Slashdot Mirror
More Insight On Longhorn's Avalon And Aero Design
Lispy writes "While monitoring the Xorg mailinglist I came across this set of WinHEC PPT-presentations (work fine in OOorg) that cover some interesting details on the underlying architecture of Aero, Aero Glass and future font rendering in Microsoft's upcoming Longhorn OS. What does the Slashdot crowd think about the overall design and its downsides, such as power consumption on notebooks?" (KPresenter works fine, too, btw.)
The date is pretty much up for grabs. Everyone has their own opinion - everyone is speculating. Some say as soon as next year, others say as late as 2007.
Last I heard, it's schedule for the day after Duke Nukem: Forever.
When certain fonts are displayed in Windows, certain characters are cut off on the left. It seems when the bounding box of text is calculated, it is incorrect, allowing some text to be cut off when displayed. I've noticed this is IE and Word especially.
Hopefuly they can fix this in the new fonts.
http://github.com/gbook/nidb
Joe Beda had said that Avalon is going to be more of an advanced UI/Visualization toolkit, while Dx will continue to contain all the other serious stuff.
Now, it looks like Avalon can do 3d on it's own and maybe more too -- what's the idea behind this anyway?
Are they trying to get a fresh new API or something? It seems unlikely, since I remember Joe and Scobles saying that they will probably be using Dx for serious graphics and game development. The redundancy seems strange.
From the presentation --
Avalon 3-D are not a replacement for Direct3D
You will find Avalon 3-D useful if:
- You want to integrate 3-D seamlessly into an Avalon app that also contains 2-D content, controls, etc.
- Platform features like Remote Desktop and multimon are high priorities for you
- You want to easily add 3-D functionality without quickly without needing to learn how the graphics hardware works
You will find DirectX useful if:
- You want access to all of the features provided by the graphics hardware
- You want to have full control over how your scene is stored and managed in memory
- Plan for interop between Direct3D and Avalon
Render Direct3D in a HWND and host within Avalon
So basically it seems to help ease the creation of bells and whistles, more than anything. Weird.
And oh, completely offtopic -- what's the deal with saying, work fine in OOorg -- shouldn't that be works fine with OO? Why the org/.org thingy?
Apparently an Aero testing suite will be released at least 18 months before Longhorn's release. Is this demo yet available? It's interesting to note that if this message is right, Longhorn will not be available until 2006.
Will MS publicly announce this Aero test, so that we can anticipate a real release date for Longhorn?
Meanwhile, in the real world outside slashdot, patenting everything has been _everyone's_ tactic for at least ten years.
I remember one place I worked, every engineer _had_ to file at least one patent a year, even if all they did was write device drivers... had to do it, though, in case someone else ever sued.
Whence? Hence. Whither? Thither.
This is /. and you're asking about how they will like a Microsoft technology? Of course, they'll hate it. Microsoft could come out with something that that's the coolest thing since Linux and /. will still hate it.
Peeeerrrrty
Look at those two screen shots. If Longhorn wan't an OS I would bang her 4 times a day.
Then again, I would get a virus, like MSclap
=)
Icons with reflection and depth
While I have to admit it sounds cool, I can't really think of a real need for this.
Quantum materiae materietur marmota monax si marmota monax materiam possit materiari?
I'm not interested in warezing it; Why steal a Ford Pinto when Linus is giving away Lamborghinis for free?
Almighty Railgun
You Speak a Lethal Gospel!
Bloody Gibs Follow.
This has been going on for at least wo weeks. You can get to the download.microsoft.com by appending "c.footprint.net" on the end of the server address. So, the link becomes:
n load/1/8/f/18f8 cee2-0b64-41f2-893d-a6f2295b40c8/TW04006_WINHEC200 4.ppt
http://download.microsoft.com.c.footprint.net/dow
Go figure. I have no idea excatly why this is happening. I'll leave that to people who care.
call it longshot
What would Edward Tufte make of this plot? Ah well, perhaps the multidimensional rotating bar graph will be of some use, should the presenter want to conceal some data.
It's somewhat disappointing that the presenter chose not to include a lens flare, though.
Perhaps "Iron Pyrites" would probably be the best mineral related name for Direct3D.
Bad analogies are like waxing a monkey with a rainbow.
He was implying that nobody buys Windows. Which is true... one of my friends said this to me "Linux is nice, but Windows is free too. If I had to buy it I'd use Linux..."
I'm the only person I know that has a legal copy (and I have NEVER had a Windows partition in my life... I used MacOS before I switched to Linux) because M$ sent me some brainwashing kit for UNIX developers or something to that effect.
My other car is first.
No, not hot.
But reading Avalon's text support it seems that Longhorn will FINALLY be able to have the same deep text support that OS X has had since at least 10.0. Yes, all the APIs are marked AVAILABLE_MAC_OS_X_VERSION_10_0_AND_LATER with most of them having support in CarbonLib 1.0, ATSUnicodeLib 8.5 (Mac OS 8.5). Nice to see longhorn might finally be catching up.
The only thing that longhorn claims it will have that ATSUI doesn't have yet is the graphics card rendering support. Ever wonder why resizing a window is so slow on OS X? ATSUI is the reason.
In what way, if any, is this different from Apple's Quartz techniques?
from the 2nd presentation (in huge capitals, orange text on blue background to make your eyes bleed. So far for userfriendlyness)
"64-bit is the future !!"
Doh. MS is missing the ball by a few 100 miles again : Billy, 64-bit is THE PRESENT. 128bit or nanocomputing is the future.
When will I end this grieving ? When will my future begin ?
Take a look at this cam of Jim Allchin's Keynote showing off Longhorn's directx capabilities.
e c/ WinHEC2004-JimAllchinKeynote.zip>
This makes it pretty clear why Apple is trying to patent transparent windows and some other elements of their UI.
<http://www.neowin.net/staff/creamhackered/winh
http://nyamenation.org/
From Logos, Trademark, and OpenOffice.org in a Nutshell:
- Sw Usr
Windows seems to be going down the road of "show fewer things but with bigger pictures", which may be great for regular folkum. Advanced users will just scrap the bells and whistles anyway for a basic, functional setup. "Dumbing down" through simplification isn't always a bad thing though. I actually like the new WinXP start menu a lot better than the classic one, albeit with small icons instead of the huge default setting. Silver Luna isn't too bad either, as long as I reduce the size of the titlebars and buttons to classic size. Again, what's the deal with Microsoft and huge buttons and icons? Are they trying to cater to the bad eyesight but too cool to wear glasses crowd?
are they creating video game or operating system?
What does the Slashdot crowd think about the overall design and its downsides, such as power consumption on notebooks?
:-P
At least the submitter understand there's no use asking for its upsides here.
Beware: In C++, your friends can see your privates!
I noted with quite a bit of interest that near the end of the Longhorn Text presentation, they claim that 1,000,000 sub-pixel antialiased (cleartype) characters can be rendered per second (8 to 11 point type, on a 96dpi LCD).
:(
Does anyone have any similar performance figures for sub-pixel AA font performance on Linux? I have a sinking feeling it might be closer to 10,000/second
I'm using gentoo primarily, and I have a box running Fedora Core 2. Bluecurve is nice, but it still can't hold a candle to Mac OS X or Longhorn (from the screenshots I've seen). With SVG now implemented in GNOME, it would be my guess that there will eventually be a big wave of pretty vector-based icon sets and themes. I sure hope we will be able to compete against the big guys (Microsoft) though.
Life is offtopic.
"Which is true... one of my friends said this to me "Linux is nice, but Windows is free too. If I had to buy it I'd use Linux..."
Are you sure he means free as in "I can dl it" vs. free as in "it comes with every computer"? In the latter case, people are buying Windows.
"Derp de derp."
I think the features they are talking about are nice.. But personally I still think their sub-pixel font anti-aliasing looks bad. http://bluehalo.homeunix.org/text/ Shows the line from the PPT and the same thing typed in OS X.
Talent hits a target no one else can hit; Genius hits a target no one else can see.
I know that foghorn is going to have a high wow-factor, but do we really want to be gaping with awe and amazement at every little tip of the veil of a system that won't be out till at least 2006?
My take is that our time is much better spent improving our prefered (open!) system, exploiting the great features that Hans Reiser has given us (which I personally find much more interesting than all the eye candy that serves to addict, distract and slow down my friends and their computers).
Extended attributes are here today. So is OpenGL. Where are the applications that exploit them? Where are the BeOS-like filesystem queries on Linux? Where are the Baldur's Gate clones? And, most of all, where is the stuff that, once and for all, asserts the superiority of the open source community, the proof that we can invent, rather than wait for the corporations to do it for us?
Please correct me if I got my facts wrong.
In the immortal words of Colonel Kurtz, "You must make a FRIEND of the horror".
.Net secret sauce coming with Longhorn, will no longer use Remoting "over the wire". Everything is going to be SOAP and web services. Sounds wonderful, right? I think so too. BUT, many consultants are busy writing all your company's apps using remoting between servers! Guess what THAT means?
What horror, you ask? A major Microsoft upgrade. We cal look forward to the following exciting experiences in the coming Longhorn apocalypse ("I love the smell of burning CPU in the morning... It smells like... Job Security!"):
1. Indigo, the new
2. All your computers are going to be landfill fodder, because Longhorn's hardware requirements are going to SMOKE 'em. Ah, well, we didn't need those 20,000 PCs anyway. And, the budget looks so much better cratered. It's like a big empty swimming pool. Makes me think of summer.
3. Performance? The users are asking about performance? Um... HUSH! Look at the pretty screens, children! Ooh, transparency!
4. Filesystem? We don't need no stinkin' filesystem. Let's put everything in a DATABASE!!!
Ok, they might not get this into Longhorn, but it's coming. All your apps that touch the filesystem? Kiss 'em goodbye.
5. More DRM. What's that? the users didn't ask for it? Let's surprise 'em; they'll be so happy!
6. A new, different and strange iteration of IE to worry about. Sigh; better set up resources for the recoding of all your web pages, just in case.
Ah, well. It should be exciting! And, who knows? Maybe the Indians will find it all just too ugly to work with and offshore all the work back here ("Oh, this is just too UGLY, you may take it back, please... No, really. No, I must insist. Oh, you are too kind, sir, but NO, I REALLY must insist... Oh you are making me very ANGRY sir, do not make me go medaeval on your unruly buttocks in the manner of Marcellus!").
Farewell! It's been a fine buncha years!
I believe this should, mind you, SHOULD raise the heat under the open source movement.
Do your OS of choice under the following:
Make it so it runs on anything from legacy hardware through current hardware.
Don't engineer it specifically as how YOU would want it done, engineer it as how you think Joe Sixpack would like it to be done. Do you know how MS keeps its market share? By making adaptive shifts to their new setups as small and painless as possible. Stop assuming everyone who wants to try Linux already knows every manual and howto available.
Screw the DMCA, reverse engineer everything (do it the old fashioned way, get 50 coders to examine 1/50th a part of the driver code, then compile accordingly, that's how it was done with IBM). That way everything can be supported.
If you want to be mainstream, start acting like you ARE mainstream. This "Lookit me, I'm a rebel!" illusion is just that. That's how Apple did it, that's how Microsoft did it. And look at them now. The rebel theme is only good as long as you expect to lose money.
I'm a MS user myself, but the DRM crap and all makes me WANT to go Linux, but the fact that not every Linux dev doesn't support EVERYTHING I want to do or use, means I'm stuck with MS until they realize this.
It's like wanting to escape from prison, while everyone else is debating the best kind of file, and what kind of cake to bake it in.
Just because you can mod me down, doesn't mean you're right. Shoes for industry!
Windows's idea of eye candy was that menus (and submenus) would all slowly fade in. The process of navigating deep into hierarchical menus was maddeningly slow--at least until everyone turned it off.
In osx, menus appear immediately, and then fade out after you select something. This is not only pretty, but functional: it gives you visual confirmation that you've selected a menu item, which can be helpful if whatever you've asked for doesn't produce obvious or instant results.
Microsoft's cargo cult design process often leads them to such mistakes. They manage to take the wrong lessons from other people's work, and conclude that what people want is snazzy looking things which tax hardware. The real lesson is that people want visual continuity and feedback in order to speed up their use. But Microsoft never seems to get as far as understanding the point of anyone else's design, just the appearance.
The text rendering technology impressed me. The new APIs did not.
When will Microsoft learn...developers don't want great big heaps of their grand designs. There's just too much of an investment to learn their way of doing things, there are too many cages around the good bits, and everything breaks when you go off the beaten path.
Example: few 'real' apps use MFC - and certainly none of Microsoft's. They expose 'Fisher Price' versions of their tools which they hand code in good old SDK.
I mean, does anyone *really* use DCOM? I guess COM has held on bascially because there isn't that much that is 'Microsoft' about it and it basically works. But what happened to ATL? DDE? ActiveX? In fact, the only useful Microsoft software tends to be the stuff they acquire (Visio, SQL Server)
Now I have not used one iota of Avalon, but I remain unconvinced it will be anything other than their typical developer traps with a bit of 'hello world' cute app as bait.
Meanwhile, Linux hacks change the world with Perl scripts. Go figure.
Does it hurt to hear them lying? Was this the only world you had?
The real meat of the whole thing starts here.
A really, really long chain of postings between the guy who wrote the article and a guy who seems to know OSX rendering pretty well.
My take on this is that OSX has tacked a lot of issues with making vector display practical, and trying to maintain a good balance between "everything is a vector and you all need new computers to run the OS" and "everything is a bitmap and you can run this on a 286". Personally I think a lot of graphic designers will be aghast at the limitations an all-vector approach will impose for things like icons - you can see guys spending days tweaking pixels. You may think you've done them a favor by taking that options away, but these are the guys that make your interfaces look good! Treat them nice, I say.
I would also say each is holding his own pretty well in this argument, it did not get too far into name calling and the like (gets more technical as you progress - my link takes you pretty much to the point wher ethey drop the childish bits). I do think the Avalon guy is a little more ignorant of what is going on in OS X rendering-wise than the OS X is of Avalon - the OS X guy for a while was unwilling to believe that anyone would actually take an approach with only vectors, but understood fully that aspect later on.
The Avalon guy has a good point that it's cheaper to send a lot of vectors to the GPU than to calculate very high DPI images for display... but I think the OS X guy has a good point that you can't have the GPU do everything.
Here's a simply summary from my read (not comprehensive):
Article guy: Maintains the vector retained model is the only scalable UI solution. UI's should only be collections of vectors (including all icons and the like) with everything, even text, being rendered by a GPU on your video card.
Also maintains that parts of OSX are not really PDF/DPS based, and therefore will not scale.
Has not yet answered if he thinks it's a good idea for the GPU to be doing typography (like kerning).
Longhorn will require new GPU's for sure, from everyone.
Avalon target is display showing about 10,000 primitives.
Avalon targeting high (300dpi+) displays, only path to good performance is feeding vectors to a video card.
OSX Guy: OSX does use PDF/DPS model correctly, elements will scale - renderer can be set to any DPI. To use the GPU for all drawing operations is madness - OS X has quick operations for things like video, and slower but much more exact operations for things like 2D operations. To use GPU for all operations is going to be a disaster as they cannot handle quality 2D operations (like exact text rendering) all that well.
Showed how saving a window as PDF yields scalable elements (not just a raw bitmap).
Maintains that OS X can support high DPI displays, you just change the rendering target.
Can rotate window contents in real time.
Expose is not supposed to be an exact vector resizing as such an operation should be very fast, not exact.
"There is more worth loving than we have strength to love." - Brian Jay Stanley