3D models for Adobe Reader 7
October 19, 2005 11:54 AM   Subscribe

Some nifty 3D models for test-driving the Universal 3D capabilities of Adobe Reader 7.
posted by nthdegx (13 comments total)
 
Intresting.
posted by delmoi at 12:31 PM on October 19, 2005


Very cool, although the speed leaves much to be desired. I was wondering how they'd handle this information in the UI and they didn't let me down - it's absolutely horrible.

I've always wondered how 3D model presentation will be perfected. Adobe has tried their had with Atmosphere and now this, and Shockwave has had some interesting work done in this area.
posted by prostyle at 1:03 PM on October 19, 2005


Yay it doesn't work in Linux. How awesomely Universal!
posted by jlub at 1:09 PM on October 19, 2005


jlub, I can't see it either with the official Adobe 7 Reader for linux. I guess we need a plugin?
posted by kuatto at 1:16 PM on October 19, 2005


To be fair, that's Adobe's fault, not a u3d problem.
posted by nthdegx at 1:24 PM on October 19, 2005


kuatto: I assume the plug-in is bundled with the Windows version, and doesn't exist for Linux. I could be wrong, of course.

nthdegx: I have no doubt that it is Adobe's fault. Could someone remind me what the purpose of Reader is again? Something about anyone being able to view your documents?
posted by jlub at 1:30 PM on October 19, 2005


Doesn't work in Safari either, gotta save it then open in Acrobat. So slow I couldn't really do much once I got it open.
posted by gunthersghost at 2:44 PM on October 19, 2005


Wonder why they chose such sucky models for their examples?
posted by signal at 9:31 PM on October 19, 2005


I had the misfortune to write an exporter for the U3D format.

The specification is a steaming pile of shit (I hope not all ECMA standards are this bad). For instance, the normative explanation of their bit-coding scheme is incorrect (section 10.4.4.5---it doesn't match the behavior of the non-normative example code, but viewers expect encoding according to the example code). There are also a few bizarro passages that seem to make no sense, like a note that "values larger than 0x1fff" lead to overflow errors on 32-bit hardware---even after taking into account that this value might be multiplied with 0xffff, there's still no overflow.

The alleged open source viewer doesn't build (even when I tried to build it with the special C++ compiler available from that website) and is Windows-only anyway. The required version of Adobe doesn't work on NT4. So forget developing for U3D on open-source platforms with free tools, or even on a pre-2000 machine that has a Microsoft OS installed.

The viewer software we used (and reportedly the one Adobe purchased to embed in Acrobat) crashes not only when faced with certain invalid data (hey, can I pwn your Windows computer by sending you a flawed PDF file?) but also when faced with "uncompressed" data which SHALL be supported by all viewers according to the standard (9.4.1.2).

I would love to go on, but thankfully my part in this finished in August and I have forgotten many of the things that made my blood boil back then.
posted by jepler at 5:51 AM on October 20, 2005


OMG, the UI, performance and quality is way worse than any of the old, now extinct VRML plugins (remember Cosmo)? That, plus you need to run it inside a PDF? Yow....
posted by kokogiak at 7:14 AM on October 20, 2005


I love it when technological advances for an open, universal format actually have the exact opposite effect of the original idea. It makes me feel all warm and tingly inside.
posted by illovich at 11:55 AM on October 20, 2005


Uh oh, adobe is getting too big for its boots.
posted by taursir at 3:58 PM on October 20, 2005


kokogiak: A few years ago, the last time we tried to ride the multimedia 3d wave, we added support for VRML output in our software.

I don't think any of us realized at the time how easy it was to write the simple ASCII text format (compression with open standard gzip optional) of a VRML file. Not compared to the insane bit-coding schemes and continuous-detail mesh quantization something something bullshit unnecessary complication. (but hey, it's somewhat smaller than the equivalent compressed vrml file, and you only have to implement it once, and that's easy once you figure out what parts of the standard you have to ignore in order to work with the existing viewers)

If you want to create your own standard, create a low barrier to entry. That Open Source implementation I mentioned earlier---the one that doesn't even compile---is several million lines big, or maybe only a few hundred thousand once you ignore the copy of libpng and libjpeg it's shipped with. It includes its own implementation of something similar to microsoft's COM (but not actually compatible with Microsoft's COM, or CORBA, or Mozilla's XPCOM). IFXCOM, or whatever it was called, also included a non-STL implementation of strings and containers. It seemed to serve only the purpose of making it difficult to re-use part of the library (such as the bit-coding scheme) in another piece of software. Easy to add U3D to your software? Fuggedaboutit.

See also the 2004 usenet post, Intel's 3D Divorce Rate [google groups]

Likely case: U3D dies a quiet death and it's another 2 or 3 years until the next 3D standard, the one that will really take root (no, not really), shows up and soon fades into obscurity with all the rest.

Impossible best case that will never happen: A new standard based on OpenGL and a cross-platform language with a sandbox model, and it actually does catch on, but strangely is never adopted by advertisers.

Worst case: Intel and Adobe get everything they want from U3D. I spend my time blocking 3D animated pop-ups instead of the normal 2D animated pop-up ads I spend my time blocking today.
posted by jepler at 7:39 PM on October 20, 2005


« Older Firefox 100 Million Downloads   |   The Psychology of Tyranny Newer »


This thread has been archived and is closed to new comments