Bigger values of 'n' to try
Looks like this task is too easy on modern hardware for those with access to GMP.
I've found a more up to date link with some even bigger numbers for those who want to extend their solution.
The value of h[n] for the biggest of these (n = 1,247,677,915) has more than 3.2 billion digits!
Regrettably, Harry Smith who did much early work on the subject has now passed away. --PureFox (talk) 18:49, 18 August 2021 (UTC)
talk about a very delayed déjà vu
It wasn't after I backed up (archived) my juggler REXX program for this (draft) task (to my archive/backup hard drive) that I discovered a copy of my (draft) task for the this sequence (I had put it aside for some reason, and then forgot about it). The old draft was created in March, 2016. Oh well. Not the first time, and it won't be the last, either. -- Gerard Schildberger (talk) 22:54, 18 August 2021 (UTC)
- I was surprised that someone hadn't made a task out of this before because it's quite a well known number sequence with an interesting background. Is there anything in your original draft that might be worth adding now? --PureFox (talk) 16:10, 19 August 2021 (UTC)
- My original draft (from 2016) didn't have as many "big integers" for stress testing, but it did have a much prettier (HTML <math>) formulas: floor and square root symbols, for two. I had copies the use of the floor from a Wikipedia site, but Rosetta Code didn't render it properly (lots of big red error messages). It took me a while to get those symbols to display properly without crashing something in Rosetta Code. -- Gerard Schildberger (talk) 17:06, 19 August 2021 (UTC)
- I did wonder about making the formulas a bit prettier but I'm always reluctant to use fancy formatting in case it doesn't render properly on all the different browsers people use. Also I've never been keen on the 'floor', 'ceil' and 'integer part' symbols which you need to examine quite closely to make sure that they're not just a value in square brackets. --PureFox (talk) 18:18, 19 August 2021 (UTC)
- Yeah, as far as all the different browsers in use, it's hard to test them all, especially if I can't even install them. At this point, I'm pretty much stuck with FireFox, which is the only browser (for me, at this point on my old software) that is working). All the other common ones (popular ones) refuse to even install, my host operating system is too old. Even Microsoft I.E. (the one that came installed with my Microsoft Windows) no longer works. Well, it works well enough to tell me that it found an error somewhere (for some websites), a hit and miss situation. -- Gerard Schildberger (talk) 18:53, 19 August 2021 (UTC)
Error in task
a[k + 1] = floor(a[k] ^ 0.5) if k is even or a[k + 1] = floor(a[k] ^ 1.5) if k is odd
a[k + 1] = floor(a[k] ^ 0.5) if a[k] is even or a[k + 1] = floor(a[k] ^ 1.5) if a[k] is odd