Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.javascript > #32609 > unrolled thread

Field Collisions Anime...

Back to article view | Back to comp.lang.javascript

Contents

  Field Collisions Anime... "Chris M. Thomasson" <invalid@invalid.invalid> - 2017-02-11 20:46 -0800
    Re: Field Collisions Anime... jonas.thornvall@gmail.com - 2017-02-11 23:00 -0800
      Re: Field Collisions Anime... "Chris M. Thomasson" <invalid@invalid.invalid> - 2017-02-13 16:15 -0800
    Re: Field Collisions Anime... jonas.thornvall@gmail.com - 2017-02-12 01:27 -0800
      Re: Field Collisions Anime... "Chris M. Thomasson" <invalid@invalid.invalid> - 2017-03-05 18:47 -0800
    Re: Field Collisions Anime... "Michael Haufe (TNO)" <tno@thenewobjective.com> - 2017-02-12 07:59 -0800
      Re: Field Collisions Anime... "Chris M. Thomasson" <invalid@invalid.invalid> - 2017-02-12 20:00 -0800
        Re: Field Collisions Anime... "Chris M. Thomasson" <invalid@invalid.invalid> - 2017-02-14 18:16 -0800
          Re: Field Collisions Anime... "Chris M. Thomasson" <invalid@invalid.invalid> - 2017-02-14 18:27 -0800

#32609 — Field Collisions Anime...

Here is the non-interactive pre-alpha test origin of a "collision 
avoidance system" based on vector fields. Its meant to be a fun, 
hopefully interesting, experiment. So far, the damn field lines just 
want to help the vehicles in circles avoid crashes. In this crude setup, 
a collision will make the anime flash, and tangentially move the 
colliding agents out of the way. The crude online example is here:

http://funwithfractals.atspace.cc/ct_gf_explore
(does this work for you at all?)

Fwiw, I first posted about this here:

https://plus.google.com/101799841244447089430/posts/JmbpLWvdJEf
(has a screenshot of the program...)

Can you please give it a go? Is it slow as shi%? Crap?

Any radical bugs in the JavaScript? ;^o

Thanks.

[toc] | [next] | [standalone]

#32611

Den söndag 12 februari 2017 kl. 05:47:02 UTC+1 skrev Chris M. Thomasson:
> Here is the non-interactive pre-alpha test origin of a "collision 
> avoidance system" based on vector fields. Its meant to be a fun, 
> hopefully interesting, experiment. So far, the damn field lines just 
> want to help the vehicles in circles avoid crashes. In this crude setup, 
> a collision will make the anime flash, and tangentially move the 
> colliding agents out of the way. The crude online example is here:
> 
> http://funwithfractals.atspace.cc/ct_gf_explore
> (does this work for you at all?)
> 
> Fwiw, I first posted about this here:
> 
> https://plus.google.com/101799841244447089430/posts/JmbpLWvdJEf
> (has a screenshot of the program...)
> 
> Can you please give it a go? Is it slow as shi%? Crap?
> 
> Any radical bugs in the JavaScript? ;^o
> 
> Thanks.

It work fine for me, but as you say flashes a bit.

[toc] | [prev] | [next] | [standalone]

#32628

On 2/11/2017 11:00 PM, jonas.thornvall@gmail.com wrote:
> Den söndag 12 februari 2017 kl. 05:47:02 UTC+1 skrev Chris M. Thomasson:
>> Here is the non-interactive pre-alpha test origin of a "collision
>> avoidance system" based on vector fields. Its meant to be a fun,
>> hopefully interesting, experiment. So far, the damn field lines just
>> want to help the vehicles in circles avoid crashes. In this crude setup,
>> a collision will make the anime flash, and tangentially move the
>> colliding agents out of the way. The crude online example is here:
>>
>> http://funwithfractals.atspace.cc/ct_gf_explore
>> (does this work for you at all?)
>>
>> Fwiw, I first posted about this here:
>>
>> https://plus.google.com/101799841244447089430/posts/JmbpLWvdJEf
>> (has a screenshot of the program...)
>>
>> Can you please give it a go? Is it slow as shi%? Crap?
>>
>> Any radical bugs in the JavaScript? ;^o
>>
>> Thanks.
>
> It work fine for me, but as you say flashes a bit.
>

Thank you for trying it out. The flashes are the points I need to 
refine, wrt choosing other field lines that do not produce collisions.

;^)

[toc] | [prev] | [next] | [standalone]

#32614

Den söndag 12 februari 2017 kl. 05:47:02 UTC+1 skrev Chris M. Thomasson:
> Here is the non-interactive pre-alpha test origin of a "collision 
> avoidance system" based on vector fields. Its meant to be a fun, 
> hopefully interesting, experiment. So far, the damn field lines just 
> want to help the vehicles in circles avoid crashes. In this crude setup, 
> a collision will make the anime flash, and tangentially move the 
> colliding agents out of the way. The crude online example is here:
> 
> http://funwithfractals.atspace.cc/ct_gf_explore
> (does this work for you at all?)
> 
> Fwiw, I first posted about this here:
> 
> https://plus.google.com/101799841244447089430/posts/JmbpLWvdJEf
> (has a screenshot of the program...)
> 
> Can you please give it a go? Is it slow as shi%? Crap?
> 
> Any radical bugs in the JavaScript? ;^o
> 
> Thanks.

But you can't see the code from browser, but i seem to remember something similar getting stucked in a loop toggle forw, back not able to leave. Because the conditions for the loop.

[toc] | [prev] | [next] | [standalone]

#32696

On 2/12/2017 1:27 AM, jonas.thornvall@gmail.com wrote:
> Den söndag 12 februari 2017 kl. 05:47:02 UTC+1 skrev Chris M. Thomasson:
>> Here is the non-interactive pre-alpha test origin of a "collision
>> avoidance system" based on vector fields. Its meant to be a fun,
>> hopefully interesting, experiment. So far, the damn field lines just
>> want to help the vehicles in circles avoid crashes. In this crude setup,
>> a collision will make the anime flash, and tangentially move the
>> colliding agents out of the way. The crude online example is here:
>>
>> http://funwithfractals.atspace.cc/ct_gf_explore
>> (does this work for you at all?)
>>
>> Fwiw, I first posted about this here:
>>
>> https://plus.google.com/101799841244447089430/posts/JmbpLWvdJEf
>> (has a screenshot of the program...)
>>
>> Can you please give it a go? Is it slow as shi%? Crap?
>>
>> Any radical bugs in the JavaScript? ;^o
>>
>> Thanks.
>
> But you can't see the code from browser, but i seem to remember something similar getting stucked in a loop toggle forw, back not able to leave. Because the conditions for the loop.
>

You can see the code from the browser.

[toc] | [prev] | [next] | [standalone]

#32616

Chris M. Thomasson wrote:
> Here is the non-interactive pre-alpha test origin of a "collision 
> avoidance system" based on vector fields. Its meant to be a fun, 
> hopefully interesting, experiment. So far, the damn field lines just 
> want to help the vehicles in circles avoid crashes. In this crude setup, 
> a collision will make the anime flash, and tangentially move the 
> colliding agents out of the way. The crude online example is here:
> 
> http://funwithfractals.atspace.cc/ct_gf_explore
> (does this work for you at all?)
> 
> Fwiw, I first posted about this here:
> 
> https://plus.google.com/101799841244447089430/posts/JmbpLWvdJEf
> (has a screenshot of the program...)
> 
> Can you please give it a go? Is it slow as shi%? Crap?
> 
> Any radical bugs in the JavaScript? ;^o
> 
> Thanks.

averaging about 16 FPS. Nearly 20% of the time spent on Garbage Collection.

There are a significant number of optimizations that could be done here, such as double-buffering, not drawing in Immediate Mode, and others.

Missing 'use strict' pragma

JSHint shows a number of minor issues you should handle: <http://jshint.com/>

Aside: You seem to write JavaScript like a C or C++ programmer which is a little strange.

[toc] | [prev] | [next] | [standalone]

#32617

On 2/12/2017 7:59 AM, Michael Haufe (TNO) wrote:
> Chris M. Thomasson wrote:
>> Here is the non-interactive pre-alpha test origin of a "collision
>> avoidance system" based on vector fields. Its meant to be a fun,
>> hopefully interesting, experiment. So far, the damn field lines just
>> want to help the vehicles in circles avoid crashes. In this crude setup,
>> a collision will make the anime flash, and tangentially move the
>> colliding agents out of the way. The crude online example is here:
>>
>> http://funwithfractals.atspace.cc/ct_gf_explore
>> (does this work for you at all?)
>>
>> Fwiw, I first posted about this here:
>>
>> https://plus.google.com/101799841244447089430/posts/JmbpLWvdJEf
>> (has a screenshot of the program...)
>>
>> Can you please give it a go? Is it slow as shi%? Crap?
>>
>> Any radical bugs in the JavaScript? ;^o
>>
>> Thanks.
>
> averaging about 16 FPS. Nearly 20% of the time spent on Garbage Collection.

Yup. Its because of all the damn arrays. The code uses them in the 
animation loop. Fwiw, this is in relation to:

https://groups.google.com/d/topic/comp.lang.javascript/dgakiMzhgv0/discussion

I did a benchmark, and the array version is slower.


>
> There are a significant number of optimizations that could be done here, such as double-buffering, not drawing in Immediate Mode, and others.
>
> Missing 'use strict' pragma
>
> JSHint shows a number of minor issues you should handle: <http://jshint.com/>

Thank you Michael. Will study the problems.


>
> Aside: You seem to write JavaScript like a C or C++ programmer which is a little strange.
>

Big time. I am actually not all that used to JavaScript. However, its a 
nice little medium to get some quick, hopefully portable, animations up, 
to see whats going on.

[toc] | [prev] | [next] | [standalone]

#32636

On 2/12/2017 8:00 PM, Chris M. Thomasson wrote:
> On 2/12/2017 7:59 AM, Michael Haufe (TNO) wrote:
>> Chris M. Thomasson wrote:
>>> Here is the non-interactive pre-alpha test origin of a "collision
>>> avoidance system" based on vector fields. Its meant to be a fun,
>>> hopefully interesting, experiment. So far, the damn field lines just
>>> want to help the vehicles in circles avoid crashes. In this crude setup,
>>> a collision will make the anime flash, and tangentially move the
>>> colliding agents out of the way. The crude online example is here:
>>>
>>> http://funwithfractals.atspace.cc/ct_gf_explore
>>> (does this work for you at all?)
>>>
>>> Fwiw, I first posted about this here:
>>>
>>> https://plus.google.com/101799841244447089430/posts/JmbpLWvdJEf
>>> (has a screenshot of the program...)
>>>
>>> Can you please give it a go? Is it slow as shi%? Crap?
>>>
>>> Any radical bugs in the JavaScript? ;^o
>>>
>>> Thanks.
>>
>> averaging about 16 FPS. Nearly 20% of the time spent on Garbage
>> Collection.
>
> Yup. Its because of all the damn arrays. The code uses them in the
> animation loop. Fwiw, this is in relation to:
>
> https://groups.google.com/d/topic/comp.lang.javascript/dgakiMzhgv0/discussion
>
>
> I did a benchmark, and the array version is slower.
>
>
>>
>> There are a significant number of optimizations that could be done
>> here, such as double-buffering, not drawing in Immediate Mode, and
>> others.
>>
>> Missing 'use strict' pragma
>>
>> JSHint shows a number of minor issues you should handle:
>> <http://jshint.com/>
>
> Thank you Michael. Will study the problems.

error 132 is serious, undefined variable.

[toc] | [prev] | [next] | [standalone]

#32637

On 2/14/2017 6:16 PM, Chris M. Thomasson wrote:
> On 2/12/2017 8:00 PM, Chris M. Thomasson wrote:
>> On 2/12/2017 7:59 AM, Michael Haufe (TNO) wrote:
>>> Chris M. Thomasson wrote:
>>>> Here is the non-interactive pre-alpha test origin of a "collision
>>>> avoidance system" based on vector fields. Its meant to be a fun,
>>>> hopefully interesting, experiment. So far, the damn field lines just
>>>> want to help the vehicles in circles avoid crashes. In this crude
>>>> setup,
>>>> a collision will make the anime flash, and tangentially move the
>>>> colliding agents out of the way. The crude online example is here:
>>>>
>>>> http://funwithfractals.atspace.cc/ct_gf_explore
>>>> (does this work for you at all?)
>>>>
>>>> Fwiw, I first posted about this here:
>>>>
>>>> https://plus.google.com/101799841244447089430/posts/JmbpLWvdJEf
>>>> (has a screenshot of the program...)
>>>>
>>>> Can you please give it a go? Is it slow as shi%? Crap?
>>>>
>>>> Any radical bugs in the JavaScript? ;^o
>>>>
>>>> Thanks.
>>>
>>> averaging about 16 FPS. Nearly 20% of the time spent on Garbage
>>> Collection.
>>
>> Yup. Its because of all the damn arrays. The code uses them in the
>> animation loop. Fwiw, this is in relation to:
>>
>> https://groups.google.com/d/topic/comp.lang.javascript/dgakiMzhgv0/discussion
>>
>>
>>
>> I did a benchmark, and the array version is slower.
>>
>>
>>>
>>> There are a significant number of optimizations that could be done
>>> here, such as double-buffering, not drawing in Immediate Mode, and
>>> others.
>>>
>>> Missing 'use strict' pragma
>>>
>>> JSHint shows a number of minor issues you should handle:
>>> <http://jshint.com/>
>>
>> Thank you Michael. Will study the problems.
>
> error 132 is serious, undefined variable.
>
>


Fixed it. The problem did not hit because that code path was not taken 
for this experimental online code. Hit refresh, and look at line 140.
____
     if (! aratio)
     {
         this.m_xstep = axes.width() / ((width > 1) ? (width - 1) : width);

         this.m_ystep = axes.height() / ((height > 1) ? (height - 1) : 
height);

140:    //this.m_xstep /= m_aratio;
         //^^^^^^^^^^^^^^^^^^^^^^^^

         this.m_xstep /= this.m_aratio;

         return;
     }
____

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.javascript

csiph-web