Scratch

LiquidMetal wrote:

Why are multi-letter ops so hard to process?

I think it's because they evaluate only the stuff in parentheses in front, not behind and in front.

Multi-letter ops should be handled by 2 lists. One with all the digits and one with everything as numbers.

I.E.

List 1:
1
2
+
4
8

List 2
12
+
48

See? List one makes list 2 happen, and you process list 2!

applejack wrote:

*ahem*

applejack wrote:

We need error checking for divide by zero, imaginary numbers and other stuff!

look here:

http://i.imgur.com/nh6rn.gif

-LiquidMetal

applejack wrote:

LiquidMetal wrote:

applejack wrote:

*ahem*

look here:

Code:

http://i.imgur.com/nh6rn.gif

-LiquidMetal

But you're allowed to have negative exponents. x^(-y) = 1/(x^y), so it's valid unless x = 0, right?

Also, the commonly asked and never properly explained question: does 0^0 return zero or one?

Lastly, Liquid, it's faster to use this

n^m = 10^(m*log(n))

for exponents. 

Last edited by hpotter134 (2011-07-07 10:09:20)

0^0 = 1.  Why is that so hard to understand?

LiquidMetal wrote:

0^0 = 1.  Why is that so hard to understand?

Because 0^ anything else = 0.

if you need help:

0^0=1
0^not 0=0
something^0=1
positive^something=e^(m*log(n)) (faster than 10^ and log, tested)
negative^integer=e^(m*log(abs(n))), add minus if m is odd
negative^not integer=imaginary

that's probably all.

Anyone there?  I need to know the status of the current project- have we fixed all the bugs in the sub-expression finder?  Have we already built in the multi-ops evaluation, or should I do that? 

Anyway, glad to finally get started on this project- it's been awhile.

No, I think I have a couple bugs I was working on, and we still need the multi-ops. LiquidMetal made something for us to do those.

-i^2

amcerbu wrote:

I don't think that we should move on until what we have already works flawlessly.  Otherwise, it will be harder finding the bugs.

hpotter134 wrote:

applejack wrote:

-i^2

amcerbu wrote:

I don't think that we should move on until what we have already works flawlessly.  Otherwise, it will be harder finding the bugs.

Nice one applejack 

My idea for multi ops is to make a list of all "string" characters, aka:

Code:

ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz

and then:

Code:

if a letter of the input is contained in the Alphabet list,
     if the last item of tokens is "str"
          replace last of engine with join : with new letter
     else
          add it to the engine and add "str" to the token list.
     [end if]
     if the next letter of the input is NOT contained in the Alphabet list
         [run test to see if it's a function or constant or something else]
         replace last of tokens with [something else]
     [end if]

That looks familiar... I know: it's from your project! 

LiquidMetal wrote:

Wait is this for detecting multi-ops in a string or evaluation?

The code I posted recently was for multi ops.

@sci, yeah, it's from my project 

It's for detecting.  The actual evaluation process is simple- after having determined what each multi-op is, simply run the appropriate equation.