Because GTKs native API is C. There are language bindings for both toolkits but their native interface is C and C++ respectivly.
1)Yes, there are C bindings for Qt. No, nobody uses them.
2)Yes, there are more language bindings for GTK than Qt.
3) No, Qt isnt C++ specific. Java, Python, XML, Perl and C bindings exist in various stages of completion.
True.. you dont have to pay to develop in GTK+ on windows. I probably went a bit too heavy on the RMS rethoric in my reply, but it was mainly to illustrate a point to Free fanatics (that the proprietary Qt is a sense freeer than the open GTK+. anyway, it's not that important..).
What makes the difference for me (just me, i'm not saying this applies to anyone else) is the beauty of Qt. It's a true dream to program with. That, for me, outweighs the fact that i have to pay if i want to port my programs to Windows one day. (My experiences with GTK+ are admittely rather limited - i've just written a few simple apps with it - just enough to realise that i found it a pain to work with)
BTW, GTK has far more bindings than just C. QT is locked hard into C++.
Minor nitpick - While it's true that GTK+ has more bindings than Qt, it's not true that the *only* choice for Qt is C++. There are wellmaintained and stable Python and Java bindings for Qt (works as well as native c++). C and Perl bindings exist, but i dont belive they're as good as native c++. Once gcc 3.0 comes out, with a stable c++ ABI it'll be easier to add language bindings to c++ libraries.
This doesnt change a thing (legally). Guess what? Any binary data can be represented as a decimal number - and therefore this number is as illegal as the original decss.c (according to US courts atleast) The fact that it is a prime doesnt change anything.
Still wickedly cool though...
Of course, you can have even more fun with numbers: don't tell the RIAA, but PI and e contains all their past, current and future songs aswell as all copyrighted material that has or will ever exist in any format you wish. Guess PI should be next on their hitlist.
Bottom line: TrollTech is acting out of spite, and picking sides. They're on some holy crusade to rid the world of the Redmond Menace, and they don't care how many innocent developers and users of Free software for Windows get caught under the tank treads.
Isnt this *exactly* what the GPL does? The exception in the GPL that allows it to be linked to proprietary OS components was only needed because there wasnt a Free operating system at that time. It seems to me that Qt is more free (in the RMS sense) than GTK+, since it activly encourages Free development on Free platforms while GTK+ can legally be used to develop proprietary apps on nonfree OSes. It would seem that the Qt licenses is closer to the goals of the FSF than GTK+ is. (See Why you shouldnt use the LGPL for your next library ) It seems to me Qt is doing the right thing while GTK is encouraging proprietary platforms (ironic, isnt it?)
Meanwhile, the GTK team is actively encouraging the development of Win32 and BeOS ports. In their eyes, no operating systems are more equal than others.
And that is why they will win.
So small technicalities like Qts superior documentation or easy to use signal model, or a nice interface builder (ok, gtk has this too), and a nice, easy to use, consistant API doesnt matter? Unless you're hellbent on using C, I personally think Qt is a better choice in most cases.
And like you said.. nobody is stopping you from forking Qt and porting it to Win32. Why would you need the support of the Trolls for that? (Remember that Qt is their only product)
And i cant belive i'm replying to the obvious troll above.. the licensing wars *should* have gone away a long time ago.
Since writing napster successors is the current opensource fad of the day, and not wanting to be left out i've started my own project: A combination of a secure instant messenger and a napster like p2p file sharing tool (buzzwords galore! Did i mention it's XML based too?).
Eventually, it'll provide pseudoanonymous filesharing along with strongly encrypted communications. Currently i'm in the prealpha stage (was happily coding along on the file indexer when i decided to reload/.) Codebase is in C++ with clients for linux/unix and windows. The IM part is pretty complete, and working fine. I'm working on the file sharing now.
If anyone reading this would be interested in helping out with this project, please email me! (henrik 'at' abelsson 'dot' com)
Or BlackAdder - that's atleast from an OSS friendly company. It also provides all of the above: Not Free, but well worth buying to sponsor theKompanys OSS work:
http://www.thekompany.com/products/blackadder/ .
*bah*
What's to say that the ETs got eyes? A picture isnt nearly as universal as we (a vision oriented species) think. What if the ETs primary sense is sound- or smell? Or even something more exotic. (also - define picture: i'm assuming you mean something like those who went out on the first probes)
The only really universal language is math. the value of e and pi will always remain constant. (+ math is a nessecity to build a reciver to listen to the signal in the first place). Also - any picute has a lot of cultural background encoded implicitly. Pictures are a bad idea. period.
First you state "Americans like big cars" then "American [sic] is one of the cleanest countries in the world". Cars are one of the major sources of pollution, about 25% of total. Big cars use more fuel, need more material, bigger roads, etc. They pollute more by any standard.
And on a second note.. The US was the country that shot down an international enviromental treaty recently. hell, most USians dont even belive that global warming exists even when most of the world scientists tell them. The US polludes more per capita than any other country. period. And worse than that, it does every little about it. I wouldnt care if you wanted to wreck your own place, except i live on the same planet too. Yes, i've lived in both US and Europe. I found the US *much* *much* worse enviromentally.
The US is a single-issue country (profit über alles) just like most developing contries. Only in long time, stable societies to you see ethical values being a major part of the debate. I'd say europe in general is 50-100 years ahead of the US in the ethical evolution of mankind. Most of europe recognizes that the enviroment, equality and social security is what a society needs to focus on for the next century. Building as many (large) cars as possible no matter what the cost isnt reasonable.
GDP/capita does not a whole story make. And it has very little to do with quality of life. (the US for example has a lot very poor people along with a few extremly rich. Gives a high GDP but not a society that's "free and equal for all")
In UN's Human development index (here)
the US is #3 after Canada and Norway. But when you compare gender equality and conditions for the poorest part of the population the US drops down to #18th place, among the last of the industrialized world. All of scandinavia for example is a much more wellrounded society to live in - one that doesnt focus exclusivly on profit at any price.
It is not even an expecially big task - qt is only some 20000 lines of code, for example.
The qt distribution is 506408 lines of c++. The core library is 351371 lines (Qt 2.2.3). You're off by a factor of 25.
Just to compare: kdelibs + kdebase is 593924 lines of code. (week old cvs cnapshot). Qt is a fairly large library. Good thing it's welldesigned.
Sheez. You clearly do not know what you're talking about.
even if everybody was immortal the birth rate wouldnt decrease to zero.. in my personal utopia humanity spreads across the galaxy (in the best startrek like fashion).:)
warning: a bunch of incoherent philosophical ramblings ahead. Proceed with caution.
The problem with god is that she's not needed and very arbitary (you can adequatly explain most everything without introducing a god. See Occams razor) Introducing the existance of god and an afterlife just because we're afraid of death makes a lot of things a lot more complex. Cloning is one of them. I belive that humanity is completly alone and doomed to freedom (ah, thank you Sartre). Life is a lot harder without a benevolent father taking responsibility and protecting you. But as long as we're doomed to complete freedom, we should do everything we can to advance ourselves (which is mostly done with technology).
Cloning and genetic modification is something we *should* do - to improve the lives of humans. Denouncing afterlife comes with a few consequences.. one is that (human) life is absolutly the most valuble thing there is. There can never be a justification to kill someone else in cold blood. Another one is the realization that we need to have genetic engineering to (in the very long run) make humans immortal. If dying is the end of existance, every effort should be made to abolish death from the world. (i'm obvoiusly not talking about something that'll happen in the next few hundred years. my regret is that i was born too early (but then again, it wouldnt be me.. anyway, that's another discussion:))
Supply of a specific product is infinite (or more accuratly, the cost of producing another copy is very small). I'm not saying that there's an infinite amount of content, just that once it's on the net everybody can copy it without it costing the author anything more (the real problem is finding a fair way of compensating the authors for the initial development cost. That's hard.)
The big problem for microsoft starts when people starts questioning the entire idea of intellectial property. If the public becomes used to software being zero cost some very frightening things happen. (remember what they say in basic econ about supply and demand. When supply is infinite, cost should approach zero). Then what's next? People expecting free music? hah!
It will be the downfall of civilization (or atleast of some very large multinationals) - and free software is bringing the idea of content freedom to everybody. That's why microsoft, RIAA, MPAA and all the others are fighting with everything they have to suppress the *idea* of free content.
KDE has excellent i18n support, probably the best of all unix desktop enviroments. It also provides most of what you need in a desktop enviroment.. I belive KDE has full support for unicode and right-left writing. So if you absolutly must avoid MS, i'd look into KDE. http://www.kde.org/il/hebrew/ is the page devoted into translating KDE into hebrew.
[taken from that site]
Version 2.0 of KDE featured several improvements in the field of Hebrew support in its interface. Among these improvements:
KDE is now based on Unicode. There is no longer a need for special fonts for Hebrew.
The translation of KDE applications is complete. Every part in the interface of KDE applications is translated into Hebrew, including all of the menus, messages, and quick help. All of the applications in the kdebase, kdeadmin, kdegames, kdegraphics, kdemultimedia, kdenetwork, kdepim, kdetoys and kdeutils packages are available translated into Hebrew.
The Konqueror web browser which comes with KDE supports displaying any type of Hebrew on the web, including logical Hebrew. However, all other KDE and KOffice applications do not yet support logical Hebrew.
How are those of us who do understand computer security and could evaluate the security of an e-commerce site supposed to determine the security of the sites we purchase products from?
Just follow these thee easy steps - this works for all present and future e-commerce sites:
Become a kickass hacker. If you are one of "us who do understand computer security" this should be a piece of cake.
Try to hack into the vendors servers. Make sure you don't get caught by law enforcement agencies.
If you succed in the hack, dont buy from that vendor, find another and repeat from step 1. Else, go ahead and buy.
Simple and idiotproof, plus you're doing the.com's a favor by pointing out their lackluster security.
Another cross platform RAD tool for Linux and Windows is Black Adder - It uses Qt and Python. At a third of the price for borland's kylix and using python (i'll take python over Pascal or C++ any day) it should be extremly interesting to see. The beta version is still missing a couple of things (writing python w/o autointenting sucks) but i can't think of any faster enviroment to develop in than Qt and Python. Both are extremly easy to learn, and does mostly what you expect it to do..
While Kylix is very interesting for Delphi users wishing to migrate from legacy OSes i think Black Adder is a better choice for the unix crowd. (It's not OSS tho - but i can understand that theKompany needs to make a living too. They've released tons of Free software, so i don't mind "sponsoring" their Free work with buying other non-Free software)
Not to mention that Black Adder is a much cooler name than Kylix:)
It is not emulating, but rather using actual windows libraries through some sort of abstration layer.
Well.. that's not entirely accurate.
Wine aims to be an second implementation of the Win32 API. This means that the overhead over using microsofts implementation only depends on the quality of the wine code (if the wine people write tighter code than the ms guys, wine wins). However, wine can use native win32 dll files, as a fallback for running unsupported api calls (that of course, isnt the only thing it needs dll support for). That's probably what you're talking about.
But the big thing about wine isn't it's binary compability with Win32, it's source compability. Soon all win32 apps will run nativly on linux, only depending on X and the wine libs after a simple recompile and a few tweaks (hopefully).
And you think SAS will use this in passenger flights if it causes interference with the flight instruments? It's obvious that SAS will make sure the wireless lan will work properly before shipping it. As for cell phones (in some other post).. The main difference between allowing an onboard LAN and disallowing cell phones is that the air lines cant control the phones and make sure it works correctly.
So no, before we see this in production use RF interference wont be a problem.
Email is not the Web. Email is a method for sending text (7-bit text, none of this 8-bit M$ASCII crap even!) between users on two systems.
Email is a method for communicating. C-o-m-m-u-n-i-c-a-t-i-o-n is the keyword, the issue isnt HTML vs ASCII - it's which method that allows me to communicate more efficently.
<rant>
And beliving that email should stay 7-bited is so narrowminded that it's scary - makes me wonder if you've ever seen the outside the US? (Assuming you're an american, because most others nations have a realistic worldview) . Repeat, Lather, Rinse: The US isn't the entire world. English isnt the only language. 99% OF ALL LANGUAGES CAN'T BE EXPRESSED IN 7-bit ASCII.
So why the fuck are you ranting about the superiority of 7bit ascii? There isnt a single advantage to using 7-bit ascii over MIME or (preferbly) Unicode.
The Mac will easily hold its own, and in certain tasks, like in photoshop etc, it is much much faster.
The reason for this is that the G4 has 1 MB L2 Cache, which the Athlons and P3's have reduced in size to push the MHz. Why does this matter?
The L2 Cache has a bandwith of ~10GB/s whereas accessing the main memory is 10 times slower, (PC133 has a bandwith of 1.08GB/s). When you're doing effects in Photoshop, a large L2 cache makes a huge difference, simply because the processor can load 1 mb chunks of the picture into the processor cache and perform the effect on it while the Athlons/PIIIs only have room for a quarter of that. In the very specialised problem that Photoshop is a huge L2 cache matters a lot more than MHz. (Most other apps benefit little from a L2 >256kb)
It would be interesting in seeing a benchmark comparing intel's Xeons (which also has a big L2) and the G4. Also, photoshop optimized for the P4, which thanks to rambus has a high memory bandwith (but small caches) would be interesting.
(As for the other apple "innovations", they're mostly interesting from a design perspective, not technical, so i'll leave them alone:) )
It's probably not that difficult technically, but for marketing and financial reasons i doubt we'll see a Sledgehammer compatible Cruose anytime soon. Transmeta will most likely wait and see how well the platform performs in the market before going out on a limb and embrace the platform. Unless microsoft ports windows and most of it's applications, i doubt AMD will succed. Transmeta probably doesnt have the financial muscle to keep two different platforms in the air at the same time either.
I can't belive how many people are missing the point of this announcement.
AMD isnt' redesigning Sledgehammer to include Transmeta tech. They're using Cruoses as development simulators to get developers to port their code to the Sledgehammer architecture *before* the silicon hits the shelves.
Today, this is NOTHING MORE than a way for AMD to ship fast enough simulators so that ppl can start coding for the Sledgehammer.
In the long term however - it'll allow transmeta to develop Sledgehammer compatible chips - but that's a long way off. (BTW, Is it only me that thinks that they targeted their chips at the mobile market as an afterthought "Oops guys, we can't get this to run fast enough. What to do?" "Hmm.. we'll call it a mobile chip.")
6. Etc. I could go on point by point, but the pattern remains. The author clearly doesn't understand the tradeoffs necessary when designing processors, and looks at one side without considering what it is being traded for.
..
Sure the Pentium 4 doesn't perform great on code not optimized for it. But neither did the 486, the Pentium, or the Pentium Pro. And which would you prefer to have right now, a 250MHz 386, or a 1GHz Pentium III?
That's just his point, which you're completly missing - Intel is trying to force programmers to write optimized code for the P4 while the competitors do everything they can to reduce the work the programmers have to do. Writing optimizing compilers is a very hard task, and almost all code is still compiled with compilers optimizing for the 486 (gcc anyone?).
It's been shown over and over again that forcing coders to do things a certain way doesnt work, because we're lazy. Of course, eventually compiler technology will catch up with the processor architecture, but i estimate compiler tech is atleast 10 years behind the silicon. Therefore sacrificing performance for legacy code is a huge mistake, simply because compilers wont be able to catch up during the P4's lifetime. As i said we've barely got Pentium optimizing compilers *now*. AMD and Transmeta understands this.
Slashdot Mirror
1)Yes, there are C bindings for Qt. No, nobody uses them.
2)Yes, there are more language bindings for GTK than Qt.
3) No, Qt isnt C++ specific. Java, Python, XML, Perl and C bindings exist in various stages of completion.
-henrik
What makes the difference for me (just me, i'm not saying this applies to anyone else) is the beauty of Qt. It's a true dream to program with. That, for me, outweighs the fact that i have to pay if i want to port my programs to Windows one day. (My experiences with GTK+ are admittely rather limited - i've just written a few simple apps with it - just enough to realise that i found it a pain to work with)
BTW, GTK has far more bindings than just C. QT is locked hard into C++.
Minor nitpick - While it's true that GTK+ has more bindings than Qt, it's not true that the *only* choice for Qt is C++. There are wellmaintained and stable Python and Java bindings for Qt (works as well as native c++). C and Perl bindings exist, but i dont belive they're as good as native c++. Once gcc 3.0 comes out, with a stable c++ ABI it'll be easier to add language bindings to c++ libraries.
-henrik
Still wickedly cool though...
Of course, you can have even more fun with numbers: don't tell the RIAA, but PI and e contains all their past, current and future songs aswell as all copyrighted material that has or will ever exist in any format you wish. Guess PI should be next on their hitlist.
-henrik
Isnt this *exactly* what the GPL does? The exception in the GPL that allows it to be linked to proprietary OS components was only needed because there wasnt a Free operating system at that time. It seems to me that Qt is more free (in the RMS sense) than GTK+, since it activly encourages Free development on Free platforms while GTK+ can legally be used to develop proprietary apps on nonfree OSes. It would seem that the Qt licenses is closer to the goals of the FSF than GTK+ is. (See Why you shouldnt use the LGPL for your next library ) It seems to me Qt is doing the right thing while GTK is encouraging proprietary platforms (ironic, isnt it?)
Meanwhile, the GTK team is actively encouraging the development of Win32 and BeOS ports. In their eyes, no operating systems are more equal than others. And that is why they will win.
So small technicalities like Qts superior documentation or easy to use signal model, or a nice interface builder (ok, gtk has this too), and a nice, easy to use, consistant API doesnt matter? Unless you're hellbent on using C, I personally think Qt is a better choice in most cases.
And like you said.. nobody is stopping you from forking Qt and porting it to Win32. Why would you need the support of the Trolls for that? (Remember that Qt is their only product)
And i cant belive i'm replying to the obvious troll above.. the licensing wars *should* have gone away a long time ago.
-henrik
Eventually, it'll provide pseudoanonymous filesharing along with strongly encrypted communications. Currently i'm in the prealpha stage (was happily coding along on the file indexer when i decided to reload /.) Codebase is in C++ with clients for linux/unix and windows. The IM part is pretty complete, and working fine. I'm working on the file sharing now.
If anyone reading this would be interested in helping out with this project, please email me! (henrik 'at' abelsson 'dot' com)
-henrik
No, i dont work for 'em. (Tho i'd like to :)
-henrik
What's to say that the ETs got eyes? A picture isnt nearly as universal as we (a vision oriented species) think. What if the ETs primary sense is sound- or smell? Or even something more exotic. (also - define picture: i'm assuming you mean something like those who went out on the first probes)
The only really universal language is math. the value of e and pi will always remain constant. (+ math is a nessecity to build a reciver to listen to the signal in the first place). Also - any picute has a lot of cultural background encoded implicitly. Pictures are a bad idea. period.
-henrik
And on a second note.. The US was the country that shot down an international enviromental treaty recently. hell, most USians dont even belive that global warming exists even when most of the world scientists tell them. The US polludes more per capita than any other country. period. And worse than that, it does every little about it. I wouldnt care if you wanted to wreck your own place, except i live on the same planet too. Yes, i've lived in both US and Europe. I found the US *much* *much* worse enviromentally.
The US is a single-issue country (profit über alles) just like most developing contries. Only in long time, stable societies to you see ethical values being a major part of the debate. I'd say europe in general is 50-100 years ahead of the US in the ethical evolution of mankind. Most of europe recognizes that the enviroment, equality and social security is what a society needs to focus on for the next century. Building as many (large) cars as possible no matter what the cost isnt reasonable.
And profit isnt the only thing that matters.
-henrik
In UN's Human development index (here) the US is #3 after Canada and Norway. But when you compare gender equality and conditions for the poorest part of the population the US drops down to #18th place, among the last of the industrialized world. All of scandinavia for example is a much more wellrounded society to live in - one that doesnt focus exclusivly on profit at any price.
-henrik
The qt distribution is 506408 lines of c++. The core library is 351371 lines (Qt 2.2.3). You're off by a factor of 25.
Just to compare: kdelibs + kdebase is 593924 lines of code. (week old cvs cnapshot). Qt is a fairly large library. Good thing it's welldesigned.
Sheez. You clearly do not know what you're talking about.
-henrik
not very many would have 10 kids tho..
-henrik
The problem with god is that she's not needed and very arbitary (you can adequatly explain most everything without introducing a god. See Occams razor) Introducing the existance of god and an afterlife just because we're afraid of death makes a lot of things a lot more complex. Cloning is one of them. I belive that humanity is completly alone and doomed to freedom (ah, thank you Sartre). Life is a lot harder without a benevolent father taking responsibility and protecting you. But as long as we're doomed to complete freedom, we should do everything we can to advance ourselves (which is mostly done with technology).
Cloning and genetic modification is something we *should* do - to improve the lives of humans. Denouncing afterlife comes with a few consequences.. one is that (human) life is absolutly the most valuble thing there is. There can never be a justification to kill someone else in cold blood. Another one is the realization that we need to have genetic engineering to (in the very long run) make humans immortal. If dying is the end of existance, every effort should be made to abolish death from the world. (i'm obvoiusly not talking about something that'll happen in the next few hundred years. my regret is that i was born too early (but then again, it wouldnt be me.. anyway, that's another discussion :))
-henrik
-henrik
It will be the downfall of civilization (or atleast of some very large multinationals) - and free software is bringing the idea of content freedom to everybody. That's why microsoft, RIAA, MPAA and all the others are fighting with everything they have to suppress the *idea* of free content.
I predict a long, hard fight..
-henrik
[taken from that site]
Version 2.0 of KDE featured several improvements in the field of Hebrew support in its interface. Among these improvements:
-henrik
Just follow these thee easy steps - this works for all present and future e-commerce sites:
Simple and idiotproof, plus you're doing the .com's a favor by pointing out their lackluster security.
-henrik
While Kylix is very interesting for Delphi users wishing to migrate from legacy OSes i think Black Adder is a better choice for the unix crowd. (It's not OSS tho - but i can understand that theKompany needs to make a living too. They've released tons of Free software, so i don't mind "sponsoring" their Free work with buying other non-Free software)
Not to mention that Black Adder is a much cooler name than Kylix :)
-henrik
Well.. that's not entirely accurate. Wine aims to be an second implementation of the Win32 API. This means that the overhead over using microsofts implementation only depends on the quality of the wine code (if the wine people write tighter code than the ms guys, wine wins). However, wine can use native win32 dll files, as a fallback for running unsupported api calls (that of course, isnt the only thing it needs dll support for). That's probably what you're talking about.
But the big thing about wine isn't it's binary compability with Win32, it's source compability. Soon all win32 apps will run nativly on linux, only depending on X and the wine libs after a simple recompile and a few tweaks (hopefully).
-henrik
So no, before we see this in production use RF interference wont be a problem.
-henrik
Email is a method for communicating. C-o-m-m-u-n-i-c-a-t-i-o-n is the keyword, the issue isnt HTML vs ASCII - it's which method that allows me to communicate more efficently.
<rant>
And beliving that email should stay 7-bited is so narrowminded that it's scary - makes me wonder if you've ever seen the outside the US? (Assuming you're an american, because most others nations have a realistic worldview) . Repeat, Lather, Rinse: The US isn't the entire world. English isnt the only language. 99% OF ALL LANGUAGES CAN'T BE EXPRESSED IN 7-bit ASCII.
So why the fuck are you ranting about the superiority of 7bit ascii? There isnt a single advantage to using 7-bit ascii over MIME or (preferbly) Unicode.
</rant>
-henrik
www.namesys.com
The reason for this is that the G4 has 1 MB L2 Cache, which the Athlons and P3's have reduced in size to push the MHz. Why does this matter?
The L2 Cache has a bandwith of ~10GB/s whereas accessing the main memory is 10 times slower, (PC133 has a bandwith of 1.08GB/s). When you're doing effects in Photoshop, a large L2 cache makes a huge difference, simply because the processor can load 1 mb chunks of the picture into the processor cache and perform the effect on it while the Athlons/PIIIs only have room for a quarter of that. In the very specialised problem that Photoshop is a huge L2 cache matters a lot more than MHz. (Most other apps benefit little from a L2 >256kb)
It would be interesting in seeing a benchmark comparing intel's Xeons (which also has a big L2) and the G4. Also, photoshop optimized for the P4, which thanks to rambus has a high memory bandwith (but small caches) would be interesting.
(As for the other apple "innovations", they're mostly interesting from a design perspective, not technical, so i'll leave them alone :) )
-henrik
-henrik
AMD isnt' redesigning Sledgehammer to include Transmeta tech. They're using Cruoses as development simulators to get developers to port their code to the Sledgehammer architecture *before* the silicon hits the shelves. Today, this is NOTHING MORE than a way for AMD to ship fast enough simulators so that ppl can start coding for the Sledgehammer.
In the long term however - it'll allow transmeta to develop Sledgehammer compatible chips - but that's a long way off. (BTW, Is it only me that thinks that they targeted their chips at the mobile market as an afterthought "Oops guys, we can't get this to run fast enough. What to do?" "Hmm.. we'll call it a mobile chip.")
-henrik
..
Sure the Pentium 4 doesn't perform great on code not optimized for it. But neither did the 486, the Pentium, or the Pentium Pro. And which would you prefer to have right now, a 250MHz 386, or a 1GHz Pentium III?
That's just his point, which you're completly missing - Intel is trying to force programmers to write optimized code for the P4 while the competitors do everything they can to reduce the work the programmers have to do. Writing optimizing compilers is a very hard task, and almost all code is still compiled with compilers optimizing for the 486 (gcc anyone?).
It's been shown over and over again that forcing coders to do things a certain way doesnt work, because we're lazy. Of course, eventually compiler technology will catch up with the processor architecture, but i estimate compiler tech is atleast 10 years behind the silicon. Therefore sacrificing performance for legacy code is a huge mistake, simply because compilers wont be able to catch up during the P4's lifetime. As i said we've barely got Pentium optimizing compilers *now*.
AMD and Transmeta understands this.
-henrik