Thought not required

dmwit

I just recently got interested in vPython for doing graphics. I thought I'd look at the code they had as examples of good form, and found this little gem:

    	if pfc.x > self.boidflock[aboid].x:
            pfc.x = (pfc.x - self.boidflock[aboid].x)*self.FACTOR
        if pfc.x < self.boidflock[aboid].x:
            pfc.x = (self.boidflock[aboid].x - pfc.x)*self.NEGFACTOR
        if pfc.y > self.boidflock[aboid].y:
            pfc.y = (pfc.y - self.boidflock[aboid].y)*self.FACTOR
        if pfc.y < self.boidflock[aboid].y:
            pfc.y = (self.boidflock[aboid].y - pfc.y)*self.NEGFACTOR
        if pfc.z > self.boidflock[aboid].z:
            pfc.z = (pfc.z - self.boidflock[aboid].z)*self.FACTOR
        if pfc.z < self.boidflock[aboid].z:
            pfc.z = (self.boidflock[aboid].z - pfc.z)*self.NEGFACTOR

<font face="Times New Roman" size="4">I mean, it isn't terrible, but I probably would have just used
		<font size="2"><font face="Courier New">pfc = (pfc - self.boidflock[aboid]) * self.FACTOR</font></font>
</font>

script_man

@dmwit said:

I just recently got interested in vPython for doing graphics. I thought I'd look at the code they had as examples of good form, and found this little gem:

    	if pfc.x > self.boidflock[aboid].x:
            pfc.x = (pfc.x - self.boidflock[aboid].x)*self.FACTOR
        if pfc.x < self.boidflock[aboid].x:
            pfc.x = (self.boidflock[aboid].x - pfc.x)*self.NEGFACTOR
        if pfc.y > self.boidflock[aboid].y:
            pfc.y = (pfc.y - self.boidflock[aboid].y)*self.FACTOR
        if pfc.y < self.boidflock[aboid].y:
            pfc.y = (self.boidflock[aboid].y - pfc.y)*self.NEGFACTOR
        if pfc.z > self.boidflock[aboid].z:
            pfc.z = (pfc.z - self.boidflock[aboid].z)*self.FACTOR
        if pfc.z < self.boidflock[aboid].z:
            pfc.z = (self.boidflock[aboid].z - pfc.z)*self.NEGFACTOR

<FONT face="Times New Roman" size=4>I mean, it isn't terrible, but I probably would have just used
		<FONT size=2><FONT face="Courier New">pfc = (pfc - self.boidflock[aboid]) * self.FACTOR</FONT></FONT>
</FONT>

True, but the sample code in a lot of packages is not written for brevity; it's written to illustrate to a product-noobie how the thing works under-the-hood. Sample code is usually less than ideal for illustrative purposes.

dmwit

But... self.NEGFACTOR ??

themagni

Some coding styles state that you can't assume that the elements in a structure will assign with an assignment of the structure. In other words,

pfc = pfc - modification

is not the same as:

pfc.x = pfc.x - modification

pfc.z = pfc.z - modification

etc.

You've missed the big bug in the example code: What happens when pfc.value is the same as self.boidflock[aboid]?

kwan3217

@dmwit said:

But... self.NEGFACTOR ??

It seems likely that self.NEGFACTOR does not equal -self.FACTOR. This sure looks like part of a flocking algorithm, where a bunch of boids are following a leader. Maybe they want to make sure that the boids stay behind the leader, so they put a heavier correction on one side than on the other.

dasmb

Man, it really sucks when you report a WTF and everybody defends it, doesn't it?

Personally, I prefer the example code...it is much more explicit, and exploiting relatively obscure syntactical decisions is a sure way to cause confusion. Impress people by doing it well, not by doing it short.

dmwit

No, self.NEGFACTOR is most definitely set explicitly to -1 * self.FACTOR. =(
~d