A serious question here: what features do you need that aren't covered by simply using 'display: table-cell' and 'display: table-row'?
You've already mentioned row and column spanning.
I have yet to see a web page that uses smart packing behavior... columns packed based on their contents, and then expanded intelligently when the window is resized. It's pretty rare to get a page that even handles window resizing at all. This should be able to be statically defined in CSS, without having to specify any absolute widths.
There is pretty limited anchor support in CSS.
Have a look at what the three layout managers in Tk can do. Ousterhout held off on a grid layout manager for a long time, trying to make do with packing and floats, but he finally adopted one.
In some text formatting languages you define the layout with no content, with named boxes, and then the content is placed in a separate container... eg ""... "...". I don't know the algorithms text formatters use for sizing non-contiguous regions, but there's gotta be code in open-source programs like TeX and Lout.
This would also solve your first problem, and more generally.
I've been using Pages and it's got all the structural problems of Word, with no nesting and primitive directly-manipulated layout. Apple would be better off, I think, making HTML a primary target of the editor rather than trying to mimic the structure of Word documents. Is the new version any better?
Will you quit babbling about what IE does or does not support... because I could care less. I *banned* IE at our office from 1997 until we got borged in 2002, for security reasons (and what happened after we were forced to use it again really made me look like a hero). Nobody should be using IE for anything but a last resort for broken sites.
If CSS supports grid layout then you should be able to show me a page that I can view in Safari, Camino, Firefox, Opera, or any other decent browser, and is still readable in IE.
Just being able to display a div as a table cell is just part of it.
HTML and CSS are different specs like Ford and Lincoln are different companies. HTML and CSS have to work together and build off each other.
There is NO grid layout support in CSS2. There is crippled packing support that, when combined with painstakingly cumbersome absolute-sized and positioned divisions, can be used to simulate very very simple grid layouts. This is so far from a real grid layout engine that it's not funny.
The flow support in CSS3 is similarly limited to a specific canned solution to a specific problem.
The result of these half-assed solutions is web pages that are actually *less* capable of adapting to different window sizes, let alone different display devices than if they were written using tables for layout... which is what makes this an HTML problem.
There are three main layout models in text formatters: cartesian coordinates, packed boxes, and grids. CSS floats (both absolute and relative) are cartesian, the traditional non-floated divisions are packed boxes (albeit a fairly primitive packer), but there's no grid layout other than tables.
The other missing thing is what you need to create automatic columnar layout, the ability to define divisions as being part of the same flow, so that text can fill across division boundaries.
All the special tags they've come up with, though, can be handled by standardising names for spans and divisions. In fact that's pretty much all they are, new aliases for and
.
Microsoft Windows has had critical problems from the start with the way applications are forced to work in a single flat namespace, making it unreasonably difficult to install multiple instances of an application on a single computer. That's what's really been driving virtualization... most of the problems virtualization solves have much simpler and more efficient solutions on UNIX... or virtually any other serious operating system.
Most of them are based simply on taking advantage of the hierarchical file system and the process hierarchy.
All a well behaved UNIX application needs to run isolated from its incompatible brothers is to inherit an environment from its parent that tells it where to find its configuration parameters and files. This can be as simple as running out of a particular directory or using an environment variable, or as complex as a "chroot" environment.
This has been standard in UNIX since it was created, and Microsoft knew about it... they had the most popular UNIX variant in the world in the early '80s, before they followed apple down the cul-de-sac by designing an OS around the GUI instead of making that simply another resource that the OS manages. Now, they're coming up with an inefficient solution that will let some small portion of their users get a fraction of the capabilities they would have had if they'd stuck with Xenix as their premier OS.
Why? Not every machine needs to be equipped for higher end gaming.
The GMA950 hurts any 3d application, and any application that pushes the limits on RAM - and not only because it eats 64M of real memory. It's not just third-party software (let alone games) that exceed the limits of what the GMA950 can do, Apple's own software uses 3d effects all over the place, so it's got to load their software OpenGL to cover for the shortcomings of the GPU regardless. And it's going to be using more and more of them over time.
I mean the original Mini's GPU was marginal, and Tiger required more than it could handle mere months after it was released... and *it* was more capable than the GMA950. It's only because they could afford to waste CPU power to inefficiently cover for the Intel GPU that they got away with it in the first place.
Your math assumes you can throw the entire swing vote in those states on your side, *and* your opponents don't figure that out.
If it's truly a proportional vote and the election is close, then your *majority* in those states is still going to have to beat the swing the other guy gets in the states you ignored.
It's certainly less ergonomic than the previous model... you put skinny keyboards in laptops because you need to save space, not because they're actually *good*. Every new laptop I've used has had a skinnier keyboard and it's been more painful to work on. I can't use the Macbook Pro keyboard at all... fifteen minutes and I'm in pain.
At least you're not stuck with them, like you are with a laptop.
I think you're stretching on the RAM, though. They have 2GB installed but you can upgrade them all to 4GB... for less than Apple would charge.
You *like* lugging those big heavy jujubes with their absolutely horrible screens (a huge step back from the Trinitron screens Apple previously used) around?
It's not about style, it's about the CRT iMacs being a dead end.
Does this mean that I can go to CompUSA and demand *they* provide me a copy of the source to any versions of Linux embedded in any routers or other devices that they sell, since they're part of distribution of those devices and therefore (by this theory) they're also specifically bound by the GPL?
The impression I get is that they're talking about buying and selling *outside* the game. If you buy a virtual broadsword for EQ gold, that's not a transaction in real money. If you buy it on eBay with US$, that is. So I don't think someone who wins a broadsword "worth" US$1000 in Warcrack is going to be liable for $1000 in capital gains or sales/gift tax... until they eBay it.
This idea that the GPL3 is designed to attack any existing deals is a fabrication. I don't know if the person who posted this article is incompetent to practice law or trying to FUD the GPL... but if even someone like me who thinks we'd be better off without the GPL can see through it, it's pretty poor FUD.
Slashdot Mirror
A serious question here: what features do you need that aren't covered by simply using 'display: table-cell' and 'display: table-row'?
... "...". I don't know the algorithms text formatters use for sizing non-contiguous regions, but there's gotta be code in open-source programs like TeX and Lout.
You've already mentioned row and column spanning.
I have yet to see a web page that uses smart packing behavior... columns packed based on their contents, and then expanded intelligently when the window is resized. It's pretty rare to get a page that even handles window resizing at all. This should be able to be statically defined in CSS, without having to specify any absolute widths.
There is pretty limited anchor support in CSS.
Have a look at what the three layout managers in Tk can do. Ousterhout held off on a grid layout manager for a long time, trying to make do with packing and floats, but he finally adopted one.
In some text formatting languages you define the layout with no content, with named boxes, and then the content is placed in a separate container... eg ""
This would also solve your first problem, and more generally.
I've been using Pages and it's got all the structural problems of Word, with no nesting and primitive directly-manipulated layout. Apple would be better off, I think, making HTML a primary target of the editor rather than trying to mimic the structure of Word documents. Is the new version any better?
Will you quit babbling about what IE does or does not support... because I could care less. I *banned* IE at our office from 1997 until we got borged in 2002, for security reasons (and what happened after we were forced to use it again really made me look like a hero). Nobody should be using IE for anything but a last resort for broken sites.
If CSS supports grid layout then you should be able to show me a page that I can view in Safari, Camino, Firefox, Opera, or any other decent browser, and is still readable in IE.
Just being able to display a div as a table cell is just part of it.
HTML and CSS are different specs like Ford and Lincoln are different companies. HTML and CSS have to work together and build off each other.
There is NO grid layout support in CSS2. There is crippled packing support that, when combined with painstakingly cumbersome absolute-sized and positioned divisions, can be used to simulate very very simple grid layouts. This is so far from a real grid layout engine that it's not funny.
The flow support in CSS3 is similarly limited to a specific canned solution to a specific problem.
The result of these half-assed solutions is web pages that are actually *less* capable of adapting to different window sizes, let alone different display devices than if they were written using tables for layout... which is what makes this an HTML problem.
There are three main layout models in text formatters: cartesian coordinates, packed boxes, and grids. CSS floats (both absolute and relative) are cartesian, the traditional non-floated divisions are packed boxes (albeit a fairly primitive packer), but there's no grid layout other than tables.
The other missing thing is what you need to create automatic columnar layout, the ability to define divisions as being part of the same flow, so that text can fill across division boundaries.
All the special tags they've come up with, though, can be handled by standardising names for spans and divisions. In fact that's pretty much all they are, new aliases for and .
Microsoft Windows has had critical problems from the start with the way applications are forced to work in a single flat namespace, making it unreasonably difficult to install multiple instances of an application on a single computer. That's what's really been driving virtualization... most of the problems virtualization solves have much simpler and more efficient solutions on UNIX... or virtually any other serious operating system.
Most of them are based simply on taking advantage of the hierarchical file system and the process hierarchy.
All a well behaved UNIX application needs to run isolated from its incompatible brothers is to inherit an environment from its parent that tells it where to find its configuration parameters and files. This can be as simple as running out of a particular directory or using an environment variable, or as complex as a "chroot" environment.
This has been standard in UNIX since it was created, and Microsoft knew about it... they had the most popular UNIX variant in the world in the early '80s, before they followed apple down the cul-de-sac by designing an OS around the GUI instead of making that simply another resource that the OS manages. Now, they're coming up with an inefficient solution that will let some small portion of their users get a fraction of the capabilities they would have had if they'd stuck with Xenix as their premier OS.
Who's going to pay 0.98 for an unpopular song when they can get a popular one more cheaply?
Putting a TFT inside an old CRT iMac?
That's got all the style of a Caddy with racing slicks, spoilers, and a plastic mill superglued to the hood.
You might as well run Gentoo on it and call it "Mac OS XI".
I didn't say I agreed with it, I simply said that was the point behind vote-swapping.
That's not where the mistake is.
It's the electoral college that's being propped up by bogus math, not the other way around.
My MacBook's GMA950 has handled everything I throw at it
No it hasn't. The extra CPU core has been doing the heavy lifting for it.
Why? Not every machine needs to be equipped for higher end gaming.
The GMA950 hurts any 3d application, and any application that pushes the limits on RAM - and not only because it eats 64M of real memory. It's not just third-party software (let alone games) that exceed the limits of what the GMA950 can do, Apple's own software uses 3d effects all over the place, so it's got to load their software OpenGL to cover for the shortcomings of the GPU regardless. And it's going to be using more and more of them over time.
I mean the original Mini's GPU was marginal, and Tiger required more than it could handle mere months after it was released... and *it* was more capable than the GMA950. It's only because they could afford to waste CPU power to inefficiently cover for the Intel GPU that they got away with it in the first place.
Your math assumes you can throw the entire swing vote in those states on your side, *and* your opponents don't figure that out.
If it's truly a proportional vote and the election is close, then your *majority* in those states is still going to have to beat the swing the other guy gets in the states you ignored.
That keyboard looks like the Macbook keyboard.
It's certainly less ergonomic than the previous model... you put skinny keyboards in laptops because you need to save space, not because they're actually *good*. Every new laptop I've used has had a skinnier keyboard and it's been more painful to work on. I can't use the Macbook Pro keyboard at all... fifteen minutes and I'm in pain.
At least you're not stuck with them, like you are with a laptop.
I think you're stretching on the RAM, though. They have 2GB installed but you can upgrade them all to 4GB... for less than Apple would charge.
Get rid of the bloody GMA950 and put something in there with decent OpenGL support.
Just bumping the processor to Core 2 Duo doesn't cut it. That's not the bottleneck. That's not what you need to speed up.
"I want something that does what I need at a reasonable price because I'm buying a computer not a fashion accessory"
That's option 1 as well, except that "does what I need" includes "not fall apart if I don't treat it like it's made of explosive eggshells".
(1) I need this to get work done and it has to run forever: Lenovo (formerly IBM) and high-end HP. Ugly is fine, as long as it's bullet-proof ugly
That's me, now if I could get that running an OS that (a) doesn't suck and (b) has an actual application base.
(a) rules out any version of Windows (certainly anything post Windows 2000) and half the Linux distros.
(b) rules out the rest of the Linux distros, as well as things like BSD and Solaris (and oddballs like whatever-the-latest-BeOS-clone-is).
You *like* lugging those big heavy jujubes with their absolutely horrible screens (a huge step back from the Trinitron screens Apple previously used) around?
It's not about style, it's about the CRT iMacs being a dead end.
They've got wireless and wired versions. The wireless ones look even MORE like Macbook keyboards.
Hence my question... what exactly is it that they're proposing to tax that isn't already covered?
In the sense that it's used in the GPL?
Does this mean that I can go to CompUSA and demand *they* provide me a copy of the source to any versions of Linux embedded in any routers or other devices that they sell, since they're part of distribution of those devices and therefore (by this theory) they're also specifically bound by the GPL?
Err, your original post was really unclear then, because I still can't read it that way.
Microsoft isn't distributing Linux. The only GPLed software Linux distributes is in Interix.
What Microsoft might distribute in the future isn't relevant to the argument that the FSF was trying to "back date" the GPL3. That didn't happen.
The impression I get is that they're talking about buying and selling *outside* the game. If you buy a virtual broadsword for EQ gold, that's not a transaction in real money. If you buy it on eBay with US$, that is. So I don't think someone who wins a broadsword "worth" US$1000 in Warcrack is going to be liable for $1000 in capital gains or sales/gift tax ... until they eBay it.
This idea that the GPL3 is designed to attack any existing deals is a fabrication. I don't know if the person who posted this article is incompetent to practice law or trying to FUD the GPL... but if even someone like me who thinks we'd be better off without the GPL can see through it, it's pretty poor FUD.
GPL3 explicitly recognizes this situation and exempts deals in force before the GPL3 was released.