Of course, one must realize that you can't just change JSON. Changes to JSON will suffer the same fate as technologies like E4X. For server-side things this change could happen, but one would lose the benefit of JSON: a single, simple object representation.
"Don't you remember Frontpage spew with missing things like... BODY tags, DOCTYPES, and other very imporant parts!"
Yeah, Frontpage sucks. That doesn't mean the IE team enjoyed writing the support for a shitty product they didn't make. Frontpage should have been more like a contentEditable or Midas application but it isn't.
"How can you render a document if you don't even know what the type of the document is intended to be?"
Well, there's the content type it's sent as via the server for one. That seems to determine it for most browsers pretty well. Microsoft does go an extra mile to scan for tags to see if it should present the page as a text/html page instead of a text/plain page (which is a choice I don't agree with personally).
What damning evidence do you have that says that they deliberately created many of those pages in order to break their competition? Embrace and extend wasn't about creating shitty documents, it was about creating proprietary features that customers couldn't live without (like marquee, button, and iframe not to mention several CSS properties). Sometimes they work, sometimes they fail.
One thing that Microsoft does very well is backwards compatibility. IE was created before the standards that they violate existed, and as such Microsoft had to figure out a good way to accomplish two goals: backwards compatibility for their existing implementations (however shoddy they were) and a way to allow for future development to happen. They did this (very well in my opinion) with IE 6 and the DOCTYPE switching.
As for the "doesn't even support PNG correctly" which version of PNG are you referring to? Obviously it's the useful one that everyone would love to use (that works pretty well in IE 7 I think) but as for following the PNG standard they _did_. Alpha channels aren't required to be an implementation of PNG. Indexed PNGs work wonderfully in IE. 24-bit color pngs with 8-bit alpha only so-so. The alpha channel works only against a background chunk which most editors leave out (google tweakpng for a useful editor that allows you to edit that chunk). You can use alpha channel PNGs if you use them in img elements with a DirectX filter, which is well-documented. Alpha PNGs as backgrounds work now in IE 7 as well.
I never said that IE wasn't junk, I certainly avoid it. However, before one complains about the work of others, it's often useful to understand the situation they were in. IE 5 was amazing at the time, even though it sold Internet Explorer down the river to supporting crappy markup and such. Some of the choices the IE team made in the mid-90s locked Microsoft in to some design decisions they are struggling to overcome.
If you want to blame people for the adoption, blame AOL, Dell, Gateway, etc. who could have shipped Opera just as easily as the default (except the price, so maybe you should blame capitalism while you're at it).
Regardless of the author's web developing skills, it's a valid metric for evaluating a web browser.
No matter how well Firefox and Opera employ W3C standards, they still need to be able to display poorly created pages just as well as valid, semantic, XHTML-driven sites.
Yes, there are a lot of people who make a lot of workarounds for a lot of browsers. Those who lament this fact should get over it. The companies involved know damn well by now what business they're involved with. Folks have got to stop belly-aching and bitching over these now decade-old problems. They're well-defined problems, which is a good thing. It takes some tricky work to keep your backwards compatibility and introduce new ways of working, ala Internet Explorer's DOCTYPE mode. If they are concerned about people introducing hokey work-arounds that they would eventually have to work around themselves, browser makers would do well to be more involved with the design community.
Okay, I think I understand. Being able to refer to another elements compiled rules value would be sorta useful, though pretty far out there I think. I can already picture the stock price of Tylenol going up for the developers implementing that;)
I think you hit on the reason for no variables (the comma rule separator). While we're on the topic of reusable values, I've always wanted to have an external color table for PNGs and CSS so I could change my color scheme across the entire theme of my project without regenerating my image files (I can sort of accomplish this with SVG, but there are so many headaches with using SVG across browsers I don't think it's worthwhile; let alone use SVG for a background image).
I'm not sure if you're just being sarcastic or not, but I'll give you the benefit of the doubt.
With regard to your first question, "arbitrary values of arbitrary tags" are you referring to selecting an element with a particular attribute set to a particular value? If you are, CSS 2.1 already supports that input[type="submit"] for example will select a submit button. (of course support isn't 100% there, but it's rather neat)
As for variables, I suppose you could use XSLT to generate your style rules, but the use of XSLT is more geared towards transforming the content of the page, not making it pretty... I'm curious what use client-side variables would be without the support of conditional statements, care to expound on your interest in CSS variables?
My question has to deal with selectors: why doesn't CSS have a parent selector? All too often I find myself wanting to refer to a parent or a sibling of the current element but cannot (I'm sure there are a bevy of workarounds for individual cases...)
ie Why can't I do something like this to refer to a table row (given a structure of <tr><td><a>) a:hover:parent:parent { background-color: #cff; }? It's a simple example, but there really isn't a way to do this without involving a client-side scripting language.
Actually, most tags can be closed in that way in HTML without problems. The script tag for some reason demands itself to be closed with an explicit end tag. img, object, br, hr all work fine with the XHTML style in HTML (they even validate). meta and link require exactly no end tag in HTML, and are invalid if you add it in.
If it was FreeBSD or OS X, you could have tried single-user mode. I assume most Linux distros have an option for that as well, but I could be wrong. Once in single-user mode, you can mount the file system as normal and edit the config file.
Like many things, it's not evil in itself, but allows for the potential for evil. From Theo's post
Their X servers also run entirely as root, while ours is now privilege seperated and running jailed as user _x11. Even so, our privilege seperated X server is talking directly
to the IO registers of a video card with much evil in it. And many newer video cards are very smart, capable, and thus dangerous. So we have concerns.
The main concern seems to be around allowing powerful, closed pieces of hardware connected to a system through a root user connection. There likely wouldn't be a cause for concern, but technically could be a source of trouble.
Some of those names existed before the 'net came about, so searching wasn't a consideration for those product names. (think Lotus Notes, if you search for Notes, you'll be SOL)
Publisher is a poor program with a perfect name: an untrained user can probably guess that Publisher is used for some sort of desktop publishing.
And of course, "Sequel" isn't a product any more than "SQL" is a product that Microsoft sells. SQL Server (which is a perfectly cromulent name to search for).
Are you sure you're not thinking of Adobe here? LiveMotion for instance? The only reason I got it was cuz I bought their Web Designer bundle 5 years ago.
I agree with your assessment of Real (especially the basic player v. 10--quite nice) but Windows Media Player for Mac has a ton of faults that aren't worth overlooking.
The first is that it was never updated, so all outstanding bugs remain outstanding.
The second is it works quite poorly as a plugin in web browsers. There are many occasions where WMP just disappears for no reason other than it's a crappy product.
The third annoyance I've got with it is that you can't properly fast-forward or rewind. At least those controls work well within Quicktime.
From the sound of it (complaining about nagware) you're not using Quicktime 7. Give it a shot.
Two browsers that have been (in my mind) at the forefront of offering accessibility features are iCab and Opera. However, nearly every screen reader works with Internet Explorer. Is there some future hope of seeing screen reader technology in a future version of Opera?
How about "Arbitrary Collection Manager" instead? Just cuz you're a typography-geek doesn't mean everyone thinks of that when they hear "type." Ever hear of a "file-type?" Well, that's what's being described here, although the things don't have to explicitly be files.
Here's the thing: it's still good business to build for IE first.
Before you all get fired-up and start yelling "standards" you gotta realize that the web doesn't quite offer up the promise of Java (write once, run anywhere).
You start with a restricted featureset, and build a product based on that. That gets you money. This money can then be used to accomplish your dream (I dunno, supporting some weird beta of a Plan 9 only browser maybe).
You put up with restrictions because it's not finished yet. It's similar to asking why do we put up with bugs when we're running something we paid for? You just do.
Well, this should be the first time an actual standard will be created for the web then.
I know what you're thinking, "the W3C! they make standards!" Well, they don't. They recommend, but don't actually create any standards at all. Realistically speaking, standards are what you find in the wild (especially in the case of the web). mTLD seems to be doing what nobody else ever had the cajones to do online: enforce a standard on people. I'm quite curious to see how well it goes over and how that translates over to other TLDs.
I'm especially curious to see if mTLD enforces proper usage of MIME types so XHTML might actually be served as an XML application.
Slashdot Mirror
Of course, one must realize that you can't just change JSON. Changes to JSON will suffer the same fate as technologies like E4X. For server-side things this change could happen, but one would lose the benefit of JSON: a single, simple object representation.
"Don't you remember Frontpage spew with missing things like... BODY tags, DOCTYPES, and other very imporant parts!"
Yeah, Frontpage sucks. That doesn't mean the IE team enjoyed writing the support for a shitty product they didn't make. Frontpage should have been more like a contentEditable or Midas application but it isn't.
"How can you render a document if you don't even know what the type of the document is intended to be?"
Well, there's the content type it's sent as via the server for one. That seems to determine it for most browsers pretty well. Microsoft does go an extra mile to scan for tags to see if it should present the page as a text/html page instead of a text/plain page (which is a choice I don't agree with personally).
What damning evidence do you have that says that they deliberately created many of those pages in order to break their competition? Embrace and extend wasn't about creating shitty documents, it was about creating proprietary features that customers couldn't live without (like marquee, button, and iframe not to mention several CSS properties). Sometimes they work, sometimes they fail.
One thing that Microsoft does very well is backwards compatibility. IE was created before the standards that they violate existed, and as such Microsoft had to figure out a good way to accomplish two goals: backwards compatibility for their existing implementations (however shoddy they were) and a way to allow for future development to happen. They did this (very well in my opinion) with IE 6 and the DOCTYPE switching.
As for the "doesn't even support PNG correctly" which version of PNG are you referring to? Obviously it's the useful one that everyone would love to use (that works pretty well in IE 7 I think) but as for following the PNG standard they _did_. Alpha channels aren't required to be an implementation of PNG. Indexed PNGs work wonderfully in IE. 24-bit color pngs with 8-bit alpha only so-so. The alpha channel works only against a background chunk which most editors leave out (google tweakpng for a useful editor that allows you to edit that chunk). You can use alpha channel PNGs if you use them in img elements with a DirectX filter, which is well-documented. Alpha PNGs as backgrounds work now in IE 7 as well.
I never said that IE wasn't junk, I certainly avoid it. However, before one complains about the work of others, it's often useful to understand the situation they were in. IE 5 was amazing at the time, even though it sold Internet Explorer down the river to supporting crappy markup and such. Some of the choices the IE team made in the mid-90s locked Microsoft in to some design decisions they are struggling to overcome.
If you want to blame people for the adoption, blame AOL, Dell, Gateway, etc. who could have shipped Opera just as easily as the default (except the price, so maybe you should blame capitalism while you're at it).
Regardless of the author's web developing skills, it's a valid metric for evaluating a web browser.
No matter how well Firefox and Opera employ W3C standards, they still need to be able to display poorly created pages just as well as valid, semantic, XHTML-driven sites.
Yes, there are a lot of people who make a lot of workarounds for a lot of browsers. Those who lament this fact should get over it. The companies involved know damn well by now what business they're involved with. Folks have got to stop belly-aching and bitching over these now decade-old problems. They're well-defined problems, which is a good thing. It takes some tricky work to keep your backwards compatibility and introduce new ways of working, ala Internet Explorer's DOCTYPE mode. If they are concerned about people introducing hokey work-arounds that they would eventually have to work around themselves, browser makers would do well to be more involved with the design community.
Fight fire with fire? Nah, that won't float around here :)
Seriously, though, good call.
Okay, I think I understand. Being able to refer to another elements compiled rules value would be sorta useful, though pretty far out there I think. I can already picture the stock price of Tylenol going up for the developers implementing that ;)
I think you hit on the reason for no variables (the comma rule separator). While we're on the topic of reusable values, I've always wanted to have an external color table for PNGs and CSS so I could change my color scheme across the entire theme of my project without regenerating my image files (I can sort of accomplish this with SVG, but there are so many headaches with using SVG across browsers I don't think it's worthwhile; let alone use SVG for a background image).
I'm not sure if you're just being sarcastic or not, but I'll give you the benefit of the doubt.
With regard to your first question, "arbitrary values of arbitrary tags" are you referring to selecting an element with a particular attribute set to a particular value? If you are, CSS 2.1 already supports that input[type="submit"] for example will select a submit button. (of course support isn't 100% there, but it's rather neat)
As for variables, I suppose you could use XSLT to generate your style rules, but the use of XSLT is more geared towards transforming the content of the page, not making it pretty... I'm curious what use client-side variables would be without the support of conditional statements, care to expound on your interest in CSS variables?
My question has to deal with selectors: why doesn't CSS have a parent selector? All too often I find myself wanting to refer to a parent or a sibling of the current element but cannot (I'm sure there are a bevy of workarounds for individual cases...)
ie Why can't I do something like this to refer to a table row (given a structure of <tr><td><a>) a:hover:parent:parent { background-color: #cff; }? It's a simple example, but there really isn't a way to do this without involving a client-side scripting language.
Thanks for your consideration.
Actually, most tags can be closed in that way in HTML without problems. The script tag for some reason demands itself to be closed with an explicit end tag. img, object, br, hr all work fine with the XHTML style in HTML (they even validate). meta and link require exactly no end tag in HTML, and are invalid if you add it in.
If it was FreeBSD or OS X, you could have tried single-user mode. I assume most Linux distros have an option for that as well, but I could be wrong. Once in single-user mode, you can mount the file system as normal and edit the config file.
Some of those names existed before the 'net came about, so searching wasn't a consideration for those product names. (think Lotus Notes, if you search for Notes, you'll be SOL)
Publisher is a poor program with a perfect name: an untrained user can probably guess that Publisher is used for some sort of desktop publishing.
And of course, "Sequel" isn't a product any more than "SQL" is a product that Microsoft sells. SQL Server (which is a perfectly cromulent name to search for).
Are you sure you're not thinking of Adobe here? LiveMotion for instance? The only reason I got it was cuz I bought their Web Designer bundle 5 years ago.
You should be able to script the Quicktime Player in the meantime.
That's not the best solution though, and I agree with the point you made about DRM from the developer's POV.
I agree with your assessment of Real (especially the basic player v. 10--quite nice) but Windows Media Player for Mac has a ton of faults that aren't worth overlooking.
The first is that it was never updated, so all outstanding bugs remain outstanding.
The second is it works quite poorly as a plugin in web browsers. There are many occasions where WMP just disappears for no reason other than it's a crappy product.
The third annoyance I've got with it is that you can't properly fast-forward or rewind. At least those controls work well within Quicktime.
From the sound of it (complaining about nagware) you're not using Quicktime 7. Give it a shot.
Two browsers that have been (in my mind) at the forefront of offering accessibility features are iCab and Opera. However, nearly every screen reader works with Internet Explorer. Is there some future hope of seeing screen reader technology in a future version of Opera?
I dunno, ask Apple...
+5? Wow.
How about "Arbitrary Collection Manager" instead? Just cuz you're a typography-geek doesn't mean everyone thinks of that when they hear "type." Ever hear of a "file-type?" Well, that's what's being described here, although the things don't have to explicitly be files.
Here's the thing: it's still good business to build for IE first.
Before you all get fired-up and start yelling "standards" you gotta realize that the web doesn't quite offer up the promise of Java (write once, run anywhere).
You start with a restricted featureset, and build a product based on that. That gets you money. This money can then be used to accomplish your dream (I dunno, supporting some weird beta of a Plan 9 only browser maybe).
You put up with restrictions because it's not finished yet. It's similar to asking why do we put up with bugs when we're running something we paid for? You just do.
Physics is the only hard science with enough balls to call something a "law."
/just sayin'
All other hard sciences are humble enough to allow for future disproof, hence the "theory."
If they already have his DNA, why not just clone him and kill the clone? That way, they'd know they have the body of the man in question.
:)
Seriously people, stop overcomplicating the issue here
Stay in business? :)
Kidding! I love my Tivo. I just hope they can stay afloat.
Well, this should be the first time an actual standard will be created for the web then.
I know what you're thinking, "the W3C! they make standards!" Well, they don't. They recommend, but don't actually create any standards at all. Realistically speaking, standards are what you find in the wild (especially in the case of the web). mTLD seems to be doing what nobody else ever had the cajones to do online: enforce a standard on people. I'm quite curious to see how well it goes over and how that translates over to other TLDs.
I'm especially curious to see if mTLD enforces proper usage of MIME types so XHTML might actually be served as an XML application.
GCC promotes vendor lock-in too.
Just try to compile the Linux kernel with some other compiler.
Still better than using Exchange server...