Forum > General

Large or huge sets

<< < (2/14) > >>

Kays:

--- Quote from: Thaddy on September 10, 2019, 10:15:11 am ---Large or huge sets are often requested. […]
--- End quote ---
Meh, your code doesn’t support set of real. :P

Thaddy:

--- Quote from: Kays on September 10, 2019, 12:06:08 pm ---
--- Quote from: Thaddy on September 10, 2019, 10:15:11 am ---Large or huge sets are often requested. […]
--- End quote ---
Meh, your code doesn’t support set of real. :P

--- End quote ---
No, because that is not mathematically a viable option for real sets.
The TSortedHashset supports that to some extend...But not like real sets, like in a Pascal set. My code does.
You need to have a certain amount of bit patterns, just like normal sets.
(I know it requires a small, just a small, bit of computer science...)
Furthermore, I never pretend to have infinite scale, just larger scale if you run out of 256....If you do not understand that plz refrain from taking part of the discussion,
I just got two very informed answers that really help.

440bx:

--- Quote from: Kays on September 10, 2019, 12:06:08 pm ---
--- Quote from: Thaddy on September 10, 2019, 10:15:11 am ---Large or huge sets are often requested. […]
--- End quote ---
Meh, your code doesn’t support set of real. :P

--- End quote ---
I presume that comment was tongue in cheek as only countable sets can be represented digitally.  The set of reals is not countable.

jamie:
There is one thing I can say about @Thaddy, good of bad, he knows how to write bloat!
 ;)

marcov:
Oops. I only read quickly initially, and it seems not to be a value type, meaning you must change business code (and streaming!). Not really an option for me. I wonder if changing the class to a record<>,  and then hacking till it works might do it. However the problem is dimensioning the static data field to the size of the enum.

Hmm. maybe abusing an ansistring/dynarray. Automated but many aspects like streaming and " IN " operator cost magnitude, remains the same

Navigation

[0] Message Index

[#] Next page

[*] Previous page

Go to full version