Topic: NXT :: descendant of Bitcoin - Updated Information - page 84. (Read 2761645 times)

They should add a linux version. As it is Java, that shouldn't be a problem.

http://nxtra.org/nxt-wallet/

Mac vs PC is kind of a religious thing (as an example). Float vs. integer should not.

Mac vs. PC? Wait what?

I wonder if this is kind of a religious thing. To float or not to float.

But we don't talk about Mac vs PC here. It's science, Mr. White. Science, bitch!

Nxt nodes forging on a Raspberry Pi is a gimmick. Someone wake me up when we have a supercomputer running Nxt.

j/k

This discussion has been interesting to watch. I'm looking forward to what Ian has to say when he returns to the thread.

Is there any user friendly wallet?

I am about a week or so away. It is straight C, almost totally self-contained. The only external libs I use are pthread, curl, and maybe zlib

If you are able to review, I will be happy to send you advanced copy

James

Floating-point math is heavily used for financial calculations on supercomputers. If u say that FP is bad then say this to thousands scientists who use it every day.

when there is a least common denominator, like cents or satoshi, integers can do errorless calculations
for allocating coins from a pool, integers have errors representing fractions, especially cumulative sums of fractions
you need to look at the source data to determine which has better accuracy
maybe nodecoin pool calculations are not "financial program"?
maybe anything that doesnt have a least common denominator, is not a "financial program"?

Edit: Never using floating point seems to be like not taking full advantage of the hardware, which to me is not optimal. Just be careful, watch error rate, rounding, comparisons, etc. Yes, more work and possible to make mistakes, but this is programming, every line we write can have a bug! Again, if there is a least common denominator without overflow danger, use integers. Otherwise, need to think it through.

ciyam has had bad experience in past with floating point
now he hates floating point, he has cool demo that proves floating point is braindead
anytime he sees floating point, he remembers horrible past
sometimes floating point is not so bad
sometimes use cases, the cool demo does not apply
floating point, floats
fixed point, rigid
fractions must be approximated, by both
floating point has error rate
integers also have errors with fractions
need least common denominator to use integers without errors
64 bit integers will overflow if fractions used are small enough
nodecoin allocation fractions can be very small
nodecoin allocation might be very big
overflow cannot be avoided with 64 bit, without larger error rate than floating point
128 bit math is not in silicon
errors are pretty small anyway, why risk using software library when CPU does it in silicon
believe me (and CfB), or believe ciyam who has not even looked at my code that does floating point

where is this code?

So can someone help James with a review of the C code for multigateway?

multigateway has no floating point

Ill just hope that CIYAM didnt know this fact before getting us all riled up

Currently, I use floating in exactly one place. For calculating allocation of nodecoins. My usage has been approved by CfB. mthcl agreed with my math analysis. Others agreed that for my use case, there is nothing wrong with using floating point.

I make engineering decision based on tradeoffs of accuracy, implementation time, code stability. My assessment was to use 64 bit floating point for allocation. This is because floating point is just as accurate, if not more, when approximating fractions.

The reason integer math is more accurate for financial applications is because there is a least common denominator (satoshi, cent) and so all calculations can be done without errors. As long as you dont overflow.

With pool allocations, you end up with fractions without any reasonable least common denominator. That means you get errors representing it with 64 bit ints or 64 bit floats. However, floats are designed to dynamically adapt to the scale of the numbers while with integers you need a fixed multiplier. The fixed multiplier needs to deal with worst case while at the same time avoiding overflow.

So, using integers for this use case would not be any more accurate and instead open to overflow, or if to avoid overflow you get higher error rates.

The other reason that integer math is preferred in crypto is that all computers end up with the same result. This is important in synchronizing blockchain. however, my code is running on controlled servers and there is no need for synchronizing.

This "never use floating point" requirement seems to be dictating usage of inferior solutions in some use cases. I prefer to think about what I am doing and choose the appropriate solution.

If you have any math based feedback, I am eager to hear it. Superstitous, ban all floating point, is not really constructive.

James

P.S. I could get better accuracy if I used 128 bit floats and near perfect accuracy using 128 bits ints, but I do not trust the code libraries for doing 128 bit math as much as floating point silicon in Intel processors. In any case, do people really need their nodecoins allocated more accurately than .0000000000001 nodecoins?

I thought the joke was that he had a team.

Dammit, that is like the seventh time - glad that this day is over 

Actually he was joking. April 1st.

pls send me some nice message to the testnet acc: 9500913153933294022

- for the video