Id Comments ak149 you've gone a bit overboard with this assignment haven't you? I like the fact that Scala and Python are about the same size code-wise. Python3 seems to be a good step forward given the efficiency increase. C is an interesting story: good speed but still a pain to get it to work correctly. az17 you are a C++ at heart aren't you? C++ compiles but segfaults Scala incomplete no results (why did you not have access to the tweets?) bmc16 well done good work! bws5 your python solution looks good but I missed Scala/Prolog and a results file cph4 good java (but do not hard code file names use a command line argument instead) scala/prolog missing dmt17 you seem to have misread the specs but I'll give you full marks for the Java versions anyway; scala/prolog missing dsh8 good work on the Java side the Scala version could have done with a bit more tlc ejd9 good work on the Scala side the C++ version could have done with a bit more tlc hjg10 I am impressed by your choice of D jag39 jdb29 I enjoyed reading your report and agree with your Scala version being cleaner a perfect example of the advantage of a reimplementation ... jlm47 if you cannot fully readin large files you'd need to iterate reading line after line plus extend the heap if the counts need more space jps22 I liked the conciseness of your solutions especially the Python version jr94 good Java version Scala has two type errors kgas1 if you cannot fully readin large files you'd need to iterate reading line after line plus extend the heap if the counts need more space kjjg1 well done lb65 cool probably the fastest C# submission! mdb29 do not hard code file names use a command line argument instead (and maybe a default value); [you were struggling a bit with Scala right] mfb3 well done especially your concise Scala solution mfma2 I have tested your Scala version on a subset and your counting is buggy mjh77 full marks for using Prolog! If you have only two clauses xyz([] ...) and xyz( [X|L] ...) then no (ugly) cut is needed thanks to first-arg indexing. Use _VarName to avoid singleton warnings. mm236 solid solution but personally I have strong preference for splitting up large methods into smaller ones. This also enables better re-use of bits and pieces later on. mrl13 do not hard code file names use a command line argument instead (and maybe a default value) mrr21 good code have you done the Scala version after the Java version (given the code structure and your work time estimates)? msb26 do not hard code file names use a command line argument instead (and maybe a default value); your Java version is much more efficient would be interesting to do some profiling to see why the Scala version is so much slower njh34 I like your python version but your Scala version looks buggy to me no15 runs fine for me maybe it was GC running in parallel; the File.lines apparently reads in the full file ... ns127 I agree having all runs take exactly the same number of seconds is surprising; use milliseconds to see some variation ... rdc12 your scala code looks buggy: line 57: handle(sb.toString()) what is "sb" ? I like the Ruby version! rgf5 good on you for at least trying to complete the Scala version rgs10 you must have misunderstood the specs (and ignored the example output): the topK of each size 1 to 20 were asked for. Other then that: good code! rlb22 solid solution but personally I have strong preference for splitting up large methods into smaller ones. This also enables better re-use of bits and pieces later on. rsb16 solid solution but personally I have strong preference for splitting up large methods into smaller ones. This also enables better re-use of bits and pieces later on. sb144 your Scala timings are missing but the code seems ok except for an off-by-one bug sdw14 your timing results look high did you use R block machines? sh212 sjv14 do not hard code file names use a command line argument instead (and maybe a default value); the message "Please provide a value for K" is rather confusing when one provides arguments like " 17 someTweets.txt" sk218 do not hard code file names use a command line argument instead (and maybe a default value); src17 your Scala timings are missing but the code seems ok swp4 your scala code looks seems to be stuck in a loop; "exception in main" is not much help it would be better to print the stacktrace or some other form of info? Do not hard code file names use a command line argument instead (and maybe a default value). tc99 are you sure it only took you 10 minutes to code the Java solution? tcg4 good work but I am sure you could simplify your code given more time. user