What did we say about never getting it right the first time?
Here is the version that works.
//Assumes, str is null terminated. void revStr(char *str) {
int start=0/*forgot to do this before*/,end=strlen(str);
char tmp;
do
{
tmp = str[start];
str[start]= str[end];
str[end] = tmp;
}while (++start/*using != will work, but will segfault on string with len of 1*/ --end); }
IIRC, they do exist. I don't know about Linux, but I'm certain that some Unixes has those on 32bit. For Windows, apperantly it has it from Win2K onward. Here is the info I dug after a quick search: http://msdn.microsoft.com/library/default .asp?URL= / ibrary/backgrnd/html/awewindata.htm
*Shudder*, I still think that I've somewhere a spare matrix I wrote a long time ago as an experiment. Gawd, talk about an ugly thing. It had about 5 methods, and 6,00 LOC.
Why would you *want* loose coupling? What is the big advantage that it has? That you can switch components?
That was never in the agenda, you know.
Beside, you want to make it so? A method of doing something like this: HKLM\StdApp\Browser\GUID = {GUID_OF_USER_BROWSER} Would give you what you want, of course, it wouldn't work on legacy application, (meaning all of the applications), and you would've to convice programmers to risk *more* potential configuration problems.
1> ISVs were prevented from using whatever product they wanted? Can you provide some proof for that, please? To the best of my knowledge, OEM might have some problems with this, but I've never heard that an Independent Software Vendor had this.
2> Because it's a support problem. Yes, the interfaces *are* publish, and it's a piece of cake to copy them and use them, the problem isn't *that*, the problem is support. You don't seem to understand, because you apperantly come from a linux set of mind, which says: Let's define a standard for web-browser, and then let a lot of other people to implement it, and let the user choose. That is all fine in theory, but in practice it breaks, because unless you provide a standard so rigid it's actually source code, you would've problems with bugs and different behaviour that would cause problems.
A, B, C, you don't seem to live in reality. Don't you understand that just getting the full API, means nothing? Just take a debugger, and trace all the calls that Office does, okay? Find me one that isn't on the official documentation, then we will talk. To my knowledge, it's a Win 2.0 story that kept rolling.
The problem is *always* implementation, just check the JVM market and the problems that differing JVMs sometimes have when they hit a bug. It's nasty, and that is on a system that was designed from the first place to be implemented by many people.
You need only to remember Write-Once-Test-Everywhere that was (still is, as a matter of fact) Java motto, to understand that just being standard compliance isn't sufficent. We aren't talking about creating some HTML output, getting this to be a W3C standard is pretty easy, unless you want to go really fancy, I'm talking about *embedding* an HTML control inside your application. That is quite a fit, you know. And even minute differences will not allow you to complete this successfully. Not to mention that while IE is designed to be modular so you can use it from other applications, most other browsers aren't. Mozilla have that capability, AFAIK, but that about it, to the best of my knowledge.
And there is well-tested in a configuration that you can expect, and then there is well tested in a configuration that you *can't* expect. You need only to look at Unix a decade ago to understand what the problem is, there wasn't *any* way that an ISV could create an application that would run successfully on all Unixes (even with recompiling it), because there isn't any way to make assumsations about it. The way Linux handle it today is by having the LSB, and putting the burden on the user, that is all fine and good to say that "you need W3C conforming browser that can be embedded, version X and above", to run this application, but this doesn't mean *squat* to the average user. And *that* is what counts.
In addition, Quicken isn't the typical application today. Quicken is a well establish product, it come on CD, it has a reputation, etc. It can *afford* to have IE on it's installation CD. But what about all those small programs? All those programmers that just wouldn't be *able* to publish their efforts because they wouldn't work outside their own computer, because there isn't anything to make assumstions on.
About lack of MSHTML & web control, yes, most MS and a *lot* of other applications *would* break if they wouldn't be present. And that doesn't imply that they aren't modular. They just depend on this module to be present, there is a difference between modularity and dependecies. MSHTML is what MS insist should be in Windows, it already offered (and was turned down) to remove the IE executable and the icon from the dekstop. But the library has to remain, because otherwise, you break a lot of software.
Let me explain, let's say that I'm an ISV with an interest in creating some sort of application for windows. My interest is in using as much of the OS's abilities as I can. Now, on XP, I've no *need* to write/aquire a media player module if I want to play sounds, I have it built-in, and I don't need to bother if I need to display HTML, nor would I've any trouble if I wanted to do some 3D animation. I still have to test thouroghfully, but there is a configuration that I can take as granted. The reason that Java is considered multi-platform and C++ not is that it offers far richer standard library than C++, same for windows.
But on the world that the states demand, I would have to do one of the following: 1> Use some browser/media player/3d that I buy, aquire, or get. That would require that I would install it with my application, this put more burden on me, while giving me no benefit, after all, if I want to use some competing product in my application, I already *can*. 2> What most of you seems to be suggesting, use whatever product the user have installed. And *that* is what all this "considerable testing" is all about. I've to test my product with every possible version of applications that I use/interact. This mean that I've to test with Mozilla/IE/Opera/SomeBuggyBrowser, XMMS/WMP/WinAmp, DX/OpenGL/Glide, etc...
And I've to check all the prossible versions, and check for bugs, etc. *Then* I'll have to check that the user actually *has* one of the supported applications on, and have to face the users' anger if their favoraite browser/media player/3d software isn't supported, etc, etc.
No, you don't seem to understand the difference between modularity and dependecies. Software dependecies are very intricate, not matter how modular they are.
You've component A, and component B, B won't function correctly without A, you might want to change A to C, but there is a chance that even the slightest change between A & C would cause B to mal-function. Result, support nightmare.
Actually, the interfaces *are* public, and any Mozilla actually *do* implements it. And if any application creator wants to use Mozilla instead of IE all he has to do is to replace a *single* string in his code, and recompile. (Or just provide a binary patch, 5 seconds work.) And it would work just as well. On course, this means twice as much testing as he had done before, and a support problem, and...
You *do* realize that the reason that there isn't ports to those systems is not that they can't do it. There *were* ports to Alpha & PPC, (as well as MIPS), the problem is that they didn't *support themselves*, so they had to be canceled. MS has to support everything that they make, and that cost money, if a product isn't making money for them, or isn't going to, they will can it, and move to the next thing.
> Generally, I'm coming down on the side that says CORBA (and possibly COM, if it's fundamentally similar to CORBA) is great and cool, but utterly and completely inappropriate for writing plug-ins, in the traditional sense.
Actually, CORBA is bad for plug-ins, but COM is great. That is beacuse COM has in-proc servers, meaning that it is on the same adress space, and is extremely fast. Most of the plugins for MS products are COM components that implements a particular interface.
NT-SP6 & Lotus Notes: Lotus fucked up really back in conforming to the way that they should've implemented it, and got away with it because of a hole in NT that SP6 fixed. It was out only a few days when SP6a came out with a version that wouldn't break Lotus' faulty implementation.
DR-DOS & Win3.11: Are you talking about using a system on an unsupported product? That *is* likely to cause problems. Or are you talking about the beta's message box?
Name *one* proven instance when MS changed the OS interface to harm Netscape. Just one, and then prove it.
You confuse modalurity with interchangibility. Windows is highly modular, but parts aren't interchangable. If you want to replace IE, you can, but you've to make an exact clone of IE.
And MS doesn't say that this is impossible, it's quite easy to do, (although can cause a support nightmare from hell). MS says that *removing* IE will break a lot of stuff.
There shouldn't be a big fuss, if you want to replace IE, you need to re-implement the interfaces, and then you would change a couple of fields in the registry, and that would be *all*.
The problem is that MS can't support this configuration, and many applications may break if there is even just a subtle change in the way it works.
> They have the power to change the interfaces that other software uses, forcing the developers of that software to update their products to work with newer versions of windows
*Big time false*. MS is *excellent* at backward compatability. If it worked on earlier version of Windows, than there is little chance that it would break on a newer version.
The only excepetion is beta API, which you shouldn't rely upon anyway.
No, it doesn't have to be Microsoft Dlls, and even now, you can replace it quite easily. Let's assume that Mozilla correctly implements all of IWebBrowser interfaces. What does it takes to replace IE with mozilla? Change several keys in the registry, and that is all.
There *isn't* any other expanse to it. If a Real wants to replace WMP, all they need is to implement all the publish interfaces, and register them on the published GUID. That is all.
I forgot to mention, the HTML component can be removed, the problem is that a *lot* of programs are using it. Including many MS programs. You can have a Windows which is just the kernel & shell, that isn't very hard to do, the problem is that there are less than 1000 programs for NT proper. And all of them are usually administration stuff. Nothing that can be of any use.
Not a single Win32 program would run on it.
If you add Win32 subsystem, then you get much better support, but you still get a lot of missing stuff. A *lot* of programs assume that the other components will be there. From the shell (All the SH* API) to IE (to Wininet.h function).
How do you upload a file programtically to FTP site on Linux? Is there a standard way? What happen if you remove it? How many stuff would break?
Slashdot Mirror
What did we say about never getting it right the first time?
/*using != will work, but will segfault on string with len of 1*/ --end);
Here is the version that works.
//Assumes, str is null terminated.
void revStr(char *str)
{
int start=0/*forgot to do this before*/,end=strlen(str);
char tmp;
do
{
tmp = str[start];
str[start]= str[end];
str[end] = tmp;
}while (++start
}
Assuming you meant a constant ammount of memory, and not finite, that isn't very hard.
//Assumes, str is null terminated.
void revStr(char *str)
{
int start,end=strlen(str);
char tmp;
do
{
tmp = str[start];
str[start]= str[end];
str[end] = tmp;
}while (++start!=--end);
}
IIRC, they do exist.t .asp?URL= / ibrary/backgrnd/html/awewindata.htm
I don't know about Linux, but I'm certain that some Unixes has those on 32bit.
For Windows, apperantly it has it from Win2K onward.
Here is the info I dug after a quick search:
http://msdn.microsoft.com/library/defaul
6GB has table?
You've better have a *really* big memory, or you'll suffer terrible performance due to pagin.
there is a compiler switch for it, btw.
*Shudder*, I still think that I've somewhere a spare matrix I wrote a long time ago as an experiment.
Gawd, talk about an ugly thing.
It had about 5 methods, and 6,00 LOC.
Why would you *want* loose coupling? What is the big advantage that it has? That you can switch components?
That was never in the agenda, you know.
Beside, you want to make it so?
A method of doing something like this:
HKLM\StdApp\Browser\GUID = {GUID_OF_USER_BROWSER}
Would give you what you want, of course, it wouldn't work on legacy application, (meaning all of the applications), and you would've to convice programmers to risk *more* potential configuration problems.
1> ISVs were prevented from using whatever product they wanted? Can you provide some proof for that, please? To the best of my knowledge, OEM might have some problems with this, but I've never heard that an Independent Software Vendor had this.
2> Because it's a support problem. Yes, the interfaces *are* publish, and it's a piece of cake to copy them and use them, the problem isn't *that*, the problem is support.
You don't seem to understand, because you apperantly come from a linux set of mind, which says:
Let's define a standard for web-browser, and then let a lot of other people to implement it, and let the user choose.
That is all fine in theory, but in practice it breaks, because unless you provide a standard so rigid it's actually source code, you would've problems with bugs and different behaviour that would cause problems.
A, B, C, you don't seem to live in reality.
Don't you understand that just getting the full API, means nothing?
Just take a debugger, and trace all the calls that Office does, okay? Find me one that isn't on the official documentation, then we will talk.
To my knowledge, it's a Win 2.0 story that kept rolling.
The problem is *always* implementation, just check the JVM market and the problems that differing JVMs sometimes have when they hit a bug.
It's nasty, and that is on a system that was designed from the first place to be implemented by many people.
You need only to remember Write-Once-Test-Everywhere that was (still is, as a matter of fact) Java motto, to understand that just being standard compliance isn't sufficent.
We aren't talking about creating some HTML output, getting this to be a W3C standard is pretty easy, unless you want to go really fancy, I'm talking about *embedding* an HTML control inside your application.
That is quite a fit, you know. And even minute differences will not allow you to complete this successfully.
Not to mention that while IE is designed to be modular so you can use it from other applications, most other browsers aren't.
Mozilla have that capability, AFAIK, but that about it, to the best of my knowledge.
And there is well-tested in a configuration that you can expect, and then there is well tested in a configuration that you *can't* expect.
You need only to look at Unix a decade ago to understand what the problem is, there wasn't *any* way that an ISV could create an application that would run successfully on all Unixes (even with recompiling it), because there isn't any way to make assumsations about it.
The way Linux handle it today is by having the LSB, and putting the burden on the user, that is all fine and good to say that "you need W3C conforming browser that can be embedded, version X and above", to run this application, but this doesn't mean *squat* to the average user.
And *that* is what counts.
In addition, Quicken isn't the typical application today. Quicken is a well establish product, it come on CD, it has a reputation, etc.
It can *afford* to have IE on it's installation CD.
But what about all those small programs? All those programmers that just wouldn't be *able* to publish their efforts because they wouldn't work outside their own computer, because there isn't anything to make assumstions on.
About lack of MSHTML & web control, yes, most MS and a *lot* of other applications *would* break if they wouldn't be present. And that doesn't imply that they aren't modular.
They just depend on this module to be present, there is a difference between modularity and dependecies.
MSHTML is what MS insist should be in Windows, it already offered (and was turned down) to remove the IE executable and the icon from the dekstop. But the library has to remain, because otherwise, you break a lot of software.
Let me explain, let's say that I'm an ISV with an interest in creating some sort of application for windows.
My interest is in using as much of the OS's abilities as I can.
Now, on XP, I've no *need* to write/aquire a media player module if I want to play sounds, I have it built-in, and I don't need to bother if I need to display HTML, nor would I've any trouble if I wanted to do some 3D animation.
I still have to test thouroghfully, but there is a configuration that I can take as granted.
The reason that Java is considered multi-platform and C++ not is that it offers far richer standard library than C++, same for windows.
But on the world that the states demand, I would have to do one of the following:
1> Use some browser/media player/3d that I buy, aquire, or get. That would require that I would install it with my application, this put more burden on me, while giving me no benefit, after all, if I want to use some competing product in my application, I already *can*.
2> What most of you seems to be suggesting, use whatever product the user have installed. And *that* is what all this "considerable testing" is all about. I've to test my product with every possible version of applications that I use/interact. This mean that I've to test with Mozilla/IE/Opera/SomeBuggyBrowser, XMMS/WMP/WinAmp, DX/OpenGL/Glide, etc...
And I've to check all the prossible versions, and check for bugs, etc.
*Then* I'll have to check that the user actually *has* one of the supported applications on, and have to face the users' anger if their favoraite browser/media player/3d software isn't supported, etc, etc.
No, you don't seem to understand the difference between modularity and dependecies.
Software dependecies are very intricate, not matter how modular they are.
You've component A, and component B, B won't function correctly without A, you might want to change A to C, but there is a chance that even the slightest change between A & C would cause B to mal-function.
Result, support nightmare.
Actually, the interfaces *are* public, and any Mozilla actually *do* implements it.
And if any application creator wants to use Mozilla instead of IE all he has to do is to replace a *single* string in his code, and recompile. (Or just provide a binary patch, 5 seconds work.)
And it would work just as well.
On course, this means twice as much testing as he had done before, and a support problem, and...
You *do* realize that the reason that there isn't ports to those systems is not that they can't do it.
There *were* ports to Alpha & PPC, (as well as MIPS), the problem is that they didn't *support themselves*, so they had to be canceled.
MS has to support everything that they make, and that cost money, if a product isn't making money for them, or isn't going to, they will can it, and move to the next thing.
Hey!
Watch out, dude, I did it more than once, and then use edline to edit.
> Generally, I'm coming down on the side that says CORBA (and possibly COM, if it's fundamentally similar to CORBA) is great and cool, but utterly and completely inappropriate for writing plug-ins, in the traditional sense.
Actually, CORBA is bad for plug-ins, but COM is great.
That is beacuse COM has in-proc servers, meaning that it is on the same adress space, and is extremely fast.
Most of the plugins for MS products are COM components that implements a particular interface.
NT-SP6 & Lotus Notes:
Lotus fucked up really back in conforming to the way that they should've implemented it, and got away with it because of a hole in NT that SP6 fixed.
It was out only a few days when SP6a came out with a version that wouldn't break Lotus' faulty implementation.
DR-DOS & Win3.11:
Are you talking about using a system on an unsupported product? That *is* likely to cause problems.
Or are you talking about the beta's message box?
Name *one* proven instance when MS changed the OS interface to harm Netscape. Just one, and then prove it.
The problem isn't in the wrapping, it's that the states wants *all* of it removed.
You confuse modalurity with interchangibility.
Windows is highly modular, but parts aren't interchangable.
If you want to replace IE, you can, but you've to make an exact clone of IE.
And MS doesn't say that this is impossible, it's quite easy to do, (although can cause a support nightmare from hell).
MS says that *removing* IE will break a lot of stuff.
There shouldn't be a big fuss, if you want to replace IE, you need to re-implement the interfaces, and then you would change a couple of fields in the registry, and that would be *all*.
The problem is that MS can't support this configuration, and many applications may break if there is even just a subtle change in the way it works.
> They have the power to change the interfaces that other software uses, forcing the developers of that software to update their products to work with newer versions of windows
*Big time false*.
MS is *excellent* at backward compatability. If it worked on earlier version of Windows, than there is little chance that it would break on a newer version.
The only excepetion is beta API, which you shouldn't rely upon anyway.
No, it doesn't have to be Microsoft Dlls, and even now, you can replace it quite easily.
Let's assume that Mozilla correctly implements all of IWebBrowser interfaces.
What does it takes to replace IE with mozilla?
Change several keys in the registry, and that is all.
There *isn't* any other expanse to it.
If a Real wants to replace WMP, all they need is to implement all the publish interfaces, and register them on the published GUID.
That is all.
I forgot to mention, the HTML component can be removed, the problem is that a *lot* of programs are using it.
Including many MS programs.
You can have a Windows which is just the kernel & shell, that isn't very hard to do, the problem is that there are less than 1000 programs for NT proper.
And all of them are usually administration stuff.
Nothing that can be of any use.
Not a single Win32 program would run on it.
If you add Win32 subsystem, then you get much better support, but you still get a lot of missing stuff.
A *lot* of programs assume that the other components will be there. From the shell (All the SH* API) to IE (to Wininet.h function).
How do you upload a file programtically to FTP site on Linux?
Is there a standard way? What happen if you remove it? How many stuff would break?
MS offered to pulled IEXplorer & the icon, the judge wanted all of it put out of the system.
> Does the start bar suck?
t p://www.winstep.net/default.asp?cat=nextstartt p://momsoft.pair.com/momshl32.htm
. com/
/ ho mepage/facelift// vshell/t p.blarg.net/users/amol/tcsh. ac.uk/~skx/win/bash-203.zip
http://www.stardock.com/products/objectbar/
ht
ht
> Same for explorer
http://www.neoplanet.com/
http://www.smarteque
http://www.netcaptor.com/
> The command line shell
http://www.cs.purdue.edu/homes/chaudhis/pompous
http://www.vandyke.com/products
http://www.bayden.com/SlickRun/
ftp://f
http://www.tardis.ed
Anything else you desire?
Windows is *extemely* modular.