Design Module Example - Tony Ballantyne

Let's take the parser of a turtle graphics implementation as an example. The program will receive commands such as REPEAT 4\[FD 100, RT 90\] and will output pictures , in this case a square. One of the modules required will be a parser. The parser will have to * receive a string * split it according to white space * tokenize it ie recognise logo commands and arguments * chunk the tokens into commands (eg fd 100 rt 90) * check that the string is properly formed (so it doesn't say fd 100 90 30 rt for example. All correct tokens but they don't make sense in that order) * form set of commands to pass to the renderer (the thing that will draw the shapes) Now you break down each of those steps so to tokenize for example * Walk through the individual parts of the command (fd, 100, rt, 90) * Check if they are valid commands or arguments * output an error if not * form a string of tokens representing the string So for the tokenize module we'll need to do the following: | | | | --- | --- | | Sketches | Not appropriate here | | Data representation | FD: FD, Forward RT: RT, Right number: number REPEAT: RE \[ : \[\[ | | Sample Data | FD 100 RT 90 FD 100, RT 90 REPEAT 4 \[FD 100\] REPEAT 4\[FD 100, RT 90\] | | Description of algorithms | Below is a basic outline. Each step needs to be broken down into substeps 1. receive a string 2. split it according to white space 3. tokenize it ie recognise logo commands and arguments 1. loop across list array 2. for each item in the array look up valid tokens 3. 4. chunk the tokens into commands (eg fd 100 rt 90) 5. check that the string is properly formed (so it doesn't say fd 100 90 30 rt for example. All correct tokens but they don't make sense in that order) 6. form set of commands to pass to the renderer (the thing that will draw the shapes) | | Code | * Include Pseudocode and/or copy and paste appropriate code from implementation here * Which one? Whichever explains the best. For a complex algorithm, pseudocode is probably best |