intresting logic: what would you do here

Post puzzles for others to solve here.

Off-topic

Postby marek stefanik » Sat Apr 13, 2024 12:02 pm

denis_berthier wrote:Hey, this has been the 21st century for 23+ years, in case you didn't notice.
Didn't notice the 20th century ended? No, I wasn't alive back then.

denis_berthier wrote:All these properties and more have been stated and proved in [CRT] and [PBCS] some 15 years ago.
Oh no, at first, everyone was "copying" you, and now they are trying to explain your work to others when they know damn well you will not do it yourself?
On your github you mention that you used to be a professor.
If in the first lecture you said: "Go read my books, see you at the exam. Byee!", I'm sure most students would drop the course and sign up to a different one. I would.

I would have preferred to write something more serious, but I'll have to wait for people who might actually be interested in such conversations.

Marek
marek stefanik
 
Posts: 359
Joined: 05 May 2021

Re: intresting logic: what would you do here

Postby StrmCkr » Sat Apr 13, 2024 8:27 pm

Your Z Chan's do hold memory Notation as your last link is blank . Ie a forcing chain similar to a nice loop that acts as a closed loop proving the initial claim is true or false. These have advantage as they can impose changes to a sub-graph based on its assumptions/ implications, they also can be reversible with the added blank digit reintroduced as its unlisted as it's off. How is it off if it isn't retaining memory of a previous event. It does as it's directional implication, a technicality in your Notation system.


Yes
Aic are standardized as xor logic gates(strong link) in digits with weak-inferences on a bidirectional graph transversal from left and right hand of each node. 2005/2006 and adopted here under eureka Notation 2008 and onward, where all forcing chain logic stopped being used except for puzzle past se 9.9 ~

XOR logic gates have 6 single digit types, (bivalve,bi-local, grouped to single, single to group, group to group, eri)
it also has extension xor gates using als, ahs, almost fish, are also defined and standardized)

A weak-inference is specifically &! Logic gate where x digit/cell cannot be true twice for a sector/cell shared between nodes.

As for Notation these xor logic gates are established using: RC, Rn, Cn, Bn space with compression Notation established decades ago

Length counts is node +weak inferences

Readability is much easier and always reversible, and million times easier to teach this.

Xyz wing ( ab=c) r1c14 - (c=a) r3c3 => r2c23<>a

Easiest ie simplest and shortest path
would be shortest link count my displayed move in the first post has 3 strong links and 2 weak inferences which is length 5.

The only parts not firmly document is how multi criteria are listed as simultaneously inferences effecting the next link ( "|" ie and or)

How to utilize ahs that technically cells instead of digits, where ahs into ls are now digits (for Notation purposes only) where aic specifically use Digits not cells.

Most of this predates your publicized work, as for your work these very questions has been a topic on and off for a long time of why haven't you converted to the standard. Every time your reply is read my work end of story/discussion.

Even your own nrc Notation stems back to here from ABs insights into rc, Rn, Cn, Bn grid view that changed how we program.
And look at logic, without references to work me and others did behind the scenes some of your own concepts wouldn't have made you books as we challenged you provided examples etc.

I've have zero disrespect for your work, and weather you know it or not I helped with more then you know I had other correspondents in my stead as my writing is lacklustre on many finer points, My names even credited in your work for some things

So please don't play me off as If I don't get your work, nueanes might lapse time to time but your stuff is unique to you and not used any where else.

I teach others, reference to you and your books, your methods as somethings you have brought are handy and insightful, it is a pain to convert your work back into the Common practices as they simply don't match 1:1.

Even more frustrating is that you don't or rarely acknowledge our work and convert them to what you do it's been a 1 sided street for a long time

this is why I said I didn't want to engage in this rhetoric, it accomplishes nothing.

Asides all this stuff i asked not to engage on : I asked

how is your longer chains used the simplists. Ie easiest method.

When there is examples above of shorter logic and less steps
(iW ring a length length 8 aic chain,[ aic+als+ahs] length 5 => reduce grid to singles in 2 steps.
Some do, some teach, the rest look it up.
stormdoku
User avatar
StrmCkr
 
Posts: 1430
Joined: 05 September 2006

Re: intresting logic: what would you do here

Postby marek stefanik » Sun Apr 14, 2024 6:09 pm

StrmCkr wrote:How to utilize ahs that technically cells instead of digits, where ahs into ls are now digits (for Notation purposes only) where aic specifically use Digits not cells.
I don't know what you mean by that (an example might make it clearer), but using the | convention Cenoman mentioned, the notation of ahs seems quite intuitive to me.

Regarding the endless dispute about lengths, I think it depends on the pattern.
For AICs we get the same (pre)ordering regardless.
For krakens I would even argue that krakens where two branches quickly merge are simpler than krakens (with the same number of links) where they don't.
For something like Patto Patto, the total number of links feels somewhat arbitrary – we need r1 as weak links, but we can use it as strong links too, instead of the cells, which would reduce the total number of links but (imo) wouldn't make the pattern any simpler.

Note on AICs: if we don't count the weak links connecting the ends of the chain and the eliminations, we get that a nice loop is 'more complex' than each of the underlying AICs, so counting them seems better to me.

Marek
marek stefanik
 
Posts: 359
Joined: 05 May 2021

Re: intresting logic: what would you do here

Postby StrmCkr » Sun Apr 14, 2024 9:12 pm

Don't get me wrong here I know how to use ahs as links manually

, my concern has been to write them out as eureka so my new solver actually does it correctly as written aspects of ahs (wasn't documented anywhere ad these used to be not used when it was written)

als
Digits (123) @ r1c23

Has 6 diffrent xor strong links generated.
(1=23) r1c23, (2=13),r1c23 (3=12)r1c23
(23=1)r1c23,(13=2)r1c23, (12=3)r1c23

aic use digits, the &! Logic between nodes is straight forward as the first digit of the node cannot be true twice for the digit.

With ahs

Cells r1c124 @ (digits 12)
Has the following 6 strong links

(R1c1=r1c24)12, (r1c2=r1c14)12, (r1c4=r1c12) 12
(R1c24=r1c1)12, (r1c14=r1c2)12, (r1c12=r1c4) 12

It's no longer using digits, instead its cells that you have to &! As the weak inference

Which makes transparency from all other links types not as clear, plus these nodes changes where digits are located inside the
Node instead of continuation of digit lookups.

Example:

(1)R5C7= R5C1 - (R1C1=R1C24)(12)

If that's correct then i see how to actually code ahs links for once.
(I'd have to modify my next cue call to also check the last node cell for peer links on the ahs table, as well as digits for the other tables)

My hangups been on aic use digit links from node to node.
With ahs if I wrote the digits first then the cells the &! Gate dosent work correctly.

Hence my? On how on write these as my presumptions was
(digit) (xor) &! (digit)(xor) : is digit linking how does it write out a cell link When these have digits then links.

Simplified ~
My present aic code links from r5c1 looks for another (1) stronglink
Keyed to the sector that r5c1 occupies,
and the next strong link (a=b) all of a is also in that sector.
~
(my tables are built with, all that information and others stuff for eliminations)
I also dont see the need to count weak inferences to determine eliminations for extra counts
( I see it as the intersection of peers of start and end, which is information carried by my links)
And since it dosent connect to a new node or (back to start) it dosent change counts.

InWhich I do Agreed aic have shorter link counts then niceloops, and drop step counts down as well.

aic have access to more eliminations, (As each node of an aic is a starting and ending point)

As they aren't a contradiction chain with singular targets for elims/assertions
(skyscraper example 4x niceloops for all elims, vs 1 aic)
Some do, some teach, the rest look it up.
stormdoku
User avatar
StrmCkr
 
Posts: 1430
Joined: 05 September 2006

Previous

Return to Puzzles