Oh, and on the subject of de-facto standards, IIRC QT is used by Borland in Kylix, and will be used in the next generation of Delphi6/C++ Builder on Windows.
Yep, Borland's CLX framework in Kylix and the just-released Delphi 6 is built on the shoulders of QT.
Another thought, Borland already have JBuilder for the Mac (being coded in Java, it's an easyish port), what odds on a Mac Delphi/Kylix/C++ port coming soon?
JBuilder is "easyish" to get to Mac because the hard part is done by whoever creates the Java VM for the Mac - not the JBuilder guys. They just wait for the VM to show up with sufficient completeness (which usually means a long wait, since JBuilder uses just about every facet of the VM).
A Mac version of Kylix/Delphi/C++ for any Mac OS would require a completely new compiler code generator that targets whatever CPU architecture the Mac is running this week. That's not exactly something that pops up overnight.
-mazor
Re:Possibly for somethings, not all though.
on
Qt for Mac
·
· Score: 1
You don't have to pay full price to get a QT development license for a particular platform. Borland's Kylix and Delphi 6 development products include QT development and deployment licenses for their respective platforms. You only need to ante up for the bigger QT license if you want to deploy your app on platforms other than the platform of the Borland tools you bought.
BTW, Liebniz' calculus is in broader use today then Newton's calculus. Same ideas, but very different notation. Newtonian notation is typically used today in the physics, but the folks up in the math department at your university use Liebniz notation.
--mazor
(certified and pedigreed in Liebniz notation)
Don't need just a feeling - it's documented in the AMD installation docs. In a recent Duron installation, the chip install sheet warned that powering up the chip without a heat sink installed could cause permanent damage to the chip in less than 15 seconds!
There's your answer, right there: Amtrak is only really useful in the NE. (Coincidentally, that's also the only place where it's particularly profitable.) Anywhere else, and the two points you're travelling between are either close enough that driving is more convenient (and probably a bit faster),
You obviously haven't been to San Francisco. The only sane way to get in or out of SFO (from San Francisco or from San Jose) is by CalTrain. The SF pennisula corridor's commuter train system is profitable enough that they're doubling the tracks and electrifying the who thing Real Soon Now to increase the speed and decrease the noise.
And paying the conductor cash for the ticket is commonplace.
When you call MS to get an auth code, your machine spits out a machine key that you read to the MS operator on the phone. They generate an auth code using the machine key you provide. They are cryptographically related. If you give MS a fake machine key, the auth code you get back won't work. If you fake the auth code, it won't match the machine key of your machine. It's kind of like challenge-response handshakes in secure logins, only done manually.
There is a legal term called the spirit of the agreement. A signed contract can be thrown out by the court if a clause in the contract is significantly at odds with the purpose of the contract and the representations and descriptions made of the contract prior to signing.
There's a simple term for the situation where one party in a contract is deliberately trying to cheat the other party by misrepresenting what the contract is about: the word is fraud. Contracts build on fraudulent representations are almost never enforcible by law.
If somebody gets $10,000 worth of equipment for $10 by price-changing the order form, all the company has to do to justify legal action is demonstrate to the court that the purchaser deliberately intended to defraud the company. Editing the html to change the hidden price fields is a pretty clear intent.
The Borland Delphi compiler (at the heart of Kylix) is no slouch at Intel codegen. For the relatively rare occasion where normal codegen isn't good enough (like, specialized pixel blitter code) you can drop into asm within the Delphi(Kylix) source code.
The problem with using low-level asm is that most coders assume assembler is the solution to all performance problems. It's not. As Michael Abrash points out over and over again in his books on code optimization, shifting from high level compiled code (like Delphi or C) to low level machine code buys you incremental improvements whereas refining the algorithm can improvem performance by orders of magnitude.
There is little about a well-designed rendering engine architecture that is tied to a particular OS or hardware platform. If you write your 3D stuff to use OpenGL, for example, it should work reasonably well on Windows and on Linux, if your graphics hardware participates in the OpenGL rendering pipelines.
The problem with OpenGL is that it forces you to work in a particular model of 3D space. Many of the successful 3D games for the PC got their edge by not using traditional approaches to 3D rendering. "Descent", for example, doesn't use 3D space at all, but uses a connected cube space to decide what's visible and what's not. That's what allowed Descent to break free of the "floor map" mentality of Doom/Quake/DukeNukem and basically produce a mini flight sim in a 3DFPS.
POSIX compliance isn't all it's cracked up to be. Linux would do well to focus on things like supporting end user locales and languages that POSIX doesn't support at all.
I mean, cmon! The POSIX "standard" locale charset doesn't even have 8 bits!
--mazor
I was a bit surprised to read about that data swapping, too. Swapping the salary of a 25 year old and a 50 year old in the same census block may have no effect on the congressional district calculations (which is the main reason for the census), but it certainly does change the correlations between age and income within that census block!
"Hey guys, the President wants to show that the country is performing at XYZ level. Swap some stats around to make the numbers come out right."
Nice story. But what does this say about Linux? People keep pointing to miraculous works by a few individuals, such as Torvalds, as a tribute to Linux. It's not a tribute to Linux, it's idol worship.
Stories like this don't build confidence that the Linux development process is fault-tolerant of human failures.
What happens to the Linux OS when Linus isn't around anymore?
It doesn't have to download the entire 500MB game fileset before the game can begin. Just replace loading scenes from the cdrom with loading scenes from the network. Add a background downloader thread that constantly grabs the next scene the player is likely to venture into, and cache that to the hard drive, and you're set for realtime play.
The file transfer bandwidth on the CDRom drives in console boxes is probably not much beyond 300kbyte/sec. That's "only" 3mbit/sec - well within cable broadband capability. Even at a third that rate, you could still run today's console games over the network with full media content.
Simple: "Improved" doesn't sell. "New" gets the attention of the 8 to 18 year old consumer.
Once people get a certain association in their brain between a label and a product, it's very hard to change it - especially for kids. That can be an advantage (name recognition) and a disadvantage (old embarassments).
When you hear "Playstation 2", even if you know nothing about it, you're bound to think that it must be incrementally related to the Playstation 1. Your views of the Playstation 1 will taint your appraisal of the Playstation 2.
On the other hand, when you hear "X Box", even if you've heard something about it, you still have a lot of unanswered questions. It's mysterious. It's unknown. Imagination fills in the gaps between news and rumor to make the unknown exactly what you hope for, and far far better than what you've got. That's what makes "new" so difficult for "improved" to compete with.
As for why companies get into these vaporware hype contests, it's all about mindshare. If you're not being talked about, you're losing market share hand over fist - perhaps not at the sales counter but definitely in the media and consumer imagination. Hype does produce financially successful products. Just ask Microsoft!
As great as this tool might look, how much flexibility will this take away? It might be a pain to write your applications in VI, letting it compile by Kylix (typing:make goes much faster than saving your file, importing it and compiling it).
So type:make already! Just run the Kylix command line compiler on the source files, as you would for gcc to compile c code from within vi, or using make, or using shell scripts. Geeze. The main difference is that Kylix will compile and link the entire application in the time it takes gcc to compile a.c file to.o!
Does it allow for all those lowlevel debugging gdb has?
Yes, it does. The Kylix IDE internal debugger works the same as the Delphi and C++Builder IDE internal debuggers, with comparable capabilities to gdb. The main difference is you don't have to memorize dozens of cryptic gdb instructions to navigate or inspect the state of your program. That, and the CPU disassembly view lets you step into machine code that you don't have source code for (gdb doesn't).
Also: Kylix generates.stabs debug info so you can debug Kylix apps in gdb if you so desire.
Does the user need to download another set of large libraries to use my applications?
Nope. Non-GUI Kylix applications require only a current glibc installed on user's machines. That's all, nothing else. GUI Kylix apps require installation of the QT library, included in the Kylix product. QT is already be installed on systems running KDE.
1. Borland announces that there will be a free (as in beer) Open Edition of Kylix, which can only be used to produce GPL apps. This will produce more software that must be freely distributed as you advocate at FSF.
2. Poster says "Hey, the tool itself isn't free as in speech, but free as in beer! It's not truly free!"
3. Your response says "Yep. All software should be free. It's morally wrong for anyone to require compensation for the use of their intellectual creations."
Was that it? In 2035 words, I might have missed something!
Not theoretically, but actually:
Many of the TeamB members are already Kylix experts. They've been in trial-by-fire training for months, since the Kylix Kick-Start last year.
They announced that an open-source edition of Kylix will follow the commercial release. The shrink-wrapped Open Edition is supposed to cost $99, but it will also be available for free download from the Borland servers. That's the "version designed for hobbyists" that you refer to and the student discount version that someone else referred to.
>> Will they start giving away an open edition of Delphi or C++ builder on Windows?
They already do. They have a free trial edition of C++ Builder 5 available for download.
>>> Will C++ builder have a free, "open edition" for Linux?
Probably, since this all follows the pattern started by JBuilder Foundation more than a year ago.
>>> Surely they don't make you include the CLX libraries if you're writing for the console.
Of course not. At LinuxWorld, they're showing Delphi Kylix producing 15k console apps. The CLX name covers the entire runtime library, but that doesn't mean the GUI stuff has to be linked into non-GUI server apps. As for C++ Kylix, who knows? It's not out yet.
Slashdot Mirror
Yep, Borland's CLX framework in Kylix and the just-released Delphi 6 is built on the shoulders of QT.
JBuilder is "easyish" to get to Mac because the hard part is done by whoever creates the Java VM for the Mac - not the JBuilder guys. They just wait for the VM to show up with sufficient completeness (which usually means a long wait, since JBuilder uses just about every facet of the VM).
A Mac version of Kylix/Delphi/C++ for any Mac OS would require a completely new compiler code generator that targets whatever CPU architecture the Mac is running this week. That's not exactly something that pops up overnight.
-mazor
-mazor
--mazor
(certified and pedigreed in Liebniz notation)
I don't intend to test this claim!
-mazor
You obviously haven't been to San Francisco. The only sane way to get in or out of SFO (from San Francisco or from San Jose) is by CalTrain. The SF pennisula corridor's commuter train system is profitable enough that they're doubling the tracks and electrifying the who thing Real Soon Now to increase the speed and decrease the noise.
And paying the conductor cash for the ticket is commonplace.
-mazor
Amen!
-mazor
-mazor
There's a simple term for the situation where one party in a contract is deliberately trying to cheat the other party by misrepresenting what the contract is about: the word is fraud. Contracts build on fraudulent representations are almost never enforcible by law.
If somebody gets $10,000 worth of equipment for $10 by price-changing the order form, all the company has to do to justify legal action is demonstrate to the court that the purchaser deliberately intended to defraud the company. Editing the html to change the hidden price fields is a pretty clear intent.
-mazor
-mazor
The problem with using low-level asm is that most coders assume assembler is the solution to all performance problems. It's not. As Michael Abrash points out over and over again in his books on code optimization, shifting from high level compiled code (like Delphi or C) to low level machine code buys you incremental improvements whereas refining the algorithm can improvem performance by orders of magnitude.
There is little about a well-designed rendering engine architecture that is tied to a particular OS or hardware platform. If you write your 3D stuff to use OpenGL, for example, it should work reasonably well on Windows and on Linux, if your graphics hardware participates in the OpenGL rendering pipelines.
The problem with OpenGL is that it forces you to work in a particular model of 3D space. Many of the successful 3D games for the PC got their edge by not using traditional approaches to 3D rendering. "Descent", for example, doesn't use 3D space at all, but uses a connected cube space to decide what's visible and what's not. That's what allowed Descent to break free of the "floor map" mentality of Doom/Quake/DukeNukem and basically produce a mini flight sim in a 3DFPS.
-mazor
I mean, cmon! The POSIX "standard" locale charset doesn't even have 8 bits! --mazor
"Hey guys, the President wants to show that the country is performing at XYZ level. Swap some stats around to make the numbers come out right."
Stories like this don't build confidence that the Linux development process is fault-tolerant of human failures.
What happens to the Linux OS when Linus isn't around anymore?
Forget the noise. Ships crossing the Atlantic at 50mph will be about as popular with whales and seals as speedboats are with manatees in Florida.
The file transfer bandwidth on the CDRom drives in console boxes is probably not much beyond 300kbyte/sec. That's "only" 3mbit/sec - well within cable broadband capability. Even at a third that rate, you could still run today's console games over the network with full media content.
Once people get a certain association in their brain between a label and a product, it's very hard to change it - especially for kids. That can be an advantage (name recognition) and a disadvantage (old embarassments).
When you hear "Playstation 2", even if you know nothing about it, you're bound to think that it must be incrementally related to the Playstation 1. Your views of the Playstation 1 will taint your appraisal of the Playstation 2.
On the other hand, when you hear "X Box", even if you've heard something about it, you still have a lot of unanswered questions. It's mysterious. It's unknown. Imagination fills in the gaps between news and rumor to make the unknown exactly what you hope for, and far far better than what you've got. That's what makes "new" so difficult for "improved" to compete with.
As for why companies get into these vaporware hype contests, it's all about mindshare. If you're not being talked about, you're losing market share hand over fist - perhaps not at the sales counter but definitely in the media and consumer imagination. Hype does produce financially successful products. Just ask Microsoft!
It's a prototype, already!
This is an overstatement. QT is only required for GUI apps. Non-GUI Kylix apps have no dependence on QT.
So type :make already! Just run the Kylix command line compiler on the source files, as you would for gcc to compile c code from within vi, or using make, or using shell scripts. Geeze. The main difference is that Kylix will compile and link the entire application in the time it takes gcc to compile a .c file to .o!
Does it allow for all those lowlevel debugging gdb has?
Yes, it does. The Kylix IDE internal debugger works the same as the Delphi and C++Builder IDE internal debuggers, with comparable capabilities to gdb. The main difference is you don't have to memorize dozens of cryptic gdb instructions to navigate or inspect the state of your program. That, and the CPU disassembly view lets you step into machine code that you don't have source code for (gdb doesn't).
Also: Kylix generates .stabs debug info so you can debug Kylix apps in gdb if you so desire.
Does the user need to download another set of large libraries to use my applications?
Nope. Non-GUI Kylix applications require only a current glibc installed on user's machines. That's all, nothing else. GUI Kylix apps require installation of the QT library, included in the Kylix product. QT is already be installed on systems running KDE.
It was $125 million, bonehead.
So, what does your reply have to do with Kylix?
1. Borland announces that there will be a free (as in beer) Open Edition of Kylix, which can only be used to produce GPL apps. This will produce more software that must be freely distributed as you advocate at FSF.
2. Poster says "Hey, the tool itself isn't free as in speech, but free as in beer! It's not truly free!"
3. Your response says "Yep. All software should be free. It's morally wrong for anyone to require compensation for the use of their intellectual creations."
Was that it? In 2035 words, I might have missed something!
Not theoretically, but actually: Many of the TeamB members are already Kylix experts. They've been in trial-by-fire training for months, since the Kylix Kick-Start last year.
They announced that an open-source edition of Kylix will follow the commercial release. The shrink-wrapped Open Edition is supposed to cost $99, but it will also be available for free download from the Borland servers. That's the "version designed for hobbyists" that you refer to and the student discount version that someone else referred to.
>> Will they start giving away an open edition of Delphi or C++ builder on Windows?
They already do. They have a free trial edition of C++ Builder 5 available for download.
>>> Will C++ builder have a free, "open edition" for Linux?
Probably, since this all follows the pattern started by JBuilder Foundation more than a year ago.
>>> Surely they don't make you include the CLX libraries if you're writing for the console.
Of course not. At LinuxWorld, they're showing Delphi Kylix producing 15k console apps. The CLX name covers the entire runtime library, but that doesn't mean the GUI stuff has to be linked into non-GUI server apps. As for C++ Kylix, who knows? It's not out yet.
--mazor