I can't even begin to tell you the number of sites required by my previous employer that required IE, and there's always a couple here and there that want ActiveX or what have you.
It's 2010. Can you list the sites here, and I'm sure someone will "evangelise" them into updating!:D
Seriously though - please do list them!
The only reason we send XHTML as text/html at all is because a certain browser doesn't understand application/xml+xhtml
..although I believe Firefox will not try to render a document served as application/xml+xhtml *until* the entire thing has loaded. That said, I can understand why it would do that if it was just "application/xml" but I would've though the +xhtml bit should trigger it to be a bit smarter! so maybe this is just a Firefox issue. And also... this is something I vaguely remember reading, so I might be talking crap and/or it might've changed in recent releases!
There's a fair bit of benefit if you can produce output [say from a server with access to.. I dunno.. weather data for example] in an XML format since that data can then be used by something else. If that format is XHTML, then you get the benefit that it can both be viewed on a browser and is machine readable. Obviously this works better now because we've separated presentation from markup (mostly!) via CSS.
Yes... I could have my server serve up HTML to browsers *and* provide an API that serves XML, and that isn't necessarily a bad idea, but being able to serve just XHTML is a good low-cost, easy to maintain, option that can be run with nothing more than a text editor.
Also... Hixie likes to bleat about stuff. I wouldn't pay *too* much attention to him!:P
I did hear that the attack was part of a targeted phishing (so called spear-phishing) campaign; that being the case, it might've been a Google bean-counter's home computer? I know that kind of seems a stretch, but I can't see Google using IE6 interally *at all* except by web-devs for browser testing, and those folks are a lot less likely to have pointed IE6 at external websites.
Also, I'd imagine that most people at Google are technically literate, but I'm guessing the accounts dept, is like accounts departments everywhere.... so you can't count [no pun intended] on them not to do something dumb.
AFAIK IE7 also changed the security setup such that legacy ActiveX controls may not work? Also, IE7 doesn't run on Windows 2000.... which is less an issue today since support for that OS is about to end, but it meant there was no MS supported upgrade path.
The UI changes aren't such a biggie, but corporates do bleat about any changes since it affects training.
So i'm not seeing what such a IE6.1 would've accomplished.
My point is that whilst IE6 didn't support newer CSS features (or PNGs) properly, and whilst it used a different box model, it also had huge bugs. The former problems are "features" and I wouldn't expect MS to change them, but the bugs were clearly bugs. They were just less apparent in 2001 because most websites used tables - not CSS - for layout.
As a web-dev, I see the biggest problem with IE6 being that it's rendering is massively buggy [the security issues don't directly affect me!]. Issues such as peek-a-boo bugs are difficult to test for; I can build a site, test it and have it all work perfectly, and then make a few tweaks later, and than discover that this triggers a weird IE6 bug. Worse, even if I have "tested"** in IE6, these things don't always reveal themselves... so in practise, I tend to find out about issues when the client complains!
IE7 introduced a bunch of changes including the obvious addition of tabs and UI changes, and the fixes and updates to the rendering engine.
IMHO, Microsoft should have fixed the rendering bugs in IE6. They should've called that IE6.1 or something. But instead, they did their usual trick where they roll bug fixes and features into one, so some people cannot easily update from IE6 to IE7. But IE6 rendering was broken and should've been fixed waaaaay before IE7 was released.
On the other-hand, if MS had done this, we would have a less flaky, but still seriously lacking, IE6 and there may be even less incentive for people to upgrade!
**"Tested" for me typically means I viewed a page once in said browser. Obviously if the budget is there then more testing may be done, but typically there isn't time for extensive testing for minor page changes.
Anyone know if the iPad can be used *without* a computer? My iPod Touch still needs a computer running iTunes when I want to update the system software... which is a pain in the arse really!
Agreed - and this makes it pointless in the first place. They'll spend a bunch of time and a chunk of cash implementing this encryption, and more money maintaining it (licensing OEMs to build decoders, etc), all of which will be passed on to the consumer.
The thing that annoys me most though is that whilst the encryption will make no practical difference to most people here since we'll know how to get around it, it makes Linux/FOSS a second class citizen to proprietry solutions. For example, a machine pre-installed with Ubuntu will likely not be able to view this DRM'd content unless someone pays for licensing.... assuming they can do that at all.
And all for what -- this makes it *very slightly* more difficult to copy content, which was illegal anyway, with ABSOLUTELY NO practical benefit to anyone.
Could they unleash it immediately on MS? Since MS has been talking about setting up server farms and managing workloads in the last year or so, and since it sounds pretty much like they're trying to do what Google do, maybe MS are the intended target.
Re flash stuttering, up until recently I was using an old Dell Dimension 3000 with a 2.8GHz Celeron + nVidia 6200 (PCI version!) and that did stutter, I was able to get around it by using the Zoom In feature of Compiz to effectively get full-screen flash.
I'm now runing a Core i5 based machine with an nVidia 9800 and flash now runs perfectly.
For the record though, I think the problem is that the flash plugin requires a ridiculous amount of CPU power to work!!
Except that Google Voice transcriptions are done server-side and don't have to be done real-time. Compared to that the phone is actually much less capable in this regard.
Are you speaking authoritatively here? I ask as the Engadget Nexus One voice recognition video kind of looks like the processing is being done locally, and obviously in this case is would need to be performed in as real-time-as-possible in order to be useful. I guess it could be that the phone pre-processes the voice before consulting a database run by Google... but I'm speculating!
...then I guess we should let music die. Music and other entertainment is not important enough by far to trade away privacy and freedom. I don't care for piracy, but I recognize that only by having complete control of what people communicate and hence their freedom of expression would it be possible to quell piracy. I hope most thinking humans would agree that this is too high a price to preserve the profitability of music.
Yeah, and he said himself that it only accounts for nearly 4 percent of GDP, so it really isnt' worth loss of freedom.
Also, at least he focused on the monitary loss rather than, I dunno, creative/cultural loss, 'cos then I might've thought he actually cared about his output!
Performance-wise, they're both blazing fast even on a modest single-core/1GB system.
Just to pick up on this point... it might be something funny with Ubuntu 9.10, or maybe I've been using Chromium a bit too much, but Firefox just seems very sluggish these days. I've recently upgraded my main computer from a 2.8GHz Celeron (2GB RAM) to a 2.6Ghz Core i5 (4GB RAM), and Firefox still takes a few seconds to start up! It's true I do run a number of extensions, and I do have a fairly rediculous number of bookmarks in FF, but even on a fresh OS install with vanilla FF, it's still not fast.
I'm not knocking Firefox btw -- I love it to bits for all it's extensions -- but the sluggishness (both startup, but also generally UI-laggyness) make it harder to love these days. So FF is kind of my web-dev/testing browser whilst Chromium is my day to day jumping around websites browser.
Standards like HTML5 and CSS3 are/will-be based on existing implementations -- these standards are never designed by committee and *then* implemented! We need working implementations that we can then refine if they aren't working as anticipated, so the implementations that exist in Firefox/Chrome/Opera are all potentially subject to change. But that's fine, because (1). if you're a web-developer and you absolutely don't want to risk having to change your code, then stick with older standards, and (2). in practise, they're probably not going to change much if at all and that means we can take advantage.
I'd rather have "release when it's done" and not "the public will do our QA for us".
What public? The public just use browsers to access websites. Going on what you said in your GP post, you seem to mixing up "beta" software with currently implementations of HTML5 and CSS3, but that is incorrect. Beta software is software that may contain bugs, so should not be used for mission critical tasks. However, the HTML5/CSS3 implementations in Firefox/Chrome/Opera are not "test" versions... they're solid implementations of the currently unfinalised standards.
As a user, I am not doing QA for anyone by using any of these browsers.
As a web-developer, if I choose to use HTML5 and/or CSS3, then I do so on the understanding that since these standards are not final, I may need to update my code in the future. But I don't have to do this -- I can stick with older standards if I so choose.
And I'd certainly rather have the web browser whose icon I just double clicked, rather than a rogue web browser injected into the UI of what I thought I was using.
It's hardly "rogue" if you've installed it is it? No one is sneaking code onto your computer!
However, if you want to use modern web apps such as Google Wave, whilst using the familiar IE interface, then the only way to do this is using the Chrome Frame plugin.
Alternatively, you could download Firefox/Chrome/Opera/whatever separately. But IE does not support the technologies required to make Google Wave (+ other apps) possible.
From Google's perspective, if they did not require HTML5/CSS3 (or SVG/Canvas or whatever it is they require) for Wave, then they would've had to use something like Flash or Silverlight.... both of which are browser plugins, same as Chrome Frame. Except of-course that Flash/Silverlight are proprietary. Oh, and not based on standards -- even unfinalised ones!
Your wording is curiously similar to the GP -- are you both related?!
Anyway, AFAIK the reason Google is pushing Chrome Frame is so that _even_ IE users can use Google WebApps. Obviously IE6 (and IE7 to a lesser extent) needs some "help" because it is so far behind all other browsers; note, this isn't a dig at IE6/7 -- they're old, so it's to be expected. But unlike other old browsers, these are still being used.
But even with IE8, the JavaScript runs a bit slow compared with everything else, there are still some CSS issues (albeit far far less then previous IE versions), and it does not support SVG or Canvas. These latter two are kind of useful when you want to build webapps that are more like current desktop apps without resorting to something proprietary like Adobe Flash or SliverLight.
HTML5 is a good example of this difference of philosophy, and certainly so is this Chrome Frame plugin which is essentially a sloppy man-in-the-middle attack vector. It's like one of those obnoxious browser toolbars that acts as an intermediary to hijack all your search queries.
If MS showed some interest in implementing Canvas/SVG, I might buy your argument. But the reason they have no interest in this is because open technologies such as these threaten Microsofts business model and do not allow them to control the market... what with being open an'all!
If MS were to implement HTML5 / CSS3 features now, that would not make the software beta. It would mean that some of the features are subject to change, so anyone using them, might have to tweak their websites in the future... but that's okay, 'cos we're all grown up developers and we understand this. But having working implementations out in the wild is what standards development is all about. This is how we discover if what _seems_ like a good idea on paper, really is a good idea or not.
As for your comments "sloppy man-in-the-middle attack vector" and "hijack all your search queries", I'm not sure where you're getting your information from, but those do sound remarkably like FUD.
I find I need it on some sites where there is flash video (that I want to watch) and also a flash advert, because any flash uses so much of my CPU, with two flash objects running, the flash video craps out.
Oh jeebus. Building a missile, bomb or anything that kills people is NOT HARD. I can get the relevant documents needed for anyone with a mild training in electronics to build a guidance system for a missile or a homing system for a rocket.
But what if I want a missile that has it's own blog? Or a bomb that tweets before it explodes? See, you *need* open source!
Works fine for me, but maybe worth noting that the current stable Ubuntu (9.04) uses OOo 3.0.1 so for the parent to be using 3.1, then they must be running something that isn't "stable"... I think!
How does this fit in with Debian now also providing a FreeBSD kernel? I'm assuming that the FreeBSD kernel uses a BSD licence... but I might well be wrong... but assuming it does, I would guess GCD would be fine?
Slashdot Mirror
I can't even begin to tell you the number of sites required by my previous employer that required IE, and there's always a couple here and there that want ActiveX or what have you.
It's 2010. Can you list the sites here, and I'm sure someone will "evangelise" them into updating! :D
Seriously though - please do list them!
The only reason we send XHTML as text/html at all is because a certain browser doesn't understand application/xml+xhtml
..although I believe Firefox will not try to render a document served as application/xml+xhtml *until* the entire thing has loaded. That said, I can understand why it would do that if it was just "application/xml" but I would've though the +xhtml bit should trigger it to be a bit smarter! so maybe this is just a Firefox issue. And also... this is something I vaguely remember reading, so I might be talking crap and/or it might've changed in recent releases!
There's a fair bit of benefit if you can produce output [say from a server with access to.. I dunno.. weather data for example] in an XML format since that data can then be used by something else. If that format is XHTML, then you get the benefit that it can both be viewed on a browser and is machine readable. Obviously this works better now because we've separated presentation from markup (mostly!) via CSS.
:P
Yes... I could have my server serve up HTML to browsers *and* provide an API that serves XML, and that isn't necessarily a bad idea, but being able to serve just XHTML is a good low-cost, easy to maintain, option that can be run with nothing more than a text editor.
Also... Hixie likes to bleat about stuff. I wouldn't pay *too* much attention to him!
I did hear that the attack was part of a targeted phishing (so called spear-phishing) campaign; that being the case, it might've been a Google bean-counter's home computer? I know that kind of seems a stretch, but I can't see Google using IE6 interally *at all* except by web-devs for browser testing, and those folks are a lot less likely to have pointed IE6 at external websites.
Also, I'd imagine that most people at Google are technically literate, but I'm guessing the accounts dept, is like accounts departments everywhere.... so you can't count [no pun intended] on them not to do something dumb.
The UI changes aren't such a biggie, but corporates do bleat about any changes since it affects training.
So i'm not seeing what such a IE6.1 would've accomplished.
My point is that whilst IE6 didn't support newer CSS features (or PNGs) properly, and whilst it used a different box model, it also had huge bugs. The former problems are "features" and I wouldn't expect MS to change them, but the bugs were clearly bugs. They were just less apparent in 2001 because most websites used tables - not CSS - for layout.
As a web-dev, I see the biggest problem with IE6 being that it's rendering is massively buggy [the security issues don't directly affect me!]. Issues such as peek-a-boo bugs are difficult to test for; I can build a site, test it and have it all work perfectly, and then make a few tweaks later, and than discover that this triggers a weird IE6 bug. Worse, even if I have "tested"** in IE6, these things don't always reveal themselves... so in practise, I tend to find out about issues when the client complains!
IE7 introduced a bunch of changes including the obvious addition of tabs and UI changes, and the fixes and updates to the rendering engine.
IMHO, Microsoft should have fixed the rendering bugs in IE6. They should've called that IE6.1 or something. But instead, they did their usual trick where they roll bug fixes and features into one, so some people cannot easily update from IE6 to IE7. But IE6 rendering was broken and should've been fixed waaaaay before IE7 was released.
On the other-hand, if MS had done this, we would have a less flaky, but still seriously lacking, IE6 and there may be even less incentive for people to upgrade!
**"Tested" for me typically means I viewed a page once in said browser. Obviously if the budget is there then more testing may be done, but typically there isn't time for extensive testing for minor page changes.
Anyone know if the iPad can be used *without* a computer? My iPod Touch still needs a computer running iTunes when I want to update the system software... which is a pain in the arse really!
Agreed - and this makes it pointless in the first place. They'll spend a bunch of time and a chunk of cash implementing this encryption, and more money maintaining it (licensing OEMs to build decoders, etc), all of which will be passed on to the consumer.
The thing that annoys me most though is that whilst the encryption will make no practical difference to most people here since we'll know how to get around it, it makes Linux/FOSS a second class citizen to proprietry solutions. For example, a machine pre-installed with Ubuntu will likely not be able to view this DRM'd content unless someone pays for licensing.... assuming they can do that at all.
And all for what -- this makes it *very slightly* more difficult to copy content, which was illegal anyway, with ABSOLUTELY NO practical benefit to anyone.
Could they unleash it immediately on MS? Since MS has been talking about setting up server farms and managing workloads in the last year or so, and since it sounds pretty much like they're trying to do what Google do, maybe MS are the intended target.
Re flash stuttering, up until recently I was using an old Dell Dimension 3000 with a 2.8GHz Celeron + nVidia 6200 (PCI version!) and that did stutter, I was able to get around it by using the Zoom In feature of Compiz to effectively get full-screen flash.
I'm now runing a Core i5 based machine with an nVidia 9800 and flash now runs perfectly.
For the record though, I think the problem is that the flash plugin requires a ridiculous amount of CPU power to work!!
Or maybe use Chrome Frame so it's invisible to the users?
Although when you compare Ballmer with most bank execs, the guy looks like a superstar!
Except that Google Voice transcriptions are done server-side and don't have to be done real-time. Compared to that the phone is actually much less capable in this regard.
Are you speaking authoritatively here? I ask as the Engadget Nexus One voice recognition video kind of looks like the processing is being done locally, and obviously in this case is would need to be performed in as real-time-as-possible in order to be useful. I guess it could be that the phone pre-processes the voice before consulting a database run by Google... but I'm speculating!
Valid point, but it's unlikely Beckham will write an article in the NYT about it!! :D
...then I guess we should let music die. Music and other entertainment is not important enough by far to trade away privacy and freedom. I don't care for piracy, but I recognize that only by having complete control of what people communicate and hence their freedom of expression would it be possible to quell piracy. I hope most thinking humans would agree that this is too high a price to preserve the profitability of music.
Yeah, and he said himself that it only accounts for nearly 4 percent of GDP, so it really isnt' worth loss of freedom.
Also, at least he focused on the monitary loss rather than, I dunno, creative/cultural loss, 'cos then I might've thought he actually cared about his output!
Performance-wise, they're both blazing fast even on a modest single-core/1GB system.
Just to pick up on this point... it might be something funny with Ubuntu 9.10, or maybe I've been using Chromium a bit too much, but Firefox just seems very sluggish these days. I've recently upgraded my main computer from a 2.8GHz Celeron (2GB RAM) to a 2.6Ghz Core i5 (4GB RAM), and Firefox still takes a few seconds to start up! It's true I do run a number of extensions, and I do have a fairly rediculous number of bookmarks in FF, but even on a fresh OS install with vanilla FF, it's still not fast.
I'm not knocking Firefox btw -- I love it to bits for all it's extensions -- but the sluggishness (both startup, but also generally UI-laggyness) make it harder to love these days. So FF is kind of my web-dev/testing browser whilst Chromium is my day to day jumping around websites browser.
/me stirs berylium ions and tastes...
/me announces "needs more cats!"
Good call -- far too subtle though!! ;)
I'd rather have "release when it's done" and not "the public will do our QA for us".
What public? The public just use browsers to access websites. Going on what you said in your GP post, you seem to mixing up "beta" software with currently implementations of HTML5 and CSS3, but that is incorrect. Beta software is software that may contain bugs, so should not be used for mission critical tasks. However, the HTML5/CSS3 implementations in Firefox/Chrome/Opera are not "test" versions... they're solid implementations of the currently unfinalised standards.
As a user, I am not doing QA for anyone by using any of these browsers.
As a web-developer, if I choose to use HTML5 and/or CSS3, then I do so on the understanding that since these standards are not final, I may need to update my code in the future. But I don't have to do this -- I can stick with older standards if I so choose.
And I'd certainly rather have the web browser whose icon I just double clicked, rather than a rogue web browser injected into the UI of what I thought I was using.
It's hardly "rogue" if you've installed it is it? No one is sneaking code onto your computer!
However, if you want to use modern web apps such as Google Wave, whilst using the familiar IE interface, then the only way to do this is using the Chrome Frame plugin.
Alternatively, you could download Firefox/Chrome/Opera/whatever separately. But IE does not support the technologies required to make Google Wave (+ other apps) possible.
From Google's perspective, if they did not require HTML5/CSS3 (or SVG/Canvas or whatever it is they require) for Wave, then they would've had to use something like Flash or Silverlight.... both of which are browser plugins, same as Chrome Frame. Except of-course that Flash/Silverlight are proprietary. Oh, and not based on standards -- even unfinalised ones!
Anyway, AFAIK the reason Google is pushing Chrome Frame is so that _even_ IE users can use Google WebApps. Obviously IE6 (and IE7 to a lesser extent) needs some "help" because it is so far behind all other browsers; note, this isn't a dig at IE6/7 -- they're old, so it's to be expected. But unlike other old browsers, these are still being used.
But even with IE8, the JavaScript runs a bit slow compared with everything else, there are still some CSS issues (albeit far far less then previous IE versions), and it does not support SVG or Canvas. These latter two are kind of useful when you want to build webapps that are more like current desktop apps without resorting to something proprietary like Adobe Flash or SliverLight.
HTML5 is a good example of this difference of philosophy, and certainly so is this Chrome Frame plugin which is essentially a sloppy man-in-the-middle attack vector. It's like one of those obnoxious browser toolbars that acts as an intermediary to hijack all your search queries.
If MS showed some interest in implementing Canvas/SVG, I might buy your argument. But the reason they have no interest in this is because open technologies such as these threaten Microsofts business model and do not allow them to control the market... what with being open an'all!
If MS were to implement HTML5 / CSS3 features now, that would not make the software beta. It would mean that some of the features are subject to change, so anyone using them, might have to tweak their websites in the future... but that's okay, 'cos we're all grown up developers and we understand this. But having working implementations out in the wild is what standards development is all about. This is how we discover if what _seems_ like a good idea on paper, really is a good idea or not.
As for your comments "sloppy man-in-the-middle attack vector" and "hijack all your search queries", I'm not sure where you're getting your information from, but those do sound remarkably like FUD.
I find I need it on some sites where there is flash video (that I want to watch) and also a flash advert, because any flash uses so much of my CPU, with two flash objects running, the flash video craps out.
Oh jeebus. Building a missile, bomb or anything that kills people is NOT HARD. I can get the relevant documents needed for anyone with a mild training in electronics to build a guidance system for a missile or a homing system for a rocket.
But what if I want a missile that has it's own blog? Or a bomb that tweets before it explodes? See, you *need* open source!
In the same way that BadAnalogyGuy is the bad analogy of the bad analogy world?!
Works fine for me, but maybe worth noting that the current stable Ubuntu (9.04) uses OOo 3.0.1 so for the parent to be using 3.1, then they must be running something that isn't "stable"... I think!
How does this fit in with Debian now also providing a FreeBSD kernel? I'm assuming that the FreeBSD kernel uses a BSD licence... but I might well be wrong... but assuming it does, I would guess GCD would be fine?