Puzzled: Two Move Checkmate

That's another interesting point. Are we playing the FEN or the diagram? You may have noticed quite a lot of posts show diagrams where white is obviously on move but the "Analysis" shows black on move. I always ignore the FEN because (1) you shouldn't open the analysis module in the first place because it will also show you the best moves which is hard to ignore (2) the all important miscellaneous conventions for composers are precisely designed to replace the FEN! In effect, they give default values for all unknown states and implicitly demand that you ignore the FEN because things like castling right are fundamentally uncertain. The only things you can use are the diagram, the stipulation (caption) and the conventions - and of course the chess rules.

The fundamental uncertainty of castling and other rights are very much like the fundamental uncertainties in quantum states which Einstein fought in the second part of his life and are treated in the same way in one of the retro-logics. Don't worry, you will not collide with these unless you try to solve retro-problems but it conventionally applies to all compositions.

Note: if you want a different value for a castling state from the default (castling rights are on) then you need to specify it in the stipulation, not in the FEN (unless of course there is a stipulation text field in the FEN).

You should click on the two fingers to the right of the analysis icon and then on PGN. This will give you the FEN - any moves included in the PGN will generally be invisible. If you click instead on the analysis icon for this puzzle it won't give you the most accurate move anyway (as previously posted) and that's probably true in general.

I would assume that we're playing the position which in general is not fully specified by the FEN, but under current basic rules the only things missing are which pieces have been touched in which order and whether anybody has resigned or a draw has been agreed. It's definitely not fully specified by the the diagram - even side to move is not shown unless it's a puzzle.  Under competition rules there are are more things missing from the FEN, but with PC=0 both the 50 move rule and repetition rules are settled. Things like flags falling, arbiters awarding the game, one player dying etc. would normally be assumed by common sense.

You can't necessarily assume that posters are familiar with the conventions for compositions. I know where to look up the rules for chess, but I have no idea where to look up the conventions for compositions. 

Thanks! I didn't know that! I always open the Analysis module to find FENs/PGNs.

You are a product of the digital era.

30 years ago, anyone would say that you can't assume that solvers can read FENs and PGNs while the conventions are intuitive and simple - for the greatest part that is. FENs and PGNs were invented as a tool for the communication of problems and games over digital devices, not as a representation tool in a human interface. The composition conventions are in use for at least 70 years (there are versions from the 1950s) and the flaws are in the definitions of FENs and PGNs because they didn't cover the full ground of the conventions as they should. For instance, there should be a "castling permission" state besides the "castling right" state. Permissions are not absolute.

Note that a composition is not always a "position" - it is a diagram + conventions which make it potentially a position cluster. The omissions in FENs and PGNs are therefore quite fundamental.

Not everything in game play is covered by the conventions or FENs/PGNs and that is why there are differences between "games" and "compositions". For instance, there is no clock and there are no arbiters inside compositions. The most important field in compositions is the free format "stipulation field" which cannot be formally interpreted by a computer algorithm. Besides stating the goal of the problem it covers any and everything not standard - in rules, conventions, goals, arts and politics . That is where we can also correct the shortcomings of FENS and PGNs. Btw, I wouldn't be surprised to learn there is a special notation format for compositions today; I do not follow the news closely.

I expect that solvers follow the simple conventions for compositions - when they solve a composition. If they don't know them and can't intuitively guess them, I am always willing to explain or provide a link to them as I have done countless times. Interestingly, there is no collectively agreed alternative for the composition conventions. You look up the FENs but most solvers will not; they operate on text captions and their private defaults - like "it is obvious here that black plays from the bottom up". There are quite a lot of really wrong FENs in posted puzzles which remain unnoticed since no moves are given and the puzzle mode is not turned on. Still, they matter to solvers like you who get data like "which castlings are legal" and "who is on move" from the FEN. By the way, the wrong "who is on move" state is randomly generated by the chess.com puzzle interface. What is particularly irritating about the FENs is that it gives you information you shouldn't have. For instance, you shouldn't know you cannot castle or can play e.p. in positions where such can be proved.

This is just the top of the iceberg when it comes to issues relating to solving puzzles and positional conditions. Not only the FENs/PGNs are poor, the puzzle interfaces are immensely poor and, unfortunately, even the miscellaneous conventions are poor. I am willing to forgive everything except the last category as the problemist community should be capable of doing better!

Addendum:

You mentioned some other conditions like the "PC=0 count", repetitions and the "touched pieces"(which we discussed before). These are insufficiently addressed by the conventions mainly because the problemists didn't care to formalize things they thought obvious - which is the same error amateurs make when they assume certain conditions in a diagram because it is "obvious". Nevertheless, I know what problemists think about them and that comes down to permissions and dual states. You may assume PC=0 until something happens which proves it otherwise. In the diagram the PC-state is unknown but something may happen during the solution which proves PC>0 in the diagram in retrospect. This type of logic (retro-strategy) is alien to you but it is necessary to resolve issues in retrograde puzzles. Note that these are not puzzles with special rules - they are just puzzles which are undecidable until you invoke a conventional retrograde logic.

And, as with the other conditions discussed, the conventions intend to cover these issues as well, however poorly. The only practical way to address them all on FEN/PGN level is to eliminate the FENs alltogether and require PGNs to always deliver complete games up to the diagram position. Which is of course completely unacceptable in the retrograde field.