit external datapath to memory, and 24-bit addresses. You could also make the CPU attach to an 8-bit data bus if you wanted or you could use a 68008 with only had 8 data pins on the package. Regardless of the external data bus width, the 32-bit data/address registers could be manipulated with 32-bit ALU operations (add.l for instance added two 32-bit values together) as well as load 32-bit data into the registers from memory (move.l, for example) which generated two address/data bus transactions (4 on the 68008 or to 8-bit wide memory in general).
The 68010 was a 68000 with a bell or whistle extra but it wasn't until the 68020 when the 68k family had 32 data pins on the package.
The bitedness of a CPU is typically measured by the largest data that can be directly manipulated by the integer ALU. For example, if it can add (at least) two 32-bit values together, then it was considered a 32-bit processor. (MMX can add two sets of four 32-bit values together but it is still 32-bit.) Yes, it is an old argument but that's the side I'm on:)
The Atari XE series was similar to the C64 in that they used one of the family of 6502 processors (not even the same one... C64 uses 6510). The graphics of each were a bit different, player/missle vs. sprites. Sound was a bit different too.
When you emulate, you have to emulate the entire machine, not just the CPU - that means any 'unexpected features' of hardware (look up the 65(C)02 sometime and check out all the 'special' behavior of even different revisions of the CPU for different instructions). Some folks wrote programs to take advantage of the logic holes (places that were don't-cares in the logic of the CPU). You also have to do this for all the other chips and stuff. It's pretty fun to do. But... in the posting mentioned in another post in this thread, they expected it would take an 866MHz P3 to emulate a C64 "perfectly" at the original 1MHz speeds.
One thing that was *really* cool about the XE machines was their memory management unit. IIRC, there was a 16K bank in the 'middle' of the 64K address space that was mapped by a single register in the MMU to memory in a 1M address space (the 130XE had 128K, you could expand it up to 1M and actually use it!). You could store routines/data in various banks of 16K throughout the memory that you had and swap them into the lower 64K with a single instruction.
I had an Atari 1040ST! (The first computer to cost less than $1/Kbyte, btw - it was $999). I loved it. I only switched to the dark side (x86) when I was finally forced to in college because of some of the software we were required to use (PSPICE) and got tired of keeping my next door neighbor up until 3am using his machine for them.
Slashdot Mirror
Anybody can write bad code. It is particularly easy to write bad C++ code and fall into the common traps, such as lots of hidden object instantiation issues, for example. Besides, I can always put empty loops all over the place and made hand-used calculators faster than a program :)
Well... The Inquirer had a blurb about this about a week ago, before anyone heard about any "memo".
Who would be the initiator of such a lawsuit?
They plan to save money, and have better software
Funny how "have better software" is always the "logical" following of "free software" to some folks. Any dolt can put out crappy code just as easily (or more easy) than someone can put out good code in the GPL realm. The fact that it is "free" (in any sense of the word) has no bearing on whether or not it is "good".
Interestingly enough...
This permits both the embrace-and-extend and the planned obsolescene (word 5 cant open word 6).
You get slammed for not having backwards compatibility BUT making sure you are backwards compatible from the start is a contributor to "bloat" in that the product keeps growing and growing, including code that may never be used.
Of course, backwards compatibility is probably better to have and just live with accusations of bloat.
But... the P3 and the P4 don't run at the same clock speeds. Last I checked, the fastest P3 I could buy is around 1.3GHz and the P4 is 3.06 GHz.
Heh... does that mean he can no longer find the email client or one that he can use? :P
I assume that you are slickly saying that he doesn't email you with problems anymore.
Intuitive also means that there is logic behind how something behaves and that you can deduce it based on past experiences and logical layout. For example, a big red X on a button that does not delete, but instead turns everything red is not intuitive.
Theres also the consistency issues: If I right-click on this thing here to get the properties for it so I can meddle with the properties, right-click will work anywhere when I want to meddle with the properties.
No... it's probably the ones who have mortages and families to feed.
I have helped start a business that developed software for Windows, Linux, VxWorks, and a few other OSs most won't have heard of. The Linux development had to leach money from the other efforts to be funded/supported because the vast majority of the people who wanted to use our product wanted to use it for free. Well, free doesn't put food on the table for the developers while you develop a product (no product = no revenue).
Once you get the product out, you then can go into a support model for revenue. The problem was the same, the Linux groups didn't want to pay for support either. In the end, our Linux efforts drew money out of our other efforts to keep it funded/supported and it was hard to justify keeping it going.
Linux programming for us became a contract programming (when there were folks who actually wanted to pay for something) and/or in-house work for companies (same thing). In either case, there was always quite a bit of uncertainty in the Linux world because we really had no idea what to expect from quarter to quarter in this mode. Support contracts were nice but they weren't enough, typically, and not enough to really support further/extended development.
First, monopoly means single source. I can get Linux from many different sources
However, there is only one OS. Software released would be for that OS only, regardless if there was a Dell, HP, or Compaq splash screens on it.
Well, you thought wrong. The movement is about creating good software.
Hmmm... then what is this Free (as in speech) stuff that we see thrown about here so much? Free (as in speech) has nothing to do with 'quality', regardless if you argue the quality or lack thereof with the Free (as in speech) software.
Once it loses domination, it will die
I have no problem with survival of the fittest via capitalism. BTW: Capitalism includes free (as in beer) services and products as long as they aren't single-sources or government provided or required.
The problem with the model is paying programmers to write the software. I would bet that a great deal of the programmers who release code into the OSS world are actually programmers during the day. I guess folks will say that it will create more demand for programmers doing in-house work for companies but most companies like the fact that they do not have (nor do they need) in-house programming staff. This leads to the support model that is widely embraced in the OSS world. The problem with the support model is that it assumes you have already paid the programmers.
Actually, Netscape gave the product away to academic sites and similar organizations but required payment for personal/commercial use. Both options were available to you via their site. You can guess how many people actually paid for the software.
I also remember Netscape having a reported 85%+ share of the browser market at one time.
Actually, the "FUD" that I know means (F)ear, (U)ncertainty, and (D)oubt. Lowering Microsoft's revenue would inject even more FUD because the market leader (for whatever reason) will be seen to be less stable. Loss of stability is unnerving to people, thus generating FUD.
Yeah, heaven forbid that someone would actually want to make money.
...or simply chooses to run Windows for free...
So, we'll be trading one monopoly for another and this is a good thing? Don't give me the BS about different distributions being choices. I thought this whole 'movement' was about choice? Why are you so adamant about killing off the Windows choice?
Branch prediction is sort of handled by the predicates. You execute both instruction streams until you say that one of the streams is invalid and negate that stream through the use of predicates.
;p
One reason why C# and VB are interesting is because the same reason as Java. You download the byte-codes and compile them on the machine you have. Using this method, the differences between the EPIC architectures aren't supposed to be too bad.
As someone else said, the onus is put on the compiler writers to optimize for the processor. Yes, this is 'hard'. The fact that compilers may not be so great kind of reflects on the skillz of the compiler writers
Unless Jack had a brother, the big Tramiel at Atari (and Commodore previously) was "Jack Tramiel".
m ie l.html
Here's an article from an old issue of STart (from 1986).
http://www.atarimagazines.com/startv1n3/JackTra
I loved my old 1040ST. I wonder if it still runs...
Actually, the "ST" in the Atari ST name meant "(S)ixteen/(T)hirty-two". I guess we are both right.
One of the charts I think (the bi-modal one) is SPECfp2k on the Y axis and SPECint2K on the X axis.
http://www.tafkac.org/celebrities/bill.gates/gates _memory.html
(Exposing the implementation details to the compiler is one reason why I think that the whole Itanium concept is a bad idea in the long run.)
Yes, one of the biggest weaknesses of the IA64 processors is that it relies on good compilers. The obvious counter-argument to this is that really good compilers can make it do well.
Eliminating all the logic that detects/forwards data dependencies should help it reach higher clockspeeds, too. If you are one of those type people who claim the Athlon is a superior processor to the P4 simply because of the Athlon's higher IPC, then you should also be loving the Itanium line too. The Itanium is designed to sustain an IPC of as wide as there are execution units (of course, dependent upon the compiler and the parallelism naturally in the code being compiled).
Why do you think Sun started/pushed Java so much? They knew that they were getting behind but they wanted a way for developers to still develop software for their platform, even if it was a simple by-product of software development for another platform. It also makes a good selling point for Java - Write it once and run it anywhere. Loosely paraphrased: Write it for your favorite platform and we can run it, too.
Hmm... never heard of the Amiga being called a 16-bit machine.
:)
The 68000 had 32-bit internal registers, 16-bit external datapath to memory, and 24-bit addresses. You could also make the CPU attach to an 8-bit data bus if you wanted or you could use a 68008 with only had 8 data pins on the package. Regardless of the external data bus width, the 32-bit data/address registers could be manipulated with 32-bit ALU operations (add.l for instance added two 32-bit values together) as well as load 32-bit data into the registers from memory (move.l, for example) which generated two address/data bus transactions (4 on the 68008 or to 8-bit wide memory in general).
The 68010 was a 68000 with a bell or whistle extra but it wasn't until the 68020 when the 68k family had 32 data pins on the package.
The bitedness of a CPU is typically measured by the largest data that can be directly manipulated by the integer ALU. For example, if it can add (at least) two 32-bit values together, then it was considered a 32-bit processor. (MMX can add two sets of four 32-bit values together but it is still 32-bit.) Yes, it is an old argument but that's the side I'm on
The Atari XE series was similar to the C64 in that they used one of the family of 6502 processors (not even the same one... C64 uses 6510). The graphics of each were a bit different, player/missle vs. sprites. Sound was a bit different too.
When you emulate, you have to emulate the entire machine, not just the CPU - that means any 'unexpected features' of hardware (look up the 65(C)02 sometime and check out all the 'special' behavior of even different revisions of the CPU for different instructions). Some folks wrote programs to take advantage of the logic holes (places that were don't-cares in the logic of the CPU). You also have to do this for all the other chips and stuff. It's pretty fun to do. But... in the posting mentioned in another post in this thread, they expected it would take an 866MHz P3 to emulate a C64 "perfectly" at the original 1MHz speeds.
One thing that was *really* cool about the XE machines was their memory management unit. IIRC, there was a 16K bank in the 'middle' of the 64K address space that was mapped by a single register in the MMU to memory in a 1M address space (the 130XE had 128K, you could expand it up to 1M and actually use it!). You could store routines/data in various banks of 16K throughout the memory that you had and swap them into the lower 64K with a single instruction.
I had an Atari 1040ST! (The first computer to cost less than $1/Kbyte, btw - it was $999). I loved it. I only switched to the dark side (x86) when I was finally forced to in college because of some of the software we were required to use (PSPICE) and got tired of keeping my next door neighbor up until 3am using his machine for them.
There are other sources of 64-bit workstations available *now* (as in for over a year). http://www.sun.com
If you really must have Linux, you can do that on one of their boxes.