Which do you think will win? Here’s an experiment showing ways that math in green can’t keep up with Ruby.
It’s important for people learning to use flowstone to get their minds around what flowstone’s math can and cannot accomplish. A moderator at flowstone, who will often get into trouble for yelling at people due to a condition that he himself labels as autism decided to post this at the DSProbotics.com forum site:
Spogg wrote:I don’t suppose you could post the lyrics…?
I’ll do some rummaging. I did digitise at least some of them only a few years ago, and that song is likely to have been one of them. In fact it ties in quite well with what @marcelodantas was just saying – the reason for me digging them out was for evidence of my mental state earlier in life to present at my autism assessment. The psychologist who diagnosed me found them remarkably useful, and I was also astonished at how clearly I had understood my autistic traits and their consequences decades before it was even suggested as a potential diagnosis.
And that’s all well and good, if he has a condition, I understand; it’s really not a problem; but what is a problem is that he’s made the experience for paying customers less enjoyable. And ultimately, he makes the company seem less credible. Here’s what he said about this experiment.
email@example.com wrote:It’s really difficult to keep up with Ruby. Ruby’s math is much more precise
Yes, Ruby uses double-precision floats throughout (as you have already been told).
Green uses single precision, and you can’t add precision without adding bits (as you have already been told).
You are simply re-iterating the exact same nonsensical theories about numeric precision as you have been on your many other similar threads. There is nothing further to say about the topic which myself, Chris, Michael, and other haven’t already pointed out to you. You have also been pointed in the direction of comprehensive learning resources on this topic – in particular the documents relating to the IEEE754 standard.
Thread locked. I will give you one last chance – reasonable questions to help you understand floating-point arithmetic may be taken, but if you create any more threads about your nonsensical “precision-and-a-half alchemy”, the consequences outlined in your final formal warning will be invoked.
He suggests that there’s another two tests that will make sense to you. But those tests don’t seem to reflect the same experience when using green math/float math. And that’s what I found so misleading. I’d looked at this schematic time and time again, and I was wondering; what is the workaround? Maybe by using multiplication or division in a slightly different way, because as I understood; single precision has an accuracy below 2, so i decided to make My own test, this one; that did not use a static number. Something that makes more sense to everyone (including myself).
And here’s a second version of that test by another user, with the alias of “chrishooker”.
second file for testing accuracy of decimals with float math
So, both of these individuals complain about the failings in math in flowstone, but totally overlook Ruby. And impressionable newcomers to the product will follow their instructions; but instead of complaining about math, do they use ruby? They don’t even make reference to the fact that their problem posed could be solved by Ruby in a hot minute. But instead it’s more fun to just whine and show a sad devotion to the old synthmaker .osm files. I felt it made more sense to show the comparison between multiplication and divide against ruby definitively, instead of preaching some age old devotion to an old product that has no hope of solving the issue at hand.
Until you see math like this; you feel so secure in the use of green math and it is hard to accept that Ruby is the better level of math to be used for filters. And this is due to the fact that so many of the schematics relied upon, cannibalized, and utilized do not use ruby because they are from an earlier version of flowstone called synthmaker. There are a few testing examples, but they involve quite abstract thinking which often involves using something called blue math and stream math which is a different thing altogether.
And the really interesting thing is, in this test; I’ve replaced, in some instances all instances of multiplication with division and multiplication from division so that you can compare.
The synthmaker site was shut down, and all of those forum posts were backed up, meaning that the old schematics were saved, and they’ve been a staple. But again, without realizing the lacking that float math has and to what extent there really won’t be the sense of urgency instilled in those new to flowstone who might be more comfortable with the idea of using simple mathematical devices known as primitives that are basic and do not require extra “coding” learning.
You see how he said many times re-iterated? Yet, to a beginner; would they have read all of these posts? and would they know anything about float level math and its failings? No. His perception is: that those new to the forum will be thinking in a similar way to his own mindset. There are no such learning documents that would make any sense to a beginner whatsoever regarding these complex ideas.
He often forgets, flowstone is not a club for “pros” it’s a forum for “beginners” who are not pros. If they knew anything about programming they would be using C+ likely and would not be using it. (however, flowstone it is a nice cheap alternative). Likely in the next few weeks he’ll be doing his usual apologizing for his psychosis on the forums but I don’t need that nonsense, lol. I’m a programmer not a psychologist. The Flowstone forum is a place for non-pro’s to talk about programming in a relaxed easy, set their own pace level.
And here’s something another user pointed out that provides something interesting for math’s accuracy in the way of “blue” code:
there are stream doubles in the codebox in Alpha,
please check if you would like to use them with more versatility,
And that’s great, but I imagine ruby is still faster for the double precision math. And hopefully there’s a way to output double precision from ruby (and then to code) also; I hope this is in the new version so that you can save a great deal of memory. Blue and ASM can use too much versus green and ruby math :).
As far as I know; this is the only standardized test that shows exactly how inaccurate green math is. Doing things like going through (minute) various instances of division and multiplication is very common, but what do you do when it throws off your math.
Enjoy this nice test 🙂
You must log in to submit a review.