I really wish the guys at Troll-Tech could work out a licensing agreement to make a version of Qt price-competitive with commercial Motif. I mean a Unix-only version exactly like Qt-free, but with a license to let you produce commercial software with it. It's an amazing product, and we could see Borland QtDelphi and C++-builder with Qt-GUI. Mmm... Commercial and free development can, indeed must, exist side-by-side.
BTW: Whoever mentioned that they like Visual C++ should check out KDevelop (www.kdevelop.org). Its interface is basically identical to VC++. --JZ
Well, I was completely down on the Amiga idea until today. Never owned one, never cared (I used video toaster once though, but didn't get very far). But the tech brief is very interesting. For a decade, Intel, MS, and computer manufacturers have been tuning the OS and hardware to go together. So finally we'll get hardware that's tuned for linux and Java. Plus DVD, multimedia, and USB improvements that can be added to the kernel. HOPEFULLY their X-server will be free too, but it's too early to tell (X could use a few more folks working on hardware-accelerated OpenGL). It's especially nice to hear that they're emphasizing developer tools. Anything we can do to speed up those ports and to finally get a platform that performs well with swing. Ok, so the chance is one in a thousand, but if a platorm actually came out that made Swing seem as fast as native graphics, I would buy it in a second. --JZ
As much as I hate Pascal the language, I really think that Delphi offers the most important contribution to Linux. Especially if you can trivially port Windows Delphi apps to Linux! That would be unbelievably great, but also quite possible, as Delphi is reasonably removed from the direct Win32 API. C++ builder will soon be in MFC-land, so that would REALLY have to diverge from the Windows version. JBuilder 3 is coming, but written in Java, and it will be competing against VisualAge from IBM. Delphi is the truly unique product out of this bunch. Oh yeah, and those of you who mentioned using Qt should check out www.kdevelop.org. KDevelop is still experimental, but actually quite good: Visual dialog builder, complete project management, class browser, integrated debugger, etc. And you don't have to use it for or with KDE if you don't want to. --JZ
So, are Amiga and QNX in a very tiny (somewhat pathetic, if you want my opinion) version of the old OS/2 vs. WinNT split? As in, "Sure, let's develop technology jointly. . . Oh wait, nevermind, how about if I just take my half of the tech and bail, leaving you along and pissed off?" How else do we explain the dual announcements of competing versions of the same product???? --JZ
The old OS/2 very much forms the core of the new one. NextStep was sold to Apple. And this article proves that people still see value in the Amiga OS (even if it IS just the name). Only GEM has been opened up. Maybe someday IBM will give up entirely on OS/2 (Wow, I just read THE WORST review of their new Warp Server, I think in Internet Week, but I can't remember) and set it free. Not that I'm saying it's a bad system. It could just be better if it could cross-fertilize code with other systems/programmers out there. --JZ
Not a quantum leap either. They would have been smarter to put that much effort into a vesion of KDE or GNOME for QNX. Either of them could be made to look a helluva lot like that desktop. And since QNX is still proprietary, it wouldn't really hurt them to have a GPLed desktop on top of it. In fact, they could have contributed to a GPLed desktop, and made their theme proprietary / copyrighted, so they'd still have a unique look and feel. Companies need to get more creative. Let's stop reinventing the wheel, damnit!
Word on the street is that version 4.0 of XF86 should finally be kickin' it. Of course I agree with you that the current X, though laudable for its ever-improving hardware support and the dedication of its developers, has become a dinosaur. And that doesn't even mention the memory leaks! (I've found X can swell up to over 30 megs of RAM or more if I don't restart it). In general, though, I would advise anyone who plans to use either GNOME or KDE to buy 32 more megs of RAM than they would for a similar Windows box. At current memory prices (about $40 for that DIMM), that's still cheaper than buying Win 98. --JZ
will be the establishment of the common CORBA mappings. I'd love to see KDE's KOM/Open Parts component model broaden to the point where I can easily write an app in GTK+ or GTK--, throw a KOM wrapper around it, and then use it seamlessly with KDE KOM parts. Imagine... A Gnumeric spreadsheet embedded in KWord! --JZ
PacketStorm was quite clearly in violation of Harvard's acceptable use policies, which, like most universities, actually permit VERY LITTLE. Their correct response should have been to cut off the server from the outside world, allow Ken to get his data, and be done with it. Their real mistake was inviting him onto the network in the first place. A site like this, operated by one person and centering on controversial issues like cracking and information security, is bound to attract unwanted attention. It should NOT be located on the net of a private, conservative university which has a repuation (albeit not a good one among people I know. ..) to uphold, trustees to satisfy, and lawyers to please. --JZ
If you have trouble with Blackdown, I'd suggest taking a look at the IBM free JVM for Linux. IBM's virtual machine is extremely well reviewed as one of the fastest on the market (comparable to HotSpot), but I think they might still be at 1.1.6.
Probably, but treat it as a beta version at best
on
Reiserfs Released
·
· Score: 1
Check out the results for the 80/20 banded file set. A key result is that Reiserfs performs MUCH faster when large numbers of files are grouped in one directory. Copying all the files from one directory under Reiserfs was 12 times faster than the same operation under ext2. Scanning all files (if located in only one directory) was improved by a factor of 6. These results drop to ratios of 2 or 3 to 1 if the files are in many subdirectories. This uses a mixed set of both large an small files (about 50 megs' worth), but emphasizes the very smallest files (12,000 of the files are under 100 bytes). Those really small files provide reiserfs with the big edge. However, its results are still quite good for medium-large files and the author admits that he's done virtually no optimizing for larger files. I suspect that, if Reiserfs catches on, we'll quickly see these optimizations in the next few months. Thus, I'd wait a little while and see how things pan out before putting it into use.
Ok, so we've all noted that peformance is awesome for 100 byte files. But check out his section on larger files too. The benchmarks of files all greater than 512k are also quite impressive: Look at the rm -rf * comparison-- Ext2: 0:02.63 Reiserfs: 0:00.16 Ratio:16 Wow, now a new user logged in as root can wipe out his/her system 16 times faster than ever! Seriously, though, a.16 second rm -rf * time for 80+ megs in a series of subdirectories is pretty damn impressive. --JZ
Ok, so we've all noted that peformance is awesome for 100 byte files. But check out his section on larger files too. The benchmarks of files all greater than 512k are also quite impressive: Look at the rm -rf * comparison-- Ext2: 0:02.63 Reiserfs: 0:00.16 Ratio:16
Re:An man...this is gonna be way too confusing!
on
Reiserfs Released
·
· Score: 1
No, it's much WORSE with B-trees than binary trees. "Damnit, did I remember to split the/usr/src node on the downward traversal last time? Or should I be splitting when I back up out of the recursion? I don't even know any more. .."
Go out to all your local bookstores, computer stores, etc., and find and destroy every copy of RH5.2, Caldera 1.3, and SuSE5.2 wherever they might be discovered by newbies. Seriously, I've actually been in stores that were selling SuSE5.2 and SuSE6.1 side-by-side, and the 5.2 COST MORE!! Same with Caldera. Plus, they have this "Linux Starter Kit", which is supposed to appeal to newbies, but only includes Caldera 1.3. These people (looking to play around with a desktio) will be sorely disappointed to discover FVWM95, an ancient version of the gimp, and maybe StarOffice 4.0. Don't get me wrong, a solid install of an older linux can run great as a server, but they just present outdated installs and desktops, perpetuating the stereotype that Linux is too hard to use. --JZ
Yes, my SuSE 6.1 is fantastic. Great install (that also installed the 4front sound drivers and saved me a helluva long time), 5 CDs, and they even throw in an old vesion of KOffice (do you know how hard it is to compile KOffice yourself?!?). Also much, much, much cheaper than RH ( I got it for $25-ish at Chumbo, well worth it).
Well, I'm happy for the Be guys. I wonder how much Microwerkz pays for each copy of the OS + Gobe. More to the point, it would be really cool if we could get linux into this kind of box (yes, I know there's a small set-top company using it, but there are a million set tops that run everything, we'll see which one wins). Linux has three key problems in this department. The first is boot time. Anybody know of projects to build an insta-boot linux for embedded devices? The second, the lack of a journalled file system, has been beaten to death. Consumers just want to be able to flick it on and flick it off. We'll see how ext3 does, but I wouldn't be surprised if it took a splinter filesystem to work with small consumer devices. They have such different needs from enterprise servers that they shouldn't necessarily have the same features. The final kicker is XF86. Great software in many ways, and everything I've heard about 4.0 sounds incredible, but it's not small. Anybody have an update on some of the attempts to create a micro-sized graphics system (GGI and Berlin? I can't really remember). . . --JZ
Try KDevelop. Still alpha, but it does integrate Kdbg for debugging. Also has a dialog editor, class browser, syntax highlighting, project management (all this in version 0.4!).
Re:Java IDEs, at least, are poor in GUI design
on
Linux IDE from Cygnus
·
· Score: 1
Well, I suppose I'd have to agree with you about the state of Java GUI design. And, luckily, Java is a lot easier when it comes to writing GUIs by hand, especially in terms of the variety of layout managers (though all of them are still fairly annoying). I found some use for JBuilder, but its layout facilities were not up to par. I'm actually a big fan of VC++'s visual layout. Fast, easy, excellent for adding member variables and functions to go with components. And, as with JavaBeans, ActiveX controls can be added visually even if the IDE has never seen them before. In terms of code resuse, I certainly agree that VB doesn't promote it. Too many properties, not enough methods. It's just not OOP.
There are three valid uses for an ide: -Class browser: See a tree with all the classes/members in your project and just hit one click to take you exactly to a method definition. Or view a graphical inheritance tree.
-Syntax completion: If you're just doing straight Unix C systems programming, no GUIs, and you have plenty of experience, sure you only need to look up the occasional command. But if you're programming against a massive GUI API (Windows, KDE, etc.) with a class library that you didn't write, syntax completion is key. Who the hell knows all three overloaded forms of some obscure command to manipulate a tree control? Who the hell really wants to? An IDE will pop up a parameter list for you instantly.
-GUI design: Look, creating dialogs is not going to be an intellectually stimulating programming task. So do it with a visual editor in 5 minutes, rather than tweaking around manually for hours. Then automatically generate the message handlers for the dialog and get started with real programming.
Good IDEs can make a serious productivity difference if they're used for the right tasks and if you KNOW HOW TO USE THEM. There is very much a mindset to each IDE, just as there is to Emacs, vi, etc. If you approach Visual C++ by thinking of it as just an even bigger, slower version of emacs with different keybindings, it will do you know good. But if you think in its visual terms, you'll get a lot more done. I personally think that programmers should be comfortable with both text editor/make and visual/IDE programming. It sharpens your skills when you come at a problem from a different angle.
I think that the heart of this deal is really the heavy COM+/DNA component. MS is SOOOOO thrilled with themselves about these technologies, they need to see them shored up in as many major programming environments as possible, especially as more and more people are starting to figure out CORBA. I'm just glad to see that they didn't do a WFC licensing plan that will screw up JBuilder. Still, since I've been programming in MFC for the past couple of weeks (and OWL years ago), I can say that it will NOT be a particularly great addition to any programming suite. Come on MS, how much crap really NEEDS to start with the initials Afx? And why do you have both pVariable and lpVariable names when all pointers in Win32 are 32 bit pointers?
Glad to see they bumped up the version number by more than.01 this time. Anybody know if there's more serious functionality planned for 1.0, or are there just 0.40 bug patches left?
Ok, this is a few years back when I was first learning to program, but I swear it's true. I had an internship with the DoD, who really didn't need any interns. I brought a copy of Borland C++ 4.0 from home and installed it on a 386/25 (or maybe 40? I don't remember). The install was beautiful, of course, 24-odd diskettes, and if something went wrong (which it did) you had to start over, then they switched me to a different, identical computer, and I had to start over. So anyways, it was a simple Windows app, but I had never touched C++ (self-taught C only), let alone Windows, so I was recompiling every 5 minutes. Well, you know, 5 minutes to debug, 45 to compile, etc. God forbid I should hit "build all" at any point. The funny part was when I had to change some little detail with the font in an MDI app. The documentation for OWL (Borland's MFC equivalent) didn't mention the specific bit, so I called their support at like $40.00 a minute, or whatever they charge. The guy stays on the phone with me for half an hour, gets his supervisor, and then says he'll call me back when he figures it out. Two hours later I get a call that he's still working on it. Never heard from him again. Go figure. --JZ
Slashdot Mirror
I really wish the guys at Troll-Tech could work out a licensing agreement to make a version of Qt price-competitive with commercial Motif. I mean a Unix-only version exactly like Qt-free, but with a license to let you produce commercial software with it. It's an amazing product, and we could see Borland QtDelphi and C++-builder with Qt-GUI. Mmm...
Commercial and free development can, indeed must, exist side-by-side.
BTW: Whoever mentioned that they like Visual C++ should check out KDevelop (www.kdevelop.org). Its interface is basically identical to VC++.
--JZ
E-Trade lets you get in with only $1000 deposit, and you don't even have to invest all of that.
--JZ
Well, I was completely down on the Amiga idea until today. Never owned one, never cared (I used video toaster once though, but didn't get very far). But the tech brief is very interesting. For a decade, Intel, MS, and computer manufacturers have been tuning the OS and hardware to go together. So finally we'll get hardware that's tuned for linux and Java. Plus DVD, multimedia, and USB improvements that can be added to the kernel. HOPEFULLY their X-server will be free too, but it's too early to tell (X could use a few more folks working on hardware-accelerated OpenGL).
It's especially nice to hear that they're emphasizing developer tools. Anything we can do to speed up those ports and to finally get a platform that performs well with swing. Ok, so the chance is one in a thousand, but if a platorm actually came out that made Swing seem as fast as native graphics, I would buy it in a second.
--JZ
As much as I hate Pascal the language, I really think that Delphi offers the most important contribution to Linux. Especially if you can trivially port Windows Delphi apps to Linux! That would be unbelievably great, but also quite possible, as Delphi is reasonably removed from the direct Win32 API. C++ builder will soon be in MFC-land, so that would REALLY have to diverge from the Windows version. JBuilder 3 is coming, but written in Java, and it will be competing against VisualAge from IBM. Delphi is the truly unique product out of this bunch.
Oh yeah, and those of you who mentioned using Qt should check out www.kdevelop.org. KDevelop is still experimental, but actually quite good: Visual dialog builder, complete project management, class browser, integrated debugger, etc. And you don't have to use it for or with KDE if you don't want to.
--JZ
I believe that the "HardHat" Linux distro has plans to incorporate RTOS extensions in their next version. I'd check them out for details.
--JZ
So, are Amiga and QNX in a very tiny (somewhat pathetic, if you want my opinion) version of the old OS/2 vs. WinNT split? As in, "Sure, let's develop technology jointly. . . Oh wait, nevermind, how about if I just take my half of the tech and bail, leaving you along and pissed off?" How else do we explain the dual announcements of competing versions of the same product????
--JZ
The old OS/2 very much forms the core of the new one. NextStep was sold to Apple. And this article proves that people still see value in the Amiga OS (even if it IS just the name). Only GEM has been opened up. Maybe someday IBM will give up entirely on OS/2 (Wow, I just read THE WORST review of their new Warp Server, I think in Internet Week, but I can't remember) and set it free. Not that I'm saying it's a bad system. It could just be better if it could cross-fertilize code with other systems/programmers out there.
--JZ
Not a quantum leap either. They would have been smarter to put that much effort into a vesion of KDE or GNOME for QNX. Either of them could be made to look a helluva lot like that desktop.
And since QNX is still proprietary, it wouldn't really hurt them to have a GPLed desktop on top of it. In fact, they could have contributed to a GPLed desktop, and made their theme proprietary / copyrighted, so they'd still have a unique look and feel.
Companies need to get more creative. Let's stop reinventing the wheel, damnit!
Word on the street is that version 4.0 of XF86 should finally be kickin' it. Of course I agree with you that the current X, though laudable for its ever-improving hardware support and the dedication of its developers, has become a dinosaur. And that doesn't even mention the memory leaks! (I've found X can swell up to over 30 megs of RAM or more if I don't restart it).
In general, though, I would advise anyone who plans to use either GNOME or KDE to buy 32 more megs of RAM than they would for a similar Windows box. At current memory prices (about $40 for that DIMM), that's still cheaper than buying Win 98.
--JZ
will be the establishment of the common CORBA mappings. I'd love to see KDE's KOM/Open Parts component model broaden to the point where I can easily write an app in GTK+ or GTK--, throw a KOM wrapper around it, and then use it seamlessly with KDE KOM parts. Imagine... A Gnumeric spreadsheet embedded in KWord!
--JZ
PacketStorm was quite clearly in violation of Harvard's acceptable use policies, which, like most universities, actually permit VERY LITTLE. Their correct response should have been to cut off the server from the outside world, allow Ken to get his data, and be done with it. .) to uphold, trustees to satisfy, and lawyers to please.
Their real mistake was inviting him onto the network in the first place. A site like this, operated by one person and centering on controversial issues like cracking and information security, is bound to attract unwanted attention. It should NOT be located on the net of a private, conservative university which has a repuation (albeit not a good one among people I know. .
--JZ
If you have trouble with Blackdown, I'd suggest taking a look at the IBM free JVM for Linux. IBM's virtual machine is extremely well reviewed as one of the fastest on the market (comparable to HotSpot), but I think they might still be at 1.1.6.
Check out the results for the 80/20 banded file set. A key result is that Reiserfs performs MUCH faster when large numbers of files are grouped in one directory. Copying all the files from one directory under Reiserfs was 12 times faster than the same operation under ext2. Scanning all files (if located in only one directory) was improved by a factor of 6. These results drop to ratios of 2 or 3 to 1 if the files are in many subdirectories. This uses a mixed set of both large an small files (about 50 megs' worth), but emphasizes the very smallest files (12,000 of the files are under 100 bytes). Those really small files provide reiserfs with the big edge.
However, its results are still quite good for medium-large files and the author admits that he's done virtually no optimizing for larger files. I suspect that, if Reiserfs catches on, we'll quickly see these optimizations in the next few months. Thus, I'd wait a little while and see how things pan out before putting it into use.
Ok, so we've all noted that peformance is awesome for 100 byte files. But check out his section on larger files too. The benchmarks of files all greater than 512k are also quite impressive: .16 second rm -rf * time for 80+ megs in a series of subdirectories is pretty damn impressive.
Look at the rm -rf * comparison--
Ext2: 0:02.63
Reiserfs: 0:00.16
Ratio:16
Wow, now a new user logged in as root can wipe out his/her system 16 times faster than ever!
Seriously, though, a
--JZ
Ok, so we've all noted that peformance is awesome for 100 byte files. But check out his section on larger files too. The benchmarks of files all greater than 512k are also quite impressive:
Look at the rm -rf * comparison--
Ext2: 0:02.63
Reiserfs: 0:00.16
Ratio:16
No, it's much WORSE with B-trees than binary trees. "Damnit, did I remember to split the /usr/src node on the downward traversal last time? Or should I be splitting when I back up out of the recursion? I don't even know any more. . ."
Go out to all your local bookstores, computer stores, etc., and find and destroy every copy of RH5.2, Caldera 1.3, and SuSE5.2 wherever they might be discovered by newbies.
Seriously, I've actually been in stores that were selling SuSE5.2 and SuSE6.1 side-by-side, and the 5.2 COST MORE!! Same with Caldera. Plus, they have this "Linux Starter Kit", which is supposed to appeal to newbies, but only includes Caldera 1.3. These people (looking to play around with a desktio) will be sorely disappointed to discover FVWM95, an ancient version of the gimp, and maybe StarOffice 4.0.
Don't get me wrong, a solid install of an older linux can run great as a server, but they just present outdated installs and desktops, perpetuating the stereotype that Linux is too hard to use.
--JZ
Yes, my SuSE 6.1 is fantastic. Great install (that also installed the 4front sound drivers and saved me a helluva long time), 5 CDs, and they even throw in an old vesion of KOffice (do you know how hard it is to compile KOffice yourself?!?).
Also much, much, much cheaper than RH ( I got it for $25-ish at Chumbo, well worth it).
Well, I'm happy for the Be guys. I wonder how much Microwerkz pays for each copy of the OS + Gobe.
More to the point, it would be really cool if we could get linux into this kind of box (yes, I know there's a small set-top company using it, but there are a million set tops that run everything, we'll see which one wins).
Linux has three key problems in this department. The first is boot time. Anybody know of projects to build an insta-boot linux for embedded devices?
The second, the lack of a journalled file system, has been beaten to death. Consumers just want to be able to flick it on and flick it off. We'll see how ext3 does, but I wouldn't be surprised if it took a splinter filesystem to work with small consumer devices. They have such different needs from enterprise servers that they shouldn't necessarily have the same features.
The final kicker is XF86. Great software in many ways, and everything I've heard about 4.0 sounds incredible, but it's not small. Anybody have an update on some of the attempts to create a micro-sized graphics system (GGI and Berlin? I can't really remember). . .
--JZ
Try KDevelop. Still alpha, but it does integrate Kdbg for debugging. Also has a dialog editor, class browser, syntax highlighting, project management (all this in version 0.4!).
Well, I suppose I'd have to agree with you about the state of Java GUI design. And, luckily, Java is a lot easier when it comes to writing GUIs by hand, especially in terms of the variety of layout managers (though all of them are still fairly annoying). I found some use for JBuilder, but its layout facilities were not up to par.
I'm actually a big fan of VC++'s visual layout. Fast, easy, excellent for adding member variables and functions to go with components. And, as with JavaBeans, ActiveX controls can be added visually even if the IDE has never seen them before.
In terms of code resuse, I certainly agree that VB doesn't promote it. Too many properties, not enough methods. It's just not OOP.
There are three valid uses for an ide:
-Class browser: See a tree with all the classes/members in your project and just hit one click to take you exactly to a method definition. Or view a graphical inheritance tree.
-Syntax completion: If you're just doing straight Unix C systems programming, no GUIs, and you have plenty of experience, sure you only need to look up the occasional command. But if you're programming against a massive GUI API (Windows, KDE, etc.) with a class library that you didn't write, syntax completion is key. Who the hell knows all three overloaded forms of some obscure command to manipulate a tree control? Who the hell really wants to? An IDE will pop up a parameter list for you instantly.
-GUI design: Look, creating dialogs is not going to be an intellectually stimulating programming task. So do it with a visual editor in 5 minutes, rather than tweaking around manually for hours. Then automatically generate the message handlers for the dialog and get started with real programming.
Good IDEs can make a serious productivity difference if they're used for the right tasks and if you KNOW HOW TO USE THEM. There is very much a mindset to each IDE, just as there is to Emacs, vi, etc. If you approach Visual C++ by thinking of it as just an even bigger, slower version of emacs with different keybindings, it will do you know good. But if you think in its visual terms, you'll get a lot more done.
I personally think that programmers should be comfortable with both text editor/make and visual/IDE programming. It sharpens your skills when you come at a problem from a different angle.
I think that the heart of this deal is really the heavy COM+/DNA component. MS is SOOOOO thrilled with themselves about these technologies, they need to see them shored up in as many major programming environments as possible, especially as more and more people are starting to figure out CORBA. I'm just glad to see that they didn't do a WFC licensing plan that will screw up JBuilder. Still, since I've been programming in MFC for the past couple of weeks (and OWL years ago), I can say that it will NOT be a particularly great addition to any programming suite. Come on MS, how much crap really NEEDS to start with the initials Afx? And why do you have both pVariable and lpVariable names when all pointers in Win32 are 32 bit pointers?
Glad to see they bumped up the version number by more than .01 this time. Anybody know if there's more serious functionality planned for 1.0, or are there just 0.40 bug patches left?
Ok, this is a few years back when I was first learning to program, but I swear it's true. I had an internship with the DoD, who really didn't need any interns. I brought a copy of Borland C++ 4.0 from home and installed it on a 386/25 (or maybe 40? I don't remember). The install was beautiful, of course, 24-odd diskettes, and if something went wrong (which it did) you had to start over, then they switched me to a different, identical computer, and I had to start over.
So anyways, it was a simple Windows app, but I had never touched C++ (self-taught C only), let alone Windows, so I was recompiling every 5 minutes. Well, you know, 5 minutes to debug, 45 to compile, etc. God forbid I should hit "build all" at any point.
The funny part was when I had to change some little detail with the font in an MDI app. The documentation for OWL (Borland's MFC equivalent) didn't mention the specific bit, so I called their support at like $40.00 a minute, or whatever they charge. The guy stays on the phone with me for half an hour, gets his supervisor, and then says he'll call me back when he figures it out. Two hours later I get a call that he's still working on it. Never heard from him again. Go figure.
--JZ