No, your life was simplified by a change of attitude. You could have accomplished the same thing by simply caring less about punctuality. Throwing away your watch may have helped you change your attitude, and the battery failing may have been the impetus, but the technology itself had no effect on you.
You can play all kinds of semantic games but you're closing your eyes and plugging your ears if you don't thing Apple does everything it can to get you to pay that $130 a year. It's called "recurring profit stream" and it's what every software company has. Rapidly obseleting old versions is a good way to do it, but it's going to piss people off. Apple doesn't apologize for it and you don't need to do it for them.
Make gcc a symlink to icc and you should be good to go for almost all apps - ICC is 90% (roughly) source-compatible with GCC, even supporting most of the extensions, and it supports all of GCCs command line arguments.
I'm a recent Firebird convert and it loads (much) faster than Mozilla. It's maybe an extra second at first startup over IE, if that - and thats without a preloader.
Well, theres civilians and then theres civilians. The Pentagon is pretty much how you'd define a military target, granted, but how about a steel factory? Think of any of the attacks on infrastructure during WW2, and the ever-popular Hiroshima.
They want to push.NET in the enterprise (thats where they make money, consumers are a secondary market). Hopefully,.NET will then trickle down to home users, and people will want to upgrade because the apps they use at work require it.
(As an aside, as a C#/.NET developer myself, app deleveopment with.NET only seems easy because MFC was so friggin bad. Using a good toolkit library makes working with friggin Windows Forms feel like ripping my own testes off. It's better than Java and thats about all I have to say for it.)
If DHCP is inherently insecure, then it shouldn't be auto-discovered out of the box. Thats a trade off between convenience and security, and while there may be very good reasons for making that decision on the side of convenienve, it's STILL A SECURITY RISK. Period.
Oh, and your network doesn't need to be compromised if you're on or near a malicious wireless network, as OS X will cheerfully auto-discover that one as well.
It's not some earth shattering "all your base" sort of flaw, but then, there really aren't very many of those. It IS, however, a real, verifiable flaw. Part of the flaw is in the design. You don't need to jump up and down defending OS X here, it makes you look like a drooling fanboy.
Local exploits are still exploits - the vast majority of Windows exploits are local only, for example.
Linking statically to wx adds about 400k, using wxMSW and MSVC++. You get a little more as you use more classes, but it grows at a fairly slow pace. The CVS version of wx splits the library into several sub-libraries, which lets you get that even smaller if, say, you don't need network libraries or the HTML control. AOL Communicator was indeed built with wxWindows, and while they did patch wx I don't believe it was any major structural changes.
It's possible that you were using the Universal build of wxWindows, which draws its own widgets and doesn't get as much attention as the other ports. If you weren't, then almost all the controls you see (including the text control) were provided by Gtk, so if they look icky it's because Gtk looks icky. It's also possible that you were using a Gtk 1 build, and your other Gtk apps are using Gtk2. wxWindows does very little custom drawing, so any zigzags you were seeing are unlikely to be a bug in it.
In fairness, almost everyone I've known has had no trouble whatsoever getting wxWindows to work on windows assuming they followed the step by step instructions. Setting up your own project file can be daunting the first time, but if really had lots of problems that weren't solved on the mailing list then odds are you either were annoying enough that nobody wanted to help you, you were making a common mistake spelled out in the docs.
The event system is a matter of preference - I should point out that while it's similiar in use to MFCs, it is by no means identical. I vastly prefer it over signals/slots.
The Qt non-commercial addition does not come with source files. It includes header files and a runtime binary. You cannot re-create that binary from the GPLed sources. By no reasonable standard does the existence of source for the GPLed version equal source for the non-commercial Windows version.
While you may be (pedantically) correct, the fact is that a standard C++ compiler cannot create a working executable from Qt sourcecode without a non-standard tool. This is hardly standard use of C++.
With the previous QT nc, only MCVC++ (6.0) and Borland free compiler were supported. Actually, I'm not even sure about the Borland one. Qt/Win32 (the commercial version) supports a whole slew of compilers, so I wouldn't make too many assumptions about what compilers are supported.
wxWindows does use native widgets. However, theres alot more to wx (and to Qt, for that matter) than just a set of widgets - there's event handling, application frameworks, all sorts of convenience classes, etc, etc, etc.
Theres also a great deal of wrapping done to present a consistent UI over inconsistent platform-specific APIs and widgets. Therefore, there is indeed a wx library which you must link to. You can, however, statically link to it which allows you to ship a single executable rather than an EXE + binary. I'm not sure if Qt allows this or not (the non-commercial version does not).
There are also, of course, the licensing differences.
.chm are compiled help files - they aren't executable (they're loaded & displayed by MS Help). HTAs have the.hta extension. This exploit takes advantage of the way MS uses filenames with CLSIDs in them to indicate "special" behaviour by the shell. For example, this is how Explorer knows to have special behavior in the Font and Temporary Internet Files folders. I'm guessing this CLSID is associated with the html viwer thats used for HTA-based parts of the OS dialogs, like the add/remove files dialog.
I'm mostly guessing here but it looks the the CLSID identifies it as an HTA (HTML application) component, which MS was hyping as all the rage in application developlment a few years back. Basically, it's like an XUL app - written in HTML and JScript. Portions of the Win2k+ UI are written using it, like the add/remove programs dialog.
Amusingly, this would make me blink because actual PDFs open automatically in IE (using the Adobe plugin) and I have to use "Save as..." to get them to disk.
Theres a couple other inconsistencies - if you do use "Save as" the filename appears to be PDF, but the filetype pre-filter (which is set to the type of file that you're downloading) is "HTML files". Interestingly, in the "open or save" dialog, the file type is blank.
I'd just like to take this time to slap microsoft for adding yet another way of associating files with applications to piss us all off. We already had enough issues with contradicting file extensions and mime types.
This has alot to do with the lack of compliant compilers, too - exception behavior and especially template support was sporadic and very platform and version dependent. The situation is much better now and you're seeing more "correct" C++ code being used, although I agree that it's still largely taught as "C with objects". Not that "C with object s" is neccesarily a bad way to program.
It's worth noting that the non-commercial versions are binary-only. Not an issue for many people, of course, but I like having source code for my libraries (if only because I like to know whats going on if something doesn't work).
Also, this non-commercial version does not appear to be available on the Trolltech website, so you'll need to get the book (and the text of the license for it doesn't appear to be available either, although I'd assume it's similiar to the 2.3 non-comm license).
I'd just like to respond by saying that while I'm sure your wife is very objective and honest with her job, not all the people who do her sort of work are (I'm originally from the northwest coast, and biased/incomplete environmental impact studies are the things of legend there). Unless theres serious oversight or a single source of testing (like the FDA for drugs), then companies (at least the sleazier ones) will gravitate toward "researchers" who can be guaranteed to provide the results they want.
That said, the LAST microsoft study that everyone on Slashdot screamed about turned out to reveal legitimate problems with the 2.2 kernel under heavy load (thundering herd problem), and was the impetous for major scheduler improvments in 2.4
That doesn't make anything I said not true. Especially the pixel-precision layout issues, which can be really annoying when you're generating dynamic content. Writing pure CSS layouts these days is just like writing table based layouts in the early 90s - you spend alot of time and energy accounting (with usually less than 100% success) for various browser issues, and with generally less than spectacular results. Theres no point in it, unless you're pedantic. Note that "Pure CSS-XHTML seperation" rarely succeeds anyway - the way you tag your content into parent/child relationships affects your layout, no matter what you do. The way that browsers are able to scale and size tables based on thier content is the boon of the web programmer.
Slashdot Mirror
No, your life was simplified by a change of attitude. You could have accomplished the same thing by simply caring less about punctuality. Throwing away your watch may have helped you change your attitude, and the battery failing may have been the impetus, but the technology itself had no effect on you.
You can play all kinds of semantic games but you're closing your eyes and plugging your ears if you don't thing Apple does everything it can to get you to pay that $130 a year. It's called "recurring profit stream" and it's what every software company has. Rapidly obseleting old versions is a good way to do it, but it's going to piss people off. Apple doesn't apologize for it and you don't need to do it for them.
Make gcc a symlink to icc and you should be good to go for almost all apps - ICC is 90% (roughly) source-compatible with GCC, even supporting most of the extensions, and it supports all of GCCs command line arguments.
Gentoo is the coolest because it boots to a kick-ass purple framebuffer terminal.
No problems with multiple monitors and Firebird for me, so it's not a generic problem.
I'm a recent Firebird convert and it loads (much) faster than Mozilla. It's maybe an extra second at first startup over IE, if that - and thats without a preloader.
Well, theres civilians and then theres civilians. The Pentagon is pretty much how you'd define a military target, granted, but how about a steel factory? Think of any of the attacks on infrastructure during WW2, and the ever-popular Hiroshima.
(As an aside, as a C#/.NET developer myself, app deleveopment with .NET only seems easy because MFC was so friggin bad. Using a good toolkit library makes working with friggin Windows Forms feel like ripping my own testes off. It's better than Java and thats about all I have to say for it.)
Dude, you forgot to slip on a bananna peel.
Oh, and your network doesn't need to be compromised if you're on or near a malicious wireless network, as OS X will cheerfully auto-discover that one as well.
It's not some earth shattering "all your base" sort of flaw, but then, there really aren't very many of those. It IS, however, a real, verifiable flaw. Part of the flaw is in the design. You don't need to jump up and down defending OS X here, it makes you look like a drooling fanboy.
Local exploits are still exploits - the vast majority of Windows exploits are local only, for example.
Linking statically to wx adds about 400k, using wxMSW and MSVC++. You get a little more as you use more classes, but it grows at a fairly slow pace. The CVS version of wx splits the library into several sub-libraries, which lets you get that even smaller if, say, you don't need network libraries or the HTML control. AOL Communicator was indeed built with wxWindows, and while they did patch wx I don't believe it was any major structural changes.
It's possible that you were using the Universal build of wxWindows, which draws its own widgets and doesn't get as much attention as the other ports. If you weren't, then almost all the controls you see (including the text control) were provided by Gtk, so if they look icky it's because Gtk looks icky. It's also possible that you were using a Gtk 1 build, and your other Gtk apps are using Gtk2. wxWindows does very little custom drawing, so any zigzags you were seeing are unlikely to be a bug in it.
The event system is a matter of preference - I should point out that while it's similiar in use to MFCs, it is by no means identical. I vastly prefer it over signals/slots.
The Qt non-commercial addition does not come with source files. It includes header files and a runtime binary. You cannot re-create that binary from the GPLed sources. By no reasonable standard does the existence of source for the GPLed version equal source for the non-commercial Windows version.
While you may be (pedantically) correct, the fact is that a standard C++ compiler cannot create a working executable from Qt sourcecode without a non-standard tool. This is hardly standard use of C++.
With the previous QT nc, only MCVC++ (6.0) and Borland free compiler were supported. Actually, I'm not even sure about the Borland one. Qt/Win32 (the commercial version) supports a whole slew of compilers, so I wouldn't make too many assumptions about what compilers are supported.
Theres also a great deal of wrapping done to present a consistent UI over inconsistent platform-specific APIs and widgets. Therefore, there is indeed a wx library which you must link to. You can, however, statically link to it which allows you to ship a single executable rather than an EXE + binary. I'm not sure if Qt allows this or not (the non-commercial version does not).
There are also, of course, the licensing differences.
The major point of wxWindows is that it uses native, not emulated widgets (unlike all other cross platform toolkits that I'm aware of).
.chm are compiled help files - they aren't executable (they're loaded & displayed by MS Help). HTAs have the .hta extension. This exploit takes advantage of the way MS uses filenames with CLSIDs in them to indicate "special" behaviour by the shell. For example, this is how Explorer knows to have special behavior in the Font and Temporary Internet Files folders. I'm guessing this CLSID is associated with the html viwer thats used for HTA-based parts of the OS dialogs, like the add/remove files dialog.
I'm mostly guessing here but it looks the the CLSID identifies it as an HTA (HTML application) component, which MS was hyping as all the rage in application developlment a few years back. Basically, it's like an XUL app - written in HTML and JScript. Portions of the Win2k+ UI are written using it, like the add/remove programs dialog.
Theres a couple other inconsistencies - if you do use "Save as" the filename appears to be PDF, but the filetype pre-filter (which is set to the type of file that you're downloading) is "HTML files". Interestingly, in the "open or save" dialog, the file type is blank.
I'd just like to take this time to slap microsoft for adding yet another way of associating files with applications to piss us all off. We already had enough issues with contradicting file extensions and mime types.
This has alot to do with the lack of compliant compilers, too - exception behavior and especially template support was sporadic and very platform and version dependent. The situation is much better now and you're seeing more "correct" C++ code being used, although I agree that it's still largely taught as "C with objects". Not that "C with object s" is neccesarily a bad way to program.
Also, this non-commercial version does not appear to be available on the Trolltech website, so you'll need to get the book (and the text of the license for it doesn't appear to be available either, although I'd assume it's similiar to the 2.3 non-comm license).
That said, the LAST microsoft study that everyone on Slashdot screamed about turned out to reveal legitimate problems with the 2.2 kernel under heavy load (thundering herd problem), and was the impetous for major scheduler improvments in 2.4
That doesn't make anything I said not true. Especially the pixel-precision layout issues, which can be really annoying when you're generating dynamic content. Writing pure CSS layouts these days is just like writing table based layouts in the early 90s - you spend alot of time and energy accounting (with usually less than 100% success) for various browser issues, and with generally less than spectacular results. Theres no point in it, unless you're pedantic. Note that "Pure CSS-XHTML seperation" rarely succeeds anyway - the way you tag your content into parent/child relationships affects your layout, no matter what you do. The way that browsers are able to scale and size tables based on thier content is the boon of the web programmer.