Yeah, that's right. Optimization wise, the calculation of force needs to be done only once for both particles. So here's some code. I downloaded giles' code and made my changes in red color. I've moved the particles.move() out of the main loop and created a separate loop. This doesn't increase the amount of calculation but will better conserve energy and momentum. Remember some of the particle pairs (+ and -) sometimes start taking off in one direction all of sudden. It's because you calculated for one particle and move it before you move the other affected particle. My way only moves particles after all calculation is over and hopfully will not get the sudden take off pairs anymore. I'd really like to have a final copy of Giles' program to show my students how someone not in the field of physics or physical science is able to do a superb job on a difficult physics problem.

I admit, as a physicist, I should have been more careful on this. My last code only calculates half interactions. I'd right if I were calculating potential energy. I need coffee, my lame excuse.

 for(int i=0; i<numparticles; i++)
 {
   
   float xaccel=0, yaccel=0;
   for(int j=i+1; j<numparticles; j++)
   {
     float ijdist = dist(particles[i].xpos, particles[i].ypos, particles[j].xpos, particles[j].ypos);
     float theta = atan2(particles[i].ypos-particles[j].ypos, particles[i].xpos-particles[j].xpos);
       
       
       
       if(ijdist>20) //attractive or repulsive forces depending on charges
       {
         xaccel+=  particles[i].charge * particles[j].charge * (force/(ijdist*ijdist)) * cos(theta);  
         yaccel+=  particles[i].charge * particles[j].charge * (force/(ijdist*ijdist)) * sin(theta);  
       }
       
       else if(ijdist>0.1)  
       {
         xaccel+=  (force/(pow(ijdist,4))) * cos(theta);  
         yaccel+=  (force/(pow(ijdist,4))) * sin(theta);  
       }  
       
       
       particles[i].xspeed+=xaccel;
       particles[i].yspeed+=yaccel;
       particles[j].xspeed-=xaccel;
       particles[j].yspeed-=yaccel;
  }
}

for (int i=0;i<numparticles; i++)
{
  if(particles[i].affected)
  {
    particles[i].move();
  }
  particles[i].render();
}

Have you tried running this? At first it seems to have more realistic behaviour than my last version. As you add more and more particles, some strange behaviour starts occurring. All the particles on the screen will suddenly jerk in one direction. Chains will form, but they don't move around spontaneously as in my last version.

When I conceived of this program I was sort of expecting the particles to sort of pack together like in an ionic solid, for example - sodium chloride (but in 2D).  Because this would minimise the potential energy. Maybe if you used the Leonard-Jones model and had the ionic attractions in the right proportion to these van der Waals -type attractions (and repulsions at close range), it would work out that way.

As an artist, I'm interetsed in creating systems with interesting behaviour.... so the physical model is a conceptual starting point, but then, if things "go wrong", I'm happy to pursue it in that direction (which is an aesthetic direction). I know my program doesn't have realistic properties as it seems to generate infinite amounts of kinetic energy. Even though there is a significant damping, which should cause the energy be dissipated, the particles continue to move and chase each other around - like they have an internal energy source. So they are more like living things in that way. If it was modelled in a physically accurate way, I would expect it to move towards some equilibrium position and then remain static. That could be very interesting - as you added more particles you could see the system re-adjust and reach a new equilibrium.

I'm glad that both of you're having some fun. The program doesn't have to do exact physics to be interesting. Yes, distant objects do interact and pull one another. On my side, I implemented simultaneous updating positions and r^4 repulsive forces. Two opposite charges with oscillate forever (true but not fun). Sometimes I can make two charge go around their common center while spining together. I'll use this next time I teach physics of electricity and magnetism. I also changed from random charge to positive/negative upon left/right mouse click.