This week's book giveaway is in the OO, Patterns, UML and Refactoring forum. We're giving away four copies of Refactoring for Software Design Smells: Managing Technical Debt and have Girish Suryanarayana, Ganesh Samarthyam & Tushar Sharma on-line! See this thread for details.

Hi ranchers, i'm having fibonacci numbers in an array of type int[]. The numbers are up to 4million. Since I generated it with a loop, i have no idea of what the 3999998 element will be. On invoking println, i knew the answer was wrong because it printed a negative number. But i'm very sure it works for like the 1st 1000 elements or very much more. Thanks in advance!

Ofcourse Java can handle numbers greater than one billion. But an int is a 32-bit number, which can store values between -2^32 and 2^32 - 1, so you won't be able to store numbers larger than 2^32 - 1 = 4,294,967,295 in an int.

You could use a long which has 64 bits, but the Fibonacci sequence grows so fast that you'll quickly run out of the range of the long type too.

To store integer numbers with an arbitrary number of digits, you can use class BigInteger.

Shamsudeen Akanbi wrote:Hi ranchers, i'm having fibonacci numbers in an array of type int[]. The numbers are up to 4million. Since I generated it with a loop, i have no idea of what the 3999998 element will be. On invoking println, i knew the answer was wrong because it printed a negative number. But i'm very sure it works for like the 1st 1000 elements or very much more. Thanks in advance!

This sounds like one of those Euler Project problems.

A nice property to remember about the Fibonacci series is that F(n)/F(n-1) approaches the Golden Ratio: (√5+1)/2 ≈ 1.618...

Therefore, results are going to progress in powers of 1.618, which is larger than √2 (1.414...), and that means that they are going to more than double for every other result. So, in order to get F(3999998) you need a number that holds at least 2 million bits (and probably a lot more).

Luckily, there is a class that can hold such values easily: BigInteger.

What may not be so easy is waiting for the result of F(3999998)...but good luck in your quest.

Winston

[Edit] Actually, the "therefore" above may be wrong; but the rest of what I said is correct.

Isn't it funny how there's always time and money enough to do it WRONG?
Articles by Winston can be found here

Stevens Miller wrote:Nice analysis, Winston! I love to see people apply mathematics to computation. Not enough of that in our field, imho.

Cheers mate...although in my case it's more 'geek' than Maths (only got to A-level). I still like it though (when I can follow it ).

And I think your use of "therefore" is okay, fwiw...

And that's where my Maths fails. I believe (but I may be wrong here) that Fibonacci produces 'progressive closest integral fractions to G' and, if that's right, then the rest would be right by inference; but if not, I think it would be a lot more difficult to prove the "therefore" bit.

But I'd love to know if I'm wrong...and therefore right.

Winston Gutkowski wrote:I believe (but I may be wrong here) that Fibonacci produces 'progressive closest integral fractions to G'...

Actually, I think I may have worked it out for myself: If F(n) == F(n-1) + F(n-2), then it doesn't matter if progressive results of F(n)/F(n-1) produce the closest fraction to G or not; simply that they produce a closer one - and that I'm pretty certain they do.

Mathematicians, please put me out of my ignorance. My tools are teenage Maths and logic.

Shamsudeen Akanbi wrote:I got an error that: BigInteger(long) has private access in BigInteger. BigInteger total=new BigInteger(0);. And an arrow was pointed at 'n' in new.

Shamsudeen, have a look at the list of constructors available when you use the BigInteger class. BigInteger(long) isn't one of them. BigInteger("0") might get you going, but read the documentation before you assume that will work for you.

Stevens Miller wrote:We don't say "mate" much here on the American Atlantic Coast (well, not as a noun, anyway). Where are you located?

Currently: Brussels (for the last 8 years); but originally a Canuck, brought up on the 'Sarf' coast of England.

Before 'La belle Belgique', had split my life pretty much equally between the UK and Canada - with 18 months in the US (NYC, IL and MN) back in '79-80 as my re-intro to 'Norf' America.

So you'll forgive any confusion (comes out in my posts sometimes) when it comes to "nationhood". Personally, I see myself as a "Brit with other bits".

Stevens Miller wrote:Cool! How's the market for programmers out there? Need any help?

What, me personally, or my "nation" (UK, I guess) as a whole?

In general, I'd say that the economy over here isn't quite as healthy as yours; but there are jobs out there for people with the "right" skills.
What's "right"? Dunno. These days, it seems to involve a lot of TLA's on your CV; and it also helps if you ain't 55 (like me) and perceived as a dinosaur.

I'll be a programmer until they curl me up and chuck me in the furnace, job or no job; but if you're looking for pastures new, have a look at the ads - you guys can get the Times or the FT same day can't you?

Britain's a wonderful country; but be prepared for some changes if you do decide to make the hop. It's more prosaic, more wordy, and much, much smaller. And it's GREEN. Living on the East Coast you may be used to it, but I lived in Vancouver, BC for most of my 22 years in Canada; just south of 300-odd thousand square miles of pine forest. And believe me, you notice it when you're on the glide path into Heathrow.

I turn 54 in October, so I guess we can look down our withered old noses at these newcomers together.

I live in the glide-path for Dulles International, so I can relate.

Say, back to Java for a moment: I'm having a heck of a time with an interface I added to a .jar file that is already working well as a source of classes for a different project. I added an interface to the .jar, and now my main project can't find the interface when I compile the main project. Any pointers for me to some guidance?

Well, solved my own problem, sort of... The "bug" was actually a problem with how I had the NetBeans build.xml executing javah. If I had known I was seeing output from javah, I'd have fixed it right away.