Talk:Parsing/RPN calculator algorithm: Difference between revisions

→‎Graduation to task: Single digit integers suffice for demonstration purposes.
(→‎Graduation to task: Single digit integers suffice for demonstration purposes.)
 
(5 intermediate revisions by 3 users not shown)
Line 4:
 
: Several programming examples are treating the input as characters instead of tokens, and this is working because there are no multi-digit numbers, and no numbers that contain a decimal point, and no numbers that are expressed in exponential notation (such as 3E27), and no numbers that have a leading sign, and no multi-character operands (such as '''**'''), etc. Should those programs be considered incorrect? (Just because they come up with the correct answer, doesn't make it correct. Hurrumph!) -- [[User:Gerard Schildberger|Gerard Schildberger]] 23:56, 14 December 2012 (UTC)
 
::The Fortran example contains the remark that dealing only with single-digit integers enables a simple scan, while a fuller multi-digit integer (or further, floating-point with exponent part) requires much more code which would obscure the presumed objective of exemplifying the parsing of RPN itself. I have a function EATREAL that takes 120 lines, while the task is achieved in less than fifty. As for revealing the state at each step, two statements: a WRITE and a FORMAT suffice. [[User:Dinosaur|Dinosaur]] ([[User talk:Dinosaur|talk]]) 10:49, 4 October 2016 (UTC)
 
== Extraneous Requirements ==
Line 34 ⟶ 36:
 
:::: (Do you think we should migrate the bulk of this discussion to the Village Pump page, since it all about the general approach rather any particular implementation?)--[[User:Abathologist|Abathologist]] ([[User talk:Abathologist|talk]]) 04:44, 21 September 2016 (UTC)
 
::::: This discussion is long enough, already, that I don't think it should go on village pump. If you feel dropping a small section there that links here, that could work.
 
::::: I don't like the idea of making it optional - that just seems lazy. The problem being that the task results would then be less comparable, and the whole point of this site is to create examples which are comparable across languages.
 
::::: It's ok to object to this aspect of the tasks, though, and to recommend that most tasks avoid such things, and to also require a bit more work on the task author's part (such as making it a part of the page title). And, I suppose, it's also all right to set up the task such that the logging is an "extra credit" part of the task rather than a core part.
 
::::: But I also want to get some other people's opinions on this - I know other people have disagreed with me on other issues, so I don't think I should be the sole voice here. --[[User:Rdm|Rdm]] ([[User talk:Rdm|talk]]) 07:58, 21 September 2016 (UTC)
 
:::::: I see your point about the optional option encouraging non-comparable implementations. I hadn't considered that. By "extra credit part" do you mean, it would not be an optional feature of the algorithm, but like a secondary, extra credit implementation? E.g., something like,
 
::::::: >Task: Implement X using Y and Z. Extra credit: implement another version of X using Y and Z that logs intermediary stages of the data.
 
:::::: If this is a correct interpretation of your suggestion, I think it is a very good idea! It addresses all of my concerns.
 
:::::: I am fine waiting for some time. I created that Village Pump page in the hopes of soliciting feedback, but none has yet been given. What other steps should I take to get a wider chorus of voices chiming? Thanks for your feedback and guidance! --[[User:Abathologist|Abathologist]] ([[User talk:Abathologist|talk]]) 03:09, 3 October 2016 (UTC)
 
::::::: Yes. Thank you. --[[User:Rdm|Rdm]] ([[User talk:Rdm|talk]]) 04:03, 3 October 2016 (UTC)
 
::::Whilst the actual logging code is probably mundane, it does show how the algorithm works. If there was a vote, I'd favour keeping this particular task as it currently is. --[[User:Tigerofdarkness|Tigerofdarkness]] ([[User talk:Tigerofdarkness|talk]]) 19:12, 3 October 2016 (UTC)
1,220

edits