Wikidata:Property proposal/factorizsation

factorisation edit

Originally proposed at Wikidata:Property proposal/Natural science

Withdrawn

Description	writing a number or another mathematical object as a product of several factors, usually smaller or simpler objects of the same kind (sample: 4 = 2*2). For natural numbers, this cover the decomposition of the number in its prime factors
Represents	factorization (Q188804)
Data type	Mathematical expression
Domain	e.g. integer
Example	42 (Q812996) → 237
Planned use	add to natural numbers
See also	Special:ListProperties/math

Motivation

Looks like we aren't really using the datatype: Special:ListProperties/math
--- Jura 11:03, 6 May 2018 (UTC)[reply]

Notified participants of WikiProject Mathematics
--- Jura 11:04, 6 May 2018 (UTC)[reply]

Discussion

Could be populated by bot? Arthur Rubin (talk) 11:19, 6 May 2018 (UTC)[reply]
- Some of it surely.
  --- Jura 11:20, 6 May 2018 (UTC)[reply]
Seems like something that is easily computed. What is the benefit of saving this on Wikidata? --Tobias1984 (talk) 11:54, 6 May 2018 (UTC)[reply]
- There seems to be some demand for it: @IvanP: Wikidata:Project_chat/Archive/2018/04#Prime_factors?.
  --- Jura 11:59, 6 May 2018 (UTC)[reply]
Support David (talk) 14:44, 6 May 2018 (UTC)[reply]
Comment Even if we wanted to host this, it's not clear to me that this is the best way to do it. Since the factors are separate things, we could link to them in separate triples (e.g. 42 (Q812996) prime_factor = 2, 42 (Q812996) prime_factor = 3, 42 (Q812996) prime_factor = 7. Wouldn't we want to link to items if there were already items for the prime factors (e.g. 42 (Q812996) prime_factor = 2 (Q200), 42 (Q812996) prime_factor = 3 (Q201), 42 (Q812996) prime_factor = 7 (Q23350))? Qualifiers could sort out multiple identical factors. --99of9 (talk) 01:41, 7 May 2018 (UTC)[reply]
- Interesting thought. It might not work well for factorizations of non-integers. Maybe we could link the factors - when possible - with a qualifier.
  --- Jura 06:48, 7 May 2018 (UTC)[reply]
  - How to deal with powers of primes? Lymantria (talk) 18:23, 21 May 2018 (UTC)[reply]
    - @Lymantria: using a qualifier. If there's not already an appropriate one, we could make one called multiplicity (Q2228257). --99of9 (talk) 01:12, 23 May 2018 (UTC)[reply]
  - It does seem tricky when trying to do an item based one. Retrieval is probably even more complex.
    --- Jura 19:20, 21 May 2018 (UTC)[reply]
    - @Jura1: I'm not sure what you mean by "It does seem tricky". Retrieval of factors with qualifiers would be trivial. Joining them into a string not much harder. --99of9 (talk) 01:15, 23 May 2018 (UTC)[reply]
      - I think the LUA code might be longer than doing factorization in LUA. So what's the approach for powers of primes?
        --- Jura 08:26, 23 May 2018 (UTC)[reply]
        It depends what the LUA is trying to do. If it wants to provide sitelinks to prime factors, my model certainly helps with that! Regarding powers, see my answer to Lymantria above. I agree with ArthurPSmith below that quantity (P1114) is a good existing qualifier to use. --99of9 (talk) 00:47, 24 May 2018 (UTC)[reply]
        I agree that using items and using quantity (P1114) to indicate the power (mandatory!) is a better way to do prime factorisation. Lymantria (talk) 06:10, 24 May 2018 (UTC)[reply]
Comment factorization of non-integers? Can you suggest some examples? Polynomials perhaps? Matrices? I don't think we have many (any?) items for specific polynomials in wikidata right now... I think I prefer 99of9's approach for this, which would work with polynomials and matrices too if we actually have items for them. ArthurPSmith (talk) 14:29, 7 May 2018 (UTC)[reply]
- Comment Factorization of non-rationals is problematic because of the question of the underlying ring in which to find the factors. Arthur Rubin (talk) 16:02, 8 May 2018 (UTC)[reply]
Comment Is this meant to be only prime factorisations, or not? Put it otherwise: should we have (2+i)(2-i) as factorisation of 5? And (sqrt{3}+sqrt{2})(sqrt{3}-sqrt{2})? Where would it end? Lymantria (talk) 20:28, 13 May 2018 (UTC)[reply]
- I'd start with prime factorization.
  --- Jura 20:31, 13 May 2018 (UTC)[reply]
- I re-wrote the description accordingly.
  --- Jura 11:41, 19 May 2018 (UTC)[reply]
Oppose Since this is now limited to integers and their prime factors I much prefer 99of9's proposal, specifically to create an item-valued "prime factor" property. ArthurPSmith (talk) 14:55, 21 May 2018 (UTC)[reply]
- If a more general one is preferred, I rephrased the description.
  --- Jura 15:03, 21 May 2018 (UTC)[reply]
No, that doesn't help, sorry. Still opposed - and please don't mark your own proposals as "ready". ArthurPSmith (talk) 20:20, 21 May 2018 (UTC)[reply]
- Would you address the problems raised above? Please avoid to comment just to comment. This isn't productive.
  --- Jura 21:14, 21 May 2018 (UTC)[reply]
@Jura1: your only response to my request above to "suggest some examples" of factorization of non-integers was to respond to limit it to "prime factorization" of the integers, to which I indicated 99of9's approach was definitely better. As Arthur Rubin and Lymantria pointed out above, there are many different possible factorizations if you are not sufficiently specific. If you have a rigorous mathematical definition of what you are trying to do here that leads to a singular mathematical formula for a specified domain (and it's not simply prime factorization of the integers, or more generally if it's not better solved by 99of9's approach), I might see a purpose to this property. Right now, it doesn't seem headed in a useful direction at all. ArthurPSmith (talk) 17:44, 22 May 2018 (UTC)[reply]
- There are two problems outlined with 99of9's hypothetical approach you seem to hypothetically prefer. Seems it isn't practically feasible.
  --- Jura 17:50, 22 May 2018 (UTC)[reply]
  - I have no idea what you are talking about. The only "problem outlined" I can see was the issue of multiple factors, which 99of9 explicitly addressed in their original comment - "Qualifiers could sort out multiple identical factors." Powers of primes are exactly that case also. I see no complication here. quantity (P1114) is the obvious qualifier to use. ArthurPSmith (talk) 19:40, 23 May 2018 (UTC)[reply]
Oppose In favor of 99of9's proposal for prime factorisation. Lymantria (talk) 06:11, 24 May 2018 (UTC)[reply]
@Jura1, Arthur Rubin, Tobias1984, IvanP: @99of9, ديفيد عادل وهبة خليل 2, Lymantria: ok, the above seems like a consensus... - see Wikidata:Property proposal/prime factor. ArthurPSmith (talk) 19:07, 24 May 2018 (UTC)[reply]
Comment I removed the "not done" as the closer didn't bother writing a conclusion and this proposal has some support.
--- Jura 09:19, 7 June 2018 (UTC)[reply]
@Jura1: sorry I thought that was obvious from the comments above. Then Oppose in favour of the alternative approach for which prime factor (P5236) was created − Pintoch (talk) 09:58, 7 June 2018 (UTC)[reply]
- It seems odd that you oppose when closing. This can give the impression you aren't neutral when closing discussions.
  --- Jura 10:01, 7 June 2018 (UTC)[reply]
  - @Jura1: I thought there was a consensus for closure, therefore I closed as not done (without voting). Now you dispute that, so I do not close anymore but vote instead. So I do not oppose when closing. That does not seem scandalous to me. − Pintoch (talk) 10:26, 7 June 2018 (UTC)[reply]