[x3d-public] Is there even a *case* for client-side 3D at all??

GL info at 3dnetproductions.com
Tue Apr 30 17:08:08 PDT 2019


 

Excellent overview John. I've looked at the Castle Game Engine as you suggested. But Unreal is what's really starting  get my attention, partly due to the fact that its use of C++ gives it a lot of power, with full access to source code extending to the engine's subsystems. It's ultimately not free, but if trying to follow the money, it's at least as good as anything. And when used to create videos it is actually royalty free, as only interactive applications carry royalties. I've been an avid Blender user for a number of years, and that will always have its place, but getting from that to Unreal feels like getting into a brand new luxury car. Like 'iam here', I also believe that graphics quality is very important. So, I'd be looking for a possible symbiosis, augmentation, or outright replacement of FBX with X3D in Unreal or something similar, in order to turn games into work to use your terminology. I call it the gamification of work, since that has always been the direction of my efforts and, to me, the ultimate outcome. Whether a "game" is streamed from a server or not, that involves final rendering on a display device. The internal operations should essentially remain largely the same.   

 

GL

________________________________________________________

* * * Interactive Multimedia - Internet Management * * *

  * *  Virtual Reality -- Application Programming  * *

    *   3D Net Productions  3dnetproductions.com   *

 

 

 

From: x3d-public [mailto:x3d-public-bounces at web3d.org] On Behalf Of John Carlson
Sent: Sunday, April 28, 2019 11:18 AM
To: iam here; X3D Graphics public mailing list
Subject: Re: [x3d-public] Is there even a *case* for client-side 3D at all??

 

We’ve seen google 3D go flash-in-pan before.  We’ve seen Sun Microsystems Wonderland demise at the hands of oracle.   3D is not commercially supportable unless you’re getting big bucks selling an authoring system (a CLIENT system, CAD, imaging, or geo), consoles (a CLIENT) or are selling *expensive* games (more CLIENTs).  People don’t make money selling 3D servers, unless they also sell software, ads, blogs, sex, or product [Amazon (product+NetFlix client), Facebook (games, headset), Google(ads)].  My server costs $120 per year, and I have a max of say, 2 users on my system at a time, and Digital Ocean is probably half of that.  Amazon for 1 months for an idling system is like $50.   That doesn’t last long for the small indie guy.  It’s much cheaper and faster to do peer-to-peer I’d imagine for multiple user games (choose best connected/powerful client as the main server). Single user games have no use for a server except for updates and uploading and viewing score data (low bandwidth activities), and displaying ads.  I am looking at moving my content to gists and blogs, so I don’t have to host it any more. Look at all the CLIENTs Netflix supports B’god.  If NetFlix got into the game selling business with Roku or Apple TV, they would be a force, but NetFlix wants to be on all gaming consoles, not competing with them.  So the best bet to knock over X3D is Adobe, frankly. People have made advances to integrate X3D into Acrobat, but so far all offers have been rejected. I think this may be due to John Warnock’s creation of a 3D language before PostScript.  None of Adobe’s forays have succeeded either (and they are in the archiving business big time.  That means your code has to last *forever*, through a *nuclear* attack, etc.). That’s right folks there have been 3D languages and APIs and standards before OpenGL/VRML/X3D (GKS, PHIGS) That’s ancient history. VRML was original developed as OpenInventor, a web-enabled layer on top of OpenGL by SiliconGraphics, probably the only two things that make it’s out of SGI’s demise (CosmoCode too?). The real standards in this arena, and where the battle is is WebGL/OpenGL, probably Vulkan (next OpenGL from AMD/Khronos), CUDA (Nvidia) and Metal (Apple) have absolute control of the mobile market (with Direct* on Xbox and Windows systems). I don’t know what Sony is using on the PS4.  Nor Nintendo.  None of the shading languages are declarative or scenegraph oriented, (higher level than the shading languages),  [ shading languages deal with matrices, textures, colors and vertices/polygons—NOT shapes and hierarchical structure suited for animation, except through polygons ] to my knowledge, which X3D is. Our competitors are things like Collada (Sony), FBX (AutoDesk), PlayCanvas (independent), Three.js (independent), Babylon.js (Microsoft), XSeen (Leonard Daly and friends), A-Frame (Mozilla) and XML3D (independent). In other words, X3D IS NOT IN THE GAMING BUSINESS!  X3D/VRML can be primarily seen as the output from animation tools and tools like Maya, Blender, Jupyter notebook, MatLab and octave.   NOT A GAME!  We are REALLY into things like CAD (AutoCAD and PTC have the market share) Geo (ESRI has the market share, I think) Medicine (HL7 is big time here, followed by FHIR, which includes X3D)  Military Simulations (DIS) and other types of visualization (and one or two games). In other words, we are REALLY into developing WORK standards, NOT GAMES.

 

Out of those which have you heard of?   Which of them offer a declarative interface like HTML? Now you know why X3D/VRML is on top of the (rather small) heap it’s on top of.  COLLADA currently has the glTF2 market, but we’re moving glTF2 into X3D. You can translate FBX to PlayCanvas’ formats, but then they aren’t FBX ultimately

 

The main *work-related* web graphics are glTF (and it’s pal COLLADA on the server), think of this at a lower level than X3D, closer to webGL, we can incorporate it. Three.js, the most popular 3D JS framework, I think, has its own external format.   We have limited interoperability with it, but see three-x3d-loader. A-Frame (incorporates glTF and Three.js I think) is next.  It has  growing number of tags. Note that all of these use WebGL, so the shading language competition is pretty much moot on the web.  Note X3D supports any shading language it’s host browser does, but the shading language may not be portable across all browsers (which is why there’s a push to move the shading language out of X3D).

 

Notice out of all of these, AutoCAD, PTC, ESRI, and and FHIR/HL7 are market, industry-standard leaders for *work* 3D content.  Google is being a fly in the ointment for those (google earth, google maps, etc., for the consumer, not professionals, SketchPad?), offering google docs like office duplicates for groups, to distract the leaders from the real market, ads and potentially for google, AI. A typical one console ad network shop makes $100,000 per month on ads.  There are ad markets where people compete to sell ads.  X3D is a *work* related format.  Not an ad related format.  I think Web3D would start cheering very loudly if Microsoft took notice of X3D/VRML, Microsoft pretty much defines the work world on computers. We could use a little embracing and extending (XML is the Extensible Markup Language, don’tcha know.  But I don’t think Microsoft is going to take on AutoCAD, PTC or Adobe with X3D any time soon, they’re too busy with HoloLens, Games and Direct* and fighting Sony to deal with something that they frankly can’t make money on except for Medical purposes (think of running on all those Windows XP systems in the hospital and doctor’s offices, eek).  EDI has not been very popular since the web, and perhaps HL7 sees this as an opportunity to jump on web tech (eek).

 

I would encourage X3D/VRML to continue to court Adobe, by creating a stack-based encoding (by incorporating tags as stack commands or dictionaries, potentially with a new ANTLR grammar or revised VRML), like PDF and PostScript.   If we had Adobe on our side creating tools for a stack-based standards based encoding, we could compete.  Also, we would integrate with the image/text/curve in PDF and PostScript with (movie)texture/hypertext/shape in X3D.

 

If you haven’t noticed it, all but heavy AI processing has moved to desktops and mobiles, with AngularJS, YouTube, Vue and React taking up 1 or 2 CPUs.  That leaves 4-6 CPUs plus the GPU free to do 3D.  Meanwhile Google and Microsoft are trying to invent new things to do with their AI GPU cloud farms which never really materialized on the market (yet). Thus copy OnLive and Sony.  The only reason I can see for Google to enter this market is to distract MS from ads and the web, where MS has recently made a coup with the GitHub procurement.  I can see google competing with Nvidia however.  Maybe that’s the intent, and MS is backing Nvidia up.  Hmm. My guess is google is building a humungous AI farm disguised as a games farm (and they need a place to use their new TPUs).  They’re going to train a new cadre of google employees to challenge Alpha Go and Alpha Zero.

 

If you’re looking for the Game (application creating) competition for X3D (not the animation competition), there’s PlayCanvas, Unity and Unreal.   That’s pretty much it, except for niche markets. X3D is very limited in the games market, as we’ve never really targeted it. However, there’s a Unity browser for X3D (yay!) and Castle Game Engine (yay!)  check them out!

 

The network is still the primary bottleneck for graphics.

 

If there was a better way to compress video (I suggest X3D 5 or something, X3D is already part of the MPEG standard), then I could see server gaming coming to fruition.

 

Here’s a company , OnLive that went belly up selling cloud gaming (does no one remember?), worked similar to NetFlix:

 

https://www.youtube.com/watch?v=fOlgpK9ME8g

 

Obviously, Google is looking for more ad revenue and to destroy MS and Sony, and MS is trying to block them from eating into Xbox revenues.  I’m surprised Sony isn’t suing both of them into the ground.

 

I get 5Mb/second here in Iowa. I guess that’s not fast enough! They should probably wait for the rollout of 5G and sell to handheld cellphone devices. And 5G will reach my town when? We can’t even get cell coverage at the elementary school parking lot barely.

 

I am also wondering what Stadia and xCloud platforms will really be used for?  NSFW?  Note that X3D supports the up and coming international standard for Humanoid Animation (grin).  Now you see our strategy?  We’re doing a end-run-around of all the porn video and image sites and Second Life to provide standard and photographed, scanned, imaged, animated human avatars to doctors who have the money to shell out for fancy HP and Dell workstations (CLIENTs) in their offices and at home.

 

Work and entertainment, not games. Remember that iam here.  We want the oculus-web users ultimately I would say.  So put your pretty penny on React and X3D working well together, or React3D. No one is going to do a much with A-Frame is my prediction. High end social users will remain in walled gardens like Facebook, High Fidelity and Second Life—they give you the whole package, unfortunately, the last two are primarily C# I believe (no 3D format to compete with).  So what to look for is 3D developments from Facebook on their website.  That’s the future of 3D on the web, sadly.  We’ll all be datamined by the number of human avatars we’ve tried on in a month.

 

I suggest we get an X3DOM/X_ITE document on this website, really quickly, glTF is already up:   <https://dwqdaiwenqi.github.io/react-3d-viewer/site/#/GLTF> https://dwqdaiwenqi.github.io/react-3d-viewer/site/#/GLTF if we want to be part of react3d and the future of the 3D web. LOL

 

In other words, Google will serve up video, not 3D or VR (too much video).

 

I don’t think that X3D taxes the graphics card too much unless you open 4-5 canvases at a time, all with animation.   You can probably tweak a libraries WebGL code to reduce polygons if necessary.

 

And please tell me why I should upload my businesses data to google just to be shown in stadia?

 

Have you ever worked in a business?  Some people don’t want to upload their critical business data to the cloud.

 

Does anyone have a review of USD format from Pixar and Apple and whether it will be a FBX killer?

 

I believe X3D can be done on the server as well, the same way you would do a game on the server.

 

John “I use High Fidelity!  You use Facebook?  Ew!”  Carlson

 

Sent from Mail <https://go.microsoft.com/fwlink/?LinkId=550986>  for Windows 10

 

From: iam <mailto:iamhereintheworld at gmail.com>  here
Sent: Saturday, April 27, 2019 12:06 PM
To: X3D Graphics public mailing <mailto:x3d-public at web3d.org>  list
Subject: [x3d-public] Is there even a *case* for client-side 3D at all??

 

Hi,

I've posted to this list before, I'm a newbie to X3D and have been (off and on) trying to get to grips with it. However, as yall may be familiar with, Google has announced the upcoming launch of their streaming gaming service, Stadia, while MS has something called xCloud. So, what I'm wondering about is, after the advent of things like this, with God knows how much power available on the server-side, *is there even a case* for things like X3D, where the number crunching will be done on the CLIENT??

It all boils down to how many folks have high-end horsepower in their homes, and how many do not, and I can safely say that here, in my country, there are HUNDREDS OF MILLIONS of people who don't have say, a spanking Nvidia 2080 on their tables at home! But I don't have hard data on the overall situation, the *worldwide* scenario.... (btw, *data* here is dirt cheap, in case anyone wondered.)

 

What's the consensus *here*, on this list? I have great respect for what you guys *DID* with X3D, but....if there's no case for actually *using* it because it gets obsoleted by the likes of Stadia, then....it'll be a sad day, I guess. Of course we don't know what Google's *pricing* will be, but....that's hardly a lifeline. Also, these are not the only 2 there are - check something like *this*: www.liquidsky.com , which looks pretty nice.

 

So - is X3D doomed? :(

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://web3d.org/pipermail/x3d-public_web3d.org/attachments/20190430/b5121603/attachment-0001.html>


More information about the x3d-public mailing list