Comments on: Word Counting in C++: Computing the Span of a Word

By: Jonathan Boccara

Jonathan Boccara — Thu, 08 Nov 2018 12:08:00 +0000

In reply to kl_pq. Yes, good idea thanks!

By: kl_pq

kl_pq — Fri, 02 Nov 2018 08:17:00 +0000

I guess I’m a bit late but I’d like to mention that I believe you neither need nor should use optional here: The WordStats instance gets created only when there is a word, and therefore there also must be an initial line number.
So if a WordStats object got created but no occurrence was ever added added, this is probably a logic error and would even justify assert()ing.

In my opinion the code gets simpler by just checking the value of nbOccurrences_: http://coliru.stacked-crooked.com/a/9c6953aa32ef4973

By: Jonathan Boccara

Jonathan Boccara — Tue, 23 Oct 2018 21:12:00 +0000

In reply to Benjamin Navarro. Yes good point! I will implement this, thanks for the suggestion.

By: Jonathan Boccara

Jonathan Boccara — Tue, 23 Oct 2018 21:09:00 +0000

In reply to jft.

Thanks for your comment. My idea of span is also to see how spread out a concept (independent of C++ concepts) is in code. So even if a term is somehow carried by several object, and all are huddled together, I’d like to know this. What’s your opinion on that?
About your latest remark, I see your point, so how about expanding span to a measure of density perhaps, equal to #occurrences/span for example?

By: Benjamin Navarro

Benjamin Navarro — Tue, 23 Oct 2018 16:44:00 +0000

I usually rely on ±std::numeric_limits::max() to initialize my min/max variables. This way, no checks are required since for the first run, std::min(std::numeric_limits::max(), value) or std::max(-std::numeric_limits::max(), value) will both return value. I edited the code to reflect this: http://coliru.stacked-crooked.com/a/5a951e60df4d7a08

By: jft

jft — Tue, 23 Oct 2018 10:20:00 +0000

" it indicates how spread out the word is in a piece of code: are all its usage located in the same area?" Yes if the usage of the word is at the same scope. But say:


for (int i = 0; i < 5; ++i)
   cout << i << " ";
for (int i = 10; i < 15; ++i)
    cout << i + 10 << " ";

The 2 i integer variables here are not the same variable. So saying that i here has a span of 4 IMO is not correct. There are essentially 2 different i variables here each with a span of 2. And if multiple structs/classes are used which have the same name for a member variable/function then I'm not sure what a span number is implying? - especially if a class declaration is separate from its definitions and its usage. What about function declarations and definitions? The span of a variable in a function declaration at the top of a program would be from its declaration line to its last line of usage - which could be hundreds and hundreds of lines between its usage in a function declaration and its first usage in the actual function. ???