Microsoft is so big, they have no choice but to defend themselves in cases like this, otherwise they'd be deluged with a stupid number of bogus patent lawsuits.
Apache and PHP don't have lots of money.
eHarmony is small enough to give in, but probably have their hands on quite a bit of venture capital.
Epicrealm are following the time-honoured American tradition of only suing the people who is likely to give them lots of money.
The only real way to get the meassure would be analyzing slashdots log.
How many times do we have to go through this? Logs are worthless for counting browser market share. There are so many different things that skew the numbers you might as well make them up.
Example: How many people do you think go to the main page, click on a story, read it, click the back button, click on another story, read it, and so on? Given an equal number of Opera and Internet Explorer users, Internet Explorer users will be recorded in the logs as having read twice as many pages as Opera users, because Opera uses RFC 2616's history mechanism and Internet Explorer doesn't.
One of Google's products is an intranet appliance for "sifting through the piles of data stored inside organisations". This would put IBM in direct competition with them in that market. Public search isn't the only thing that Google does, you know.
This isn't a patent on dynamic page generation, but about a dynamic server farm where a primary "web server" distributes dynamic page generate requests to one or more "page servers", and where each page server can maintain a cached version of the output of the dynamic page request.
So it's basically dynamic page generation + Squid then? IIRC, this is exactly what Slashdot has been doing for years. You just have to understand that "page server" is their term for "caching proxy", and that it's been part of the HTTP specification since the mid-to-late 90s.
I don't mean to sound frustrated, but one reason most of You (the slashdot crowd, for example, not you in particular) don't associate "reasonable" answers with MS employees is that you often *disregard* our reasonable answers, or write them off, and just remember the times we say boneheaded things (keeping in mind that *all* people at *some* time will say something boneheaded).
There's a major difference between occasionally saying something stupid, and systematically disparaging your competition using the most pejorative, misleading statements you can think of at every available opportunity.
You might be reasonable. The majority of Microsoft employees might be reasonable. But when it comes to high-profile statements to the press, your top executives seem to all want to lie and cheat. That's why people don't associate Microsoft with being reasonable.
'VM', in the context of the Linux kernel, refers to its virtual memory manager, not virtual machines. It's incredibly misleading to read about 'stronger VM' like this.
Wasn't usermode Linux integrated into the 2.6 kernel anyway? What improvements in virtualisation is TFA referring to? It seemed remarkably short on details.
You should attend my LinuxWorld session this week J
For those that don't know, the letter 'J', in the WingDings font, is a smiley. This is utterly broken behaviour (because it's the character 'J', when an actual smiley character exists, U+263A that works no matter what font is used). This is usually caused by somebody pasting text written in Word into an HTML document or textarea.
And how many websites do you know that actually use this? And how many browsers even support it? AFAIK, multipart (AKA server-push) died with Netscape 4.
So when you said:
As long as web standards insist on the heavyweight request-response model...
...what you meant was:
As long as browsers don't implement the lightweight parts of the web standards...
It seems to me you are placing blame in the wrong place.
In any case, you can use remote scripting to retrieve a stream of data without having to poll for each update. Use an inline frame with sleep()s on the server. You can probably do this with XMLHttpRequest too, but I'm not certain of the cross-browser compatibility of partial results with that method. Of course, the few hundred bytes you save is an optimisation that rarely pays off.
They don't work well with the browser interface, breaking all sorts of different things, and they don't use standard controls. Not very usable.
Uneducated bullshit.
Is that so? I guess the fact that many browser controls simply stop working when faced with a Flash application is just a case of me being uneducated then? If I was smarter, the application would magically fix itself, would it? The things I routinely do in web applications would magically start working in Flash applications too if only I knew more?
Have an example: When I'm using my desktop computer, I use Firefox. This is configured to open links in a new window when I click on them with the middle mouse button. When I'm using my laptop, I use Safari. This is configured to open links in a new tab when I click on them with the middle mouse button. This works just fine in the web applications I use, and has never worked once for me with Flash applications. You are saying that my lack of knowledge broke my middle mouse button?
Flash apps are web apps more than any other technology out there.
Flash apps are web apps more than HTML+CSS+Javascript apps? In what sense?
Instead of verbose, unwieldy XML-based communication we now have fast, binary communication in Flash remoting,
That doesn't make it any more of a web application. If anything, the opposite is true. You call XML verbose an unwieldy, I call it easy to hack and consume. It's no coincidence most of the major Internet technologies - e.g. HTTP, SMTP, HTML, POP, IMAP - are all text based and easy to hack.
For the last few years there have also been standard interface components
They are standard to Flash. They are not standard to my browser or my operating environment. They work in different ways.
Once developed, they DO NOT require testing across browsers. Any browser that supports the plugin will behave the same.
I know from past experience that this is simply not true. Did you miss the bit where I mentioned Flash applications working in one version but not in newer versions? And what "the plugin" are you talking about? You do realise that there are multiple implementations, don't you? I've seen Flash applications work on Windows but break on Linux too.
Instead of addressing my statement you rambled on with some more nonsense like: "Flash is'nt installed everywhere"
Instead of addressing your statement? Your statement said it worked in all browsers. That's not true, so I pointed out it wasn't true. That was directly addressing your statement. Sounds to me like you are having a tantrum because I pointed out you were wrong.
Flash applications aren't web applications though. They are Flash applications that happen to be delivered through your web browser. They don't work well with the browser interface, breaking all sorts of different things, and they don't use standard controls. Not very usable.
Oh, and they work across all browsers, and don't require testing to make sure.
Wrong. Flash isn't installed everywhere, isn't available for all platforms, and I've seen Flash applications built for one version break in newer versions. It's just as much of a testing mess as, oh, I don't know, every other platform ever invented. Flash is certainly no magic bullet.
To my way of thinking, AJAX apps are an unmaintainable mess. You have none of the benefits of modern language design (modularity, OO structure)
Just because Javascript doesn't have 'class' as a keyword, it doesn't mean it's not OO. It's actually quite a nice OO language, in the style of Self rather than C++/Java.
your presentation layer is horribly intertwined with your program logic.
In what sense?
And the same code gets cut and pasted all over the show.
Bad developer, not bad language.
Oh, and most of the AJAX programmers I know write strictly for IE, which renders the whole discussion fairly moot anyway.
No, it means most of the AJAX programmers you know aren't very good or have their hands tied by management. No wonder you have a dim view of AJAX.
I think the only reason Java applets left a bad taste was because the platform was not mature when people started trying to use it for everything
No, the reason why Java applets left a bad taste is the same reason Flash does - it works against the web browser, not with it.
Javascript approaches (including "AJAX") work by modifying the current page. It keeps the same browser interface that people are comfortable with, and all their existing controls work. Think about it - what's the major complaint about badly-coded AJAX web applications? Breaking addressability (and therefore the back button, sending links to friends, etc).
Flash and Java, on the other hand, make very little attempt to work within the confines of the existing interface. Instead, they cut a hole in the current page, and replace it with a custom interface. A custom interface that breaks the existing interface. Does find-as-you-type work? Does middle-clicking to open in a new tab work? Does right-clicking to open in a new window work? There are so many different ways in which users' current interface outright breaks when using Flash and Java applets, that you might as well throw away the browser and ship a native executable.
The same thing could happen to AJAX and Web services - people pushing it before it's well-structured and complete enough to handle real-world application needs and the underlying Internet speed (read: 10-100Mbps to the home that's already there to the corporate desktop) is there.
I'm not sure why you are singling out speed here. AJAX is being used to speed up interfaces - it is an improvement, not an extra burden.
It's not yet certain that something with 2,000 screen forms and thousands of database tables and requiring thousands of procedures would work as well in a browser environment. I'm not saying it can't, I just haven't seen any proof yet.
The number of database tables and procedures is completely irrelevant, they aren't a UI issue. And as for the number of screen forms, why would a large number be any more difficult for the web than it is for traditional applications?
While Dr. Smith admits the image is about two years out of date, he also says he doesn't 'want to provide any easy assistance to anyone who wants to interfere with the site.' Citing the precedent of the blocks of colour over the White House and Treasury buildings, he's critical of their own security, adding 'there's a small area near the middle of the site which is quite secure, but the bulk of our site isn't all that secure' and is easily visible from the road and commercial airline flights.
If he doesn't want to "provide any easy assistance to anyone who wants to interfere with the site", then why is he publically pointing out the weak spots of their security?
Most of the web references to AJAX that I've seen correctly point out the importance of checking the browser version
Do not do this; it is broken.
If you want to use the XMLHttpRequest object, then the correct thing to do is to see if the XMLHttpRequest object is available. The browser version has nothing to do with that.
If you see me using Internet Explorer 6.0 and say to yourself "oh, I can use XMLHttpRequest because Internet Explorer is on my list of web browsers that support XMLHttpRequest", then your code is going to break when I have ActiveX switched off.
If you see me using FooBar 1.2.3 and say to yourself "oh, I can't use XMLHttpRequest because FooBar 1.2.3 isn't on my list of web browsers that support XMLHttpRequest because I've never heard of it", then your code isn't going to use XMLHttpRequest, even though it might be a simple shell on top of Gecko.
In both cases, you've done the wrong thing. The correct approach is to check to see if the object is available ("object detection" instead of "browser detection"), because that's what really matters. Browser detection is a fragile, archaic practice that competent developers stopped using back in the 90s. If you see it in a tutorial, then it's a good way of knowing that the tutorial is shit and you need to find a better one.
As long as web standards insist on the heavyweight request-response model, they will never achieve the snappiness, responsiveness, and flexibility that can be achieved with proper applications.
Er, AJAX is all about reducing that "heavyweight request-response model".
Now, imagine implementing the same sort of application in an environment where the only possible communication is you making an HTTP request and receiving an XML response.
Straw-man. HTTP isn't limited to a polling mechanism, e.g. multipart/replace.
AJAX doesn't make it easy to develop cross-platform web applications.
No, but it does improve the quality of web applications. The web is still a hideous interface compared with native applications, but AJAX is a significant step towards closing that gap.
Look at all the browser incompatibilities in the developing of Gmail and more recently MSN's start.com page.
That's more a case of the developers not being very good with Javascript than intrinsic difficulties. At least in the case of Google, I haven't looked at the start.com code.
We need to re-standarize Javascript or at least make sure all the browsers implement a 100% compatible version.
Already happened, it's ECMA-262.
The incompatibilities in client-side scripting aren't down to the language support; it's the objects provided by the host that are lacking, for example browser support for the various DOM specifications is pretty dire.
And i don't think that will work since not even HTML is properly rendered by any browser at all.
True, but at least we have a fairly reliable and useful subset of HTML that we can use to build stuff. CSS and Javascript also have a fairly reliable subset that work's across multiple browsers, the difference is that the subset is a lot smaller and therefore less useful. Browser vendors are slowly filling in the gaps in their support though, so this problem is being solved. 99% support isn't 100% support, but it's a lot more useful than 50% support.
According to the article, Microsoft is already creating a proprietary toolkit for AJAX.
No. According to the article, Microsoft is already creating a toolkit for AJAX. You added the proprietary bit, the article didn't say one way or the other.
Perhaps they hope their toolkit will become the standard.
You are turning the meaning of "standards-based" upside down. It means it uses standards, not that it is a standard. For example, it's based on the ECMA-262 standard. And it will be cross browser compatible.
So do you actually have anything of value to say, or does your comment boil down to a simple "yeah, but Microsoft are evil!" rant?
When I started using Linux in 1999, a lot of the programs were still using toolkits that predated Motif and were inferior to Windows 3.1.
Define "a lot of the programs". The KDE project was started in 1996 and had a usable 1.0 release with lots of accompanying applications in 1998. The GTK+ and Qt toolkits had been in use for years beforehand. I switched to Linux on the desktop full-time in 1998, and I think I used two or three programs that used Motif, it was definitely legacy at that point, so I don't know where you got the impression that even pre-Motif toolkits were widespread.
Now that Microsoft have cancelled all the planned new features for Vista, the only thing left for them to remove is stuff that works fine in current versions of Windows.
Absolutely they need a database to compete with local knowledge.
"Tesco" isn't some big brain residing at the centre of the UK dishing out stock. Tesco don't need to compete with local knowledge because they have local knowledge. The local store managers have it. The local store managers don't need a database to tell them what local people want to buy because they see it every day they go to work.
Slashdot Mirror
Hey, if the editors can dupe, why not you?
Dupe.
Microsoft is so big, they have no choice but to defend themselves in cases like this, otherwise they'd be deluged with a stupid number of bogus patent lawsuits.
Apache and PHP don't have lots of money.
eHarmony is small enough to give in, but probably have their hands on quite a bit of venture capital.
Epicrealm are following the time-honoured American tradition of only suing the people who is likely to give them lots of money.
How many times do we have to go through this? Logs are worthless for counting browser market share. There are so many different things that skew the numbers you might as well make them up.
Example: How many people do you think go to the main page, click on a story, read it, click the back button, click on another story, read it, and so on? Given an equal number of Opera and Internet Explorer users, Internet Explorer users will be recorded in the logs as having read twice as many pages as Opera users, because Opera uses RFC 2616's history mechanism and Internet Explorer doesn't.
One of Google's products is an intranet appliance for "sifting through the piles of data stored inside organisations". This would put IBM in direct competition with them in that market. Public search isn't the only thing that Google does, you know.
Don't be so negative!
So it's basically dynamic page generation + Squid then? IIRC, this is exactly what Slashdot has been doing for years. You just have to understand that "page server" is their term for "caching proxy", and that it's been part of the HTTP specification since the mid-to-late 90s.
There's a major difference between occasionally saying something stupid, and systematically disparaging your competition using the most pejorative, misleading statements you can think of at every available opportunity.
You might be reasonable. The majority of Microsoft employees might be reasonable. But when it comes to high-profile statements to the press, your top executives seem to all want to lie and cheat. That's why people don't associate Microsoft with being reasonable.
'VM', in the context of the Linux kernel, refers to its virtual memory manager, not virtual machines. It's incredibly misleading to read about 'stronger VM' like this.
Wasn't usermode Linux integrated into the 2.6 kernel anyway? What improvements in virtualisation is TFA referring to? It seemed remarkably short on details.
For those that don't know, the letter 'J', in the WingDings font, is a smiley. This is utterly broken behaviour (because it's the character 'J', when an actual smiley character exists, U+263A that works no matter what font is used). This is usually caused by somebody pasting text written in Word into an HTML document or textarea.
So when you said:
It seems to me you are placing blame in the wrong place.
In any case, you can use remote scripting to retrieve a stream of data without having to poll for each update. Use an inline frame with sleep()s on the server. You can probably do this with XMLHttpRequest too, but I'm not certain of the cross-browser compatibility of partial results with that method. Of course, the few hundred bytes you save is an optimisation that rarely pays off.
Is that so? I guess the fact that many browser controls simply stop working when faced with a Flash application is just a case of me being uneducated then? If I was smarter, the application would magically fix itself, would it? The things I routinely do in web applications would magically start working in Flash applications too if only I knew more?
Have an example: When I'm using my desktop computer, I use Firefox. This is configured to open links in a new window when I click on them with the middle mouse button. When I'm using my laptop, I use Safari. This is configured to open links in a new tab when I click on them with the middle mouse button. This works just fine in the web applications I use, and has never worked once for me with Flash applications. You are saying that my lack of knowledge broke my middle mouse button?
Flash apps are web apps more than HTML+CSS+Javascript apps? In what sense?
That doesn't make it any more of a web application. If anything, the opposite is true. You call XML verbose an unwieldy, I call it easy to hack and consume. It's no coincidence most of the major Internet technologies - e.g. HTTP, SMTP, HTML, POP, IMAP - are all text based and easy to hack.
They are standard to Flash. They are not standard to my browser or my operating environment. They work in different ways.
I know from past experience that this is simply not true. Did you miss the bit where I mentioned Flash applications working in one version but not in newer versions? And what "the plugin" are you talking about? You do realise that there are multiple implementations, don't you? I've seen Flash applications work on Windows but break on Linux too.
Instead of addressing your statement? Your statement said it worked in all browsers. That's not true, so I pointed out it wasn't true. That was directly addressing your statement. Sounds to me like you are having a tantrum because I pointed out you were wrong.
That's stupid web developers, not stupid AJAX. When you use AJAX properly, all the stuff you describe works just fine.
Flash applications aren't web applications though. They are Flash applications that happen to be delivered through your web browser. They don't work well with the browser interface, breaking all sorts of different things, and they don't use standard controls. Not very usable.
Wrong. Flash isn't installed everywhere, isn't available for all platforms, and I've seen Flash applications built for one version break in newer versions. It's just as much of a testing mess as, oh, I don't know, every other platform ever invented. Flash is certainly no magic bullet.
Just because Javascript doesn't have 'class' as a keyword, it doesn't mean it's not OO. It's actually quite a nice OO language, in the style of Self rather than C++/Java.
In what sense?
Bad developer, not bad language.
No, it means most of the AJAX programmers you know aren't very good or have their hands tied by management. No wonder you have a dim view of AJAX.
No, the reason why Java applets left a bad taste is the same reason Flash does - it works against the web browser, not with it.
Javascript approaches (including "AJAX") work by modifying the current page. It keeps the same browser interface that people are comfortable with, and all their existing controls work. Think about it - what's the major complaint about badly-coded AJAX web applications? Breaking addressability (and therefore the back button, sending links to friends, etc).
Flash and Java, on the other hand, make very little attempt to work within the confines of the existing interface. Instead, they cut a hole in the current page, and replace it with a custom interface. A custom interface that breaks the existing interface. Does find-as-you-type work? Does middle-clicking to open in a new tab work? Does right-clicking to open in a new window work? There are so many different ways in which users' current interface outright breaks when using Flash and Java applets, that you might as well throw away the browser and ship a native executable.
I'm not sure why you are singling out speed here. AJAX is being used to speed up interfaces - it is an improvement, not an extra burden.
The number of database tables and procedures is completely irrelevant, they aren't a UI issue. And as for the number of screen forms, why would a large number be any more difficult for the web than it is for traditional applications?
If he doesn't want to "provide any easy assistance to anyone who wants to interfere with the site", then why is he publically pointing out the weak spots of their security?
Schizophrenia has nothing to do with split personalities, you are thinking of disassociative identity disorder.
Do not do this; it is broken.
If you want to use the XMLHttpRequest object, then the correct thing to do is to see if the XMLHttpRequest object is available. The browser version has nothing to do with that.
If you see me using Internet Explorer 6.0 and say to yourself "oh, I can use XMLHttpRequest because Internet Explorer is on my list of web browsers that support XMLHttpRequest", then your code is going to break when I have ActiveX switched off.
If you see me using FooBar 1.2.3 and say to yourself "oh, I can't use XMLHttpRequest because FooBar 1.2.3 isn't on my list of web browsers that support XMLHttpRequest because I've never heard of it", then your code isn't going to use XMLHttpRequest, even though it might be a simple shell on top of Gecko.
In both cases, you've done the wrong thing. The correct approach is to check to see if the object is available ("object detection" instead of "browser detection"), because that's what really matters. Browser detection is a fragile, archaic practice that competent developers stopped using back in the 90s. If you see it in a tutorial, then it's a good way of knowing that the tutorial is shit and you need to find a better one.
Er, AJAX is all about reducing that "heavyweight request-response model".
Straw-man. HTTP isn't limited to a polling mechanism, e.g. multipart/replace.
No, but it does improve the quality of web applications. The web is still a hideous interface compared with native applications, but AJAX is a significant step towards closing that gap.
That's more a case of the developers not being very good with Javascript than intrinsic difficulties. At least in the case of Google, I haven't looked at the start.com code.
Already happened, it's ECMA-262.
The incompatibilities in client-side scripting aren't down to the language support; it's the objects provided by the host that are lacking, for example browser support for the various DOM specifications is pretty dire.
True, but at least we have a fairly reliable and useful subset of HTML that we can use to build stuff. CSS and Javascript also have a fairly reliable subset that work's across multiple browsers, the difference is that the subset is a lot smaller and therefore less useful. Browser vendors are slowly filling in the gaps in their support though, so this problem is being solved. 99% support isn't 100% support, but it's a lot more useful than 50% support.
Oh, please. Enough with the FUD.
No. According to the article, Microsoft is already creating a toolkit for AJAX. You added the proprietary bit, the article didn't say one way or the other.
You are turning the meaning of "standards-based" upside down. It means it uses standards, not that it is a standard. For example, it's based on the ECMA-262 standard. And it will be cross browser compatible.
So do you actually have anything of value to say, or does your comment boil down to a simple "yeah, but Microsoft are evil!" rant?
Define "a lot of the programs". The KDE project was started in 1996 and had a usable 1.0 release with lots of accompanying applications in 1998. The GTK+ and Qt toolkits had been in use for years beforehand. I switched to Linux on the desktop full-time in 1998, and I think I used two or three programs that used Motif, it was definitely legacy at that point, so I don't know where you got the impression that even pre-Motif toolkits were widespread.
Now that Microsoft have cancelled all the planned new features for Vista, the only thing left for them to remove is stuff that works fine in current versions of Windows.
"Tesco" isn't some big brain residing at the centre of the UK dishing out stock. Tesco don't need to compete with local knowledge because they have local knowledge. The local store managers have it. The local store managers don't need a database to tell them what local people want to buy because they see it every day they go to work.