Tutorials, wiki and all the great good

A place to discuss everything related to Newton Dynamics.

Moderators: Sascha Willems, walaber

Tutorials, wiki and all the great good

Postby shybovycha » Wed Dec 16, 2015 4:22 pm

Hello there!

It's been a while since I was here. And what's more sad than that is the fact, Wiki tutorials page was not updated for four years now...

Image

And when I visited the Github Wiki - there are just links to the old Wiki...

"That's not good" - was what I thought, so I started writing and today I updated the old Wiki page, adding my brand new tutorial on Newton, Irrlicht and all the stuff. <off-topic mode>Would be happy for any of your feedback on that!</off-topic mode>

But the disturbing question still remains: how popular Newton GD is now? Does the reduced interest in this awesome engine caused by the growing popularity of Unity, Unreal Engine and the friends?
Image
User avatar
shybovycha
 
Posts: 52
Joined: Fri Oct 23, 2009 6:15 am
Location: Poland

Re: Tutorials, wiki and all the great good

Postby Julio Jerez » Thu Dec 17, 2015 11:13 am

shybovycha wrote:Hello there!
But the disturbing question still remains: how popular Newton GD is now? Does the reduced interest in this awesome engine caused by the growing popularity of Unity, Unreal Engine and the friends?


That might be a small contribution, but the rejection of the engine but the general public begun long before that. and it has to do with a negative campaign started but some individuals.

today the is a very reduce set of individual remind engine users, and from time to time I get some professional institutions sending me private message telling they disappointed of the engine tey are using and the want to try Newton.
One problem is that they still come in total distrust because they were burned and they continue believing that the will get burned. So they are very skeptical and they ask me to convince them.

One last such company was a Geological exploration Team from Canada, they do not want to be seen as forum member but the ask me to give then papers that show how Newton works.
I am not doing that anymore, I did it for 11 years now and it has not work, so continuing is just insanity.
It is so bad that even people who use the engine are afraid to say they use it because the get a back slash from people that do not like Newton not because they use but because what they hear?

In the end, Open Source or not, adding new features, writing the integration for other people, or me writing plug ins for other engines, the results are Negative anyway.
so yes the engine was rejected and it was a failure that cost me a great deal.
Julio Jerez
Moderator
Moderator
 
Posts: 12249
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: Tutorials, wiki and all the great good

Postby FSA » Thu Dec 17, 2015 5:02 pm

So does it mean Newton ist dead? I'm still using it.
User avatar
FSA
 
Posts: 318
Joined: Wed Dec 21, 2011 9:47 am

Re: Tutorials, wiki and all the great good

Postby Julio Jerez » Thu Dec 17, 2015 6:24 pm

Not is not dead, people are using it, but there are very few new comers.
I will always, work and fix problems, but I have not more expectations like I use to before.

for 99 out of one hundred the propaganda has usefully convex people that a Game Physic Library can doe now have to be correct.
I give you an example, for almost 10 years new support arbitrary scaling collision shapes, since last 1.53, I am reading the Unity 5.3 engine info, and I saw a video that recently PhysX 3.xx now supports baking Free Scaling, for some reason that seems a powerful feature in Unity.
Basically some that we do for a very long time is neglected, but we the competition cache up, they give a fancy name and come how is a powerful feature.
Same go for stuff like Vonrnoi Fractal, Closet distance, et.
Julio Jerez
Moderator
Moderator
 
Posts: 12249
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: Tutorials, wiki and all the great good

Postby shybovycha » Fri Dec 18, 2015 7:38 am

Marketing, damn it...

Image

I have an idea on how to make audience a bit larger, but it needs to be reviewed.

As far as I know, even popular graphic engines now became pessimistic about their future - both Irrlicht and Ogre3D have been updated really slowly and mostly with small bugfixes, but no major releases have been there since I lastly used them (four years!..).

Check this out: OGRE 3D, stable version is 1.9.0 (22 November 2013); Irrlicht, last major version is 1.8 (08 November 2012).

And the thing, which is really attractive in all those mainstream gamedev tools (like Unity or Unreal Engine) is the effort, needed to create an MVP (minimum-value product).

From what I've seen, people now try to simplify development as much as possible (not game development only). And production of a simple game looks literally like this:

1. install engine
2. download/purchase assets
3. put them on a level
4. write a few scripts/set a few properties for your assets (all these are done in WYSIWYG editor)
5. click Build for Android/WebGL/OS X/Windows/XBox
6. distribute

Whilst when you are making a game without using the all-inclusive toolset, you will probably run into lot of troubles, writing same basic things, as everyone else did for many years (like putting together YOUR_GRAPHICS_ENGINE and YOUR_PHYSICS_ENGINE, adding scripting and user input, etc.).

I know, this is crazy idea, too much for one person or even one small community, but I think this is the way of promoting Newton:

may be it's time for Newton to get involved in a full-stack game development platform?

(I called it "engine", "toolset", "game development platform", but this is all about the same all-inclusive tool, like Unity)
Image
User avatar
shybovycha
 
Posts: 52
Joined: Fri Oct 23, 2009 6:15 am
Location: Poland

Re: Tutorials, wiki and all the great good

Postby d.l.i.w » Fri Dec 18, 2015 2:53 pm

No, just no.
Seeing how much manpower would be required to make a reasonable "toolset", it would be a big waste of time with little chance of success.
There already are many frameworks for game creation and all of them have a hard time to compete with the few big ones like Unity. There is really no reason to create one more, especially because it's not the "quickly clicked together" kind of games that would really promote a physics engine. On the contrary - it might even look like Newton is suited for this kind of games only.
Actually mobile phones / Steam is flooded by such low to medium quality games, I don't think there is anyone really excited about the next thousand to come.

For high quality games a really powerful framework is needed and I doubt you could reach that level easily. Look at Ogre or Irrlicht, both are widely known and have (had?) a big community. Still very few widely known commercial games use those. Rather the development pace slowed down considerably and both have trouble to keep up with graphics evolution. I guess this is true for Bullet, too.

Even putting that aside, in the end - and because of marking - it's not necessarily good quality that counts, but big names and HYPE.

Back to Newton:
While this may sound very negative, I strongly believe that Newton has big potential and can easily compete with the "big" physics engines technically-wise. So IMO it's more important to stay focused and continually evolve the Newton library (just like in the past), than trying to create "the next big thing".

Usually a good way to attract more users is good documentation and clear instructions on how to get started. This is something that could be improved.
The red links in the Newton wiki at Github are not optimal, big parts of the (old) wiki are out of date. There is little to no reference to the current Newton version 3.x.

As a side note:
It might be irritating for some to find Newton under "MADEAPPS" on Github, asking themselves what the relation is.

Certainly well made integration for some game engine / frameworks like OgreNewt also can attract users. Maybe integration into Unity would be helpful, too.

Lastly:
It might not be that obvious, but one big strength of Newton is the clean C API, which exposes the whole core functionality. None of the other "big" physics engines have this.
This is not so much important for people programming in C(++), but major advantage for anyone using any other programming language.

(And there are people out there that still use ODE instead of Bullet because of this!)
d.l.i.w
 
Posts: 81
Joined: Mon Sep 26, 2011 4:35 am

Re: Tutorials, wiki and all the great good

Postby JoeJ » Fri Dec 18, 2015 5:57 pm

shybovycha wrote:may be it's time for Newton to get involved in a full-stack game development platform?


Doing this would mean: Spend all time to copy something that is already there, instead of inventing and improving worlds best physics engine - no good trade and also it hurts existing Newton users.

It would make more sense to make plug ins for UE4 and Unity.
A robust Interaction model like in Penumbra should convince a lot of their users, even if Epic / Unity themselves might not be happy at first.
An out of the box interaction model would be my priority, because:
* Lots tried to copy, no one got it right.
* It requires stable physics - Newtons selling point number one.
* Starting UE4 editor means: Play an Ego Shooter. 1st person games really need physics interaction, just shooting becomes boring. Read some internet comments and you see lots of people are bored by CoD15 or AC10 - and those bored people are the arget audience of indie devs, and indie devs are 99% of UE and Unity users.

If an indie dev can download Newton from its market place, and immediately can pick up / stack bodies, pull a rope etc. and all is stable, he will love it.
User avatar
JoeJ
 
Posts: 1453
Joined: Tue Dec 21, 2010 6:18 pm

Re: Tutorials, wiki and all the great good

Postby FSA » Sat Dec 19, 2015 6:49 am

If an indie dev can download Newton from its market place, and immediately can pick up / stack bodies, pull a rope etc. and all is stable, he will love it.


Totally. Maybe at first Newton should provide some demos, like stable picking, bullet interaction (for ego shooter stuff) and not the technics so much.
Integrating Newton in Unity / UE is the best thing I heard.
User avatar
FSA
 
Posts: 318
Joined: Wed Dec 21, 2011 9:47 am

Re: Tutorials, wiki and all the great good

Postby shybovycha » Sat Dec 19, 2015 12:41 pm

I was not saying "let's develop a game engine with Newton" - I was talking about involving or encapsulating/plugging in Newton into some of engines which are already there. I can imagine the effort, needed to create an alternative to those tools like UE or Unity. But it would be better showing off some exciting demos like "Shooter with Newton in Unity" or "Flight simulator with Newton in UE", rather than complicated "all-demos-at-once" app, doesn't it?

This idea appeared in my head suddenly, when I came to the refactoring part of my tutorial - it turned out, after five or six chapters all I got is a bunch of code, with almost no architecture, which allows to create and manipulate a few objects and allows for only two physic shapes - sphere and trimesh (static only). At that point, my tutorial' first paragraph, "motivation" looks insane...
Image
User avatar
shybovycha
 
Posts: 52
Joined: Fri Oct 23, 2009 6:15 am
Location: Poland

Re: Tutorials, wiki and all the great good

Postby Julio Jerez » Mon Dec 21, 2015 2:09 pm

letter123 wrote:Totally. Maybe at first Newton should provide some demos, like stable picking, bullet interaction (for ego shooter stuff) and not the technics so much.
Integrating Newton in Unity / UE is the best thing I heard.

I agree, this sound like a good idea, on the Holidays break I am going to read this doc.
http://docs.unity3d.com/Manual
see if I can get the basic of Unity and then see if I can make a plugin.
I registered and open an account, and I intellect the engine.
so far for what I can see Unity is very clean an well thought interface.
I like the editor better that Unreal.
I need to do my learning curve time, but I hope I can get some going.

what I found out is that single person project do no really work. I can not get people involved then I till be the same story as Ogre, Irrlitch and Open Scene Graph
few people become fund of the and the use for their personal project, and not one extend it, so in the end it dies a quite death.

edit: so far for what I can see Unity is very well designed and very intuitive and not overbearing like UE. I like that the have these assets packages that can be used by the end user.
This is actually a big help, I remember when I did the GameStudio plug in, it became very popular because they has similar features that the common user was able to apply the engine and no feel like it was a paste stuff.
so far so good.
Julio Jerez
Moderator
Moderator
 
Posts: 12249
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: Tutorials, wiki and all the great good

Postby Julio Jerez » Mon Dec 21, 2015 2:26 pm

how many people are Unity Users?
Julio Jerez
Moderator
Moderator
 
Posts: 12249
Joined: Sun Sep 14, 2003 2:18 pm
Location: Los Angeles

Re: Tutorials, wiki and all the great good

Postby shybovycha » Mon Dec 21, 2015 5:25 pm

Julio Jerez wrote:how many people are Unity Users?


I started my learning yesterday =)
Image
User avatar
shybovycha
 
Posts: 52
Joined: Fri Oct 23, 2009 6:15 am
Location: Poland

Re: Tutorials, wiki and all the great good

Postby Stucuk » Thu Jan 07, 2016 9:19 pm

A Unity plugin would be a good idea as it seems to be the main engine people use these days. Unreal seems to be chosen more by the AAA studios with Unity being picked by the indy's.

The lack of information on the Wiki could be turning people away. If i was looking at two physics engines and one had documentation and one didn't i would more likely pick the one with documentation even if it was the worst of the two.
User avatar
Stucuk
 
Posts: 801
Joined: Sat Mar 12, 2005 3:54 pm
Location: Scotland

Re: Tutorials, wiki and all the great good

Postby lperkins2 » Wed Jan 20, 2016 3:38 am

The lack of documentation is a huge issue. I don't mind reading source code once in a while when I'm not certain how things actually work, but reading it for everything is tedious. Even just getting the comments from Newton.cpp into Newton.h would help tremendously. If I use pre-built libraries, or clean up /usr/local/src, I won't have easy access to Newton.cpp, and I only even discovered the comments in it accidentally.

I could probably migrate the comments into Newton.h if you would like, but I think more than that is needed. What I think is most needed is some minimal examples, as well as an overview of the features and how to use them. Specifically, a demo that creates a world, adds a couple of boxes on an intercept course, lets them collide and prints something about the collision would be a good start.

I know there is an impressive demo application, the problem is that there's a lot of cruft in there that has nothing to do with the physics engine (keyboard/mouse input, rendering context, menus, other demos), so you have to sift through the application source (which doesn't even build reliably on linux) to figure out what each demo is actually doing.

The actual existing tutorial files are a little better, but the CNewtonTutorials are still actually using C++. All in all, between the lack of documentation, the lack of users on the forums, the outdated make files, and things like the C tutorial being in C++, it feels like a dead project, which decreases the odds of a developer choosing it.
lperkins2
 
Posts: 39
Joined: Fri Jul 03, 2015 4:16 am

Re: Tutorials, wiki and all the great good

Postby godlike » Wed Jan 20, 2016 6:20 am

lperkins2 wrote: I could probably migrate the comments into Newton.h if you would like, but I think more than that is needed.


I proposed something similar to this some time ago. Add doxygen comments to Newton.h. If there is a traction to something like this I'd love to help out.
User avatar
godlike
 
Posts: 58
Joined: Sun Mar 16, 2014 3:48 am

Next

Return to General Discussion

Who is online

Users browsing this forum: No registered users and 14 guests

cron