Talk:ABC problem

Solution?

Had fun with this one last night after playing with my daughter. Should I include a solution? --Jking (talk) 14:46, 8 January 2014 (UTC)

You probably need to have a solution at hand to write a good task. If you've got it, why not add it as the first example? (I do) --Paddy3118 (talk) 20:34, 8 January 2014 (UTC)

Added.--Jking (talk) 03:46, 9 January 2014 (UTC)

Examples OK?

It appears to me that, given the set of blocks on the page, the result of a test on "BOOK" should be true?

Also it is possible to make a zero-length word with whatever letters you have. In which case shouldn't the results of can_make_word "" be True? --Tikkanz (talk) 19:42, 8 January 2014 (UTC)

I think you can spell the empty string with zero blocks. So the first test should return true. The reasoning is similar to the Python function all([]) returning True. Another clue comes from the Python iterative solution, that has needed a special case for the empty string. Often in algorithms if you have a special case it needs a special justification.-bearophile (talk)

Hi Bearophile, I think you are premature in your changes. The recursive Python solution has no such special case handling and returns False. I will revert your task change pending further discussion... (P.S. did you read my comments below on the null case)? --Paddy3118 (talk) 13:50, 9 January 2014 (UTC)

Hi, you need one B and two O's but the only blocks with either a B or an O are the two blocks BO and OB - you can't spell three letters with two blocks! --Paddy3118 (talk) 20:30, 8 January 2014 (UTC)

(The difficult null case). I take it as you cannot spell the null string with blocks as no block has a face without any characters. --Paddy3118 (talk) 20:37, 8 January 2014 (UTC)

OK, I was/am confused then about what the rules for using a block are? If a block is chosen, can only one of letters on the block can be used in the word? If so perhaps the following should be added as an additional constraint. "Only one letter can be used from each block chosen".

I guess that is a logical way to explain the null string generating "False", but using no blocks from the collection can also be thought of as "using the given set of blocks". Maybe better specification is required for this case too? --Tikkanz (talk)

can_make_word("") == False

Currently the original task creator set the answer for a null string as being False. Various people think that it could and should be True. There are three ways I think of dealing with this:

Keep the original False.
Switch to True.
Remove the need for the null string case from the examples and silently allow either True or False for this case.

It would be good to hear from the original task creator on this issue as well from other interested parties. --Paddy3118 (talk) 18:44, 9 January 2014 (UTC)

Well the question is, "Can you make this word given this collection of blocks?" The answer doesn't make sense without valid input and there are numerous ways to handle it in different languages. I chose the "false" value arbitrarily since it could be readily interpreted as, "no, I cannot spell any words if you don't give me one to spell." The case for "true" is equally valid. If it would help avoid bike-shedding over the issue I'd opt for removing the case altogether.--Jking (talk) 20:20, 9 January 2014 (UTC)

Not a set of blocks

The block (F S) appears twice so I changed the description to use the word collection rather than a set of blocks. I have also added a specific example "CONFUSED" that will fail if their is only one (F S) block available. --Paddy3118 (talk) 20:49, 8 January 2014 (UTC)

Why not change one of the toy blocks to a (S F)? That is the case for (B O). -- Gerard Schildberger (talk) 01:42, 9 January 2014 (UTC)

There could be duplicate blocks. I didn't mean "set" in the mathematical sense which would be confusing. Collection seems appropriate. --Jking (talk) 02:55, 9 January 2014 (UTC)