Talk:The multi-faceted reach of the OEIS
Coding music and other artwork
I think it would be nice to collect information, opinions and maybe reach concensus on various questions regarding OEIS records of non-mathematical artwork, such as:
(3) Paintings or other graphical artwork
Some of my own views:
Concerning (1): I understand it is nice to have some "MIDI" type of coding music (to allow listening to it), but at the same time this makes the sequence somehow arbitrary and obfuscates (IMO) the mathematical content which definitely is present in any musical work.
So I'd be in favour to record (in parallel) the "pure" information about "frequencies" (pitch), e.g. by converting Helmholtz pitch notation to numerical values, using c = 0, c# = 1, d = 2, c' = 12,... C = -12, ...
Concerning (2): I'd be favourable to record some classical texts (Caesar, Shakespeare, ...) using some coding system which could be considered as not-too-much-arbitrarily-chosen standard, like A=1,B=2,... or ASCII or UTF.
I did not forget that there is certainly much literature of equivalent value in non-latin/greek/similar alphabets (e.g., Arabic, Chinese, ...), but I do not have a valid opinion about how to cope with this.
Maybe each of the sub-topics would merit a Wiki page (and related discussion page) on its own, at least it is clear that the discussion should not take place on the SeqFan list, that's why I created this talk page to start with. — M. F. Hasler 13:22, 12 June 2012 (UTC)
- In the above text, I added the link to 'scientific pitch notation' on Wikipedia. — Daniel Forgues 21:15, 12 June 2012 (UTC)
- In regards to paintings, I was more thinking of cases where the painter consciously put a number or sequence of numbers into the painting, or maybe even if there is halfway credible theory that this is the case (e.g., if someone claims that Da Vinci put 7phi into the Mona Lisa), rather than encoding the artwork in the OEIS. Alonso del Arte 01:21, 13 June 2012 (UTC)
I strongly suggest to add specific keywords
- "text" for sequences representing text (1=A, 2=B or ASCII or ...)
- "music" (or rather "midi") for [MIDI- or similarly encoded] "music" sequences. — M. F. Hasler 13:35, 12 June 2012 (UTC)
- George Paschos contributed the java source code File:Seq.java (i.e. program to play out sequences as DirectDound (not MIDI) to your soundcard). Should we have DirectDound music-based sequences? — Daniel Forgues 21:15, 12 June 2012 (UTC)
As a third facet of artwork that possibly might be encoded on OEIS, one could think of pictures. However, I have not a valid opinion about a "good" coding system for such, which preferable should have the least possible arbitrariness. Are there any ideas about this? To start with, there could be two "naturally arising" approaches:
- Code the picture as a fixed grid of points.
- Use two different primes as side lengths, making the decomposition (almost) unambiguous: A finite sequence of length prime(m)*prime(n) can only be arranged in two ways (landscape(preferred) or portrait) as a 2-dimensional grid.
- To remove any doubts, one could also use two initial terms (e.g., a(0) and a(1), or a(-2) and a(-1)) to specify the dimensions. Then list the points, read line by line. (Anything else would be more contrived, IMO.)
- Code the picture as a square matrix in standard OEIS convention (antidiagonals upwards, IIRC). Use 0 for "transparent", then a picture of finite size will be a sequence ending in repeated zeros.
- Code the picture by successively refining a grid, e.g.:
- a(0) = mean value of gray over the whole picture, a(1)..a(4) = 2x2 (relative(?)) refinement, a(5)..a(20) = 4x4 refinement (each of the preceding 2x2 tiles again subdivided in 2x2 tiles), etc.
- Maybe pre-pend two terms to fix the ratio of the picture (say, a(-2) & a(-1) = width and height of the artwork in mm or any more adequate unit, to be specified in a comment).
- Alternatively, the dimensions (be it in pixels, of the format , or absolute size of the "stretchable" format ) could be given as "additional parameters" of the corresponding keyword, e.g. "KEYWORD: sign,pict(640,400),nice".
As said above, suggestions by anyone having ideas on the subject are welcome / requested. — M. F. Hasler 14:07, 12 June 2012 (UTC)
- For black-and-white images, we could have a sequence of grayscale pixel values (0..255).
- For color images, we could have a sequence of concatenated RGB sub-pixel triplets values (0..255, 0..255, 0..255). Or we could have a sequence of redscale (R) values (0..255), followed by a sequence of greenscale (G) values (0..255), then followed by a sequence of bluescale (B) values (0..255). Alternatively, we could have a sequence of brightness (B) values (0..255) (which readily gives a black-and-white version of the image), followed by a sequence of hue (H) values (0..360), then followed by a sequence of saturation (S) values (0..255). — Daniel Forgues 21:15, 12 June 2012 (UTC)
Response from Neil Sloane, Jun 12 2012
Rather than adding more keywords, which is a complicated process, I suggest that we expand the entry in the Index, Section Mu, under music, sequences related to, and then add links for the appropriate sequences. The link is this:
<a href="/index/Mu#music">Index entries for sequences based on music</a>
To start the ball rolling, I added this link to A001491, a Beethoven sequence.
We can do the same thing for sequence based on "text"s.
I added under video games that
- Tetris has A000988(4) = 7 pieces.
Actually what I wanted to do was add a similar remark under a "law" section regarding Tetris Holding & The Tetris Company v Xio Interactive (NJ district court, Civil Action No. 09-6115) where the question of copyright and trade dress is raised. In particular, p. 27 considers the seven one-sided tetrominos and the question of whether their use is a protected feature.
But I wasn't sure if this should be added, and in any event it's not clear where to add law/jurisprudence to the list.
Charles R Greathouse IV 15:40, 21 June 2012 (UTC)
- I can't imagine a reason for this to have a law subsection, but I can imagine subsections for economics and politics, like sequences pertaining to various voting schemes. Alonso del Arte 16:13, 21 June 2012 (UTC)
- I expect there are more connections between the OEIS and law than there are between the OEIS and calligraphy or naval engineering.
- Social choice theory (jury theorem, Condorcet paradox, Arrow's theorem) has many potentials for OEIS sequences though I'm not sure how much of this is realized. Not sure I'd consider this polisci except for the purpose of this list, though.
- Charles R Greathouse IV 18:44, 21 June 2012 (UTC)
- I had been thinking of adding Law and justice and Politics (we now have Political sciences, whereas politics is anything but science...) sections, but I was not convinced it was relevant. I also thought of adding Religion and spirituality but I also wasn't sure if it was relevant. — Daniel Forgues 05:12, 22 June 2012 (UTC)
Reaching: Star Trek
A010701: There's an episode of Star Trek: The Next Generation in which the Enterprise is destroyed several times. The loop is broken when Data realizes that all the 3s he's been seeing mean that Commander Riker's suggestion for avoiding destruction is the right suggestion. Alonso del Arte 14:49, 15 July 2012 (UTC)
Physic particle masses
I am unsure how A003675 and A254290 are supposed to differ (or not). One shows 1.008664904 and the other 1.00866491. The NIST reference confirms the latter. Should the sequences be consolidated? Similarly, A003677 does not seem to agree with 1.007276466 from the NIST reference 1. Similarly A057720 looks like 2 add'l digits can be added 2.--Bill McEachen 00:44, 1 February 2016 (UTC)
- Thanks for spotting the duplicate entries! I've consolidated them. Also, there were some more duplicates among natural constants, see here. Regarding keeping the constants up-to-date, it is a task for an enthusiast of physical constants in the OEIS. In the particular case of A057720, only one digit should have been added, as the next digit may be either 8 or 7. More importantly, the offset was wrong. --Andrey Zabolotskiy 15:02, 22 June 2017 (UTC)
Game shows, Numerical Integration
- also, what of numerical integration? I see A058319, A093735/736 as a few. By the way, the A058319 sequence does not match what is seen at 1, though I have not researched it further.--Bill McEachen 14:54, 20 January 2018 (UTC)
- a separate reference is here: 2, and unless I'm mistaken, the coeff sequence should be symmetrical. If the intent of A058319 was for any subsequent mirroring to be obvious/implied, I suggest a note to this effect.--Bill McEachen 18:36, 20 January 2018 (UTC)
- All of those should be added (if they haven't been already). Good finds! - Charles R Greathouse IV 06:40, 22 January 2018 (UTC)