Check ordering of exercises with stubs vs without stubs

[petertseng]

Here are the exercises with stubs, with their order in config.json and the number of lines in the stub:

 1      hello-world  23
 2             leap  12
 3            clock  26
 4       gigasecond  10
 5        raindrops   7
 6              bob   3
 7         triangle  17
10       food-chain   3
19          hamming   3
23    crypto-square   3
30       word-count   9
36  circular-buffer   3
40       custom-set   3
61    tree-building 109
64           paasio   3
68       tournament   3
71           ledger 227

A few points of interest to me:

• difference-of-squares, secret-handshake, do not have despite coming before food-chain which does
• a big gap between hamming and food-chain: should hamming be moved earlier, or should its stub be deleted?
• it's understandable that tree-building and ledger have stub files since they are refactoring-focused exercises...
• but what about all the others that come after (my arbitrary cutoff of) hamming? And the three-line ones have nothing more than a testVersion and package name, so don't seem terribly necessary.

If the idea behind the stub files is to give people new to Go a starting point, it makes the most sense that the first N exercises would have stub files and then we can stop. So I wonder if we should close the gaps in the early exercises.

Footnote, script run to produce these results:

exercises="
    hello-world
    leap
    clock
    gigasecond
    raindrops
    bob
    triangle
    difference-of-squares
    secret-handshake
    food-chain
    house
    pascals-triangle
    series
    queen-attack
    grains
    etl
    scrabble-score
    parallel-letter-frequency
    hamming
    sum-of-multiples
    pythagorean-triplet
    bank-account
    crypto-square
    luhn
    largest-series-product
    sieve
    palindrome-products
    bracket-push
    anagram
    word-count
    binary
    allergies
    rna-transcription
    roman-numerals
    say
    circular-buffer
    robot-name
    react
    counter
    custom-set
    atbash-cipher
    phone-number
    strain
    pig-latin
    prime-factors
    nth-prime
    diffie-hellman
    octal
    beer-song
    ocr-numbers
    wordy
    nucleotide-count
    grade-school
    matrix
    saddle-points
    trinary
    hexadecimal
    meetup
    binary-search
    binary-search-tree
    tree-building
    kindergarten-garden
    simple-cipher
    paasio
    pov
    minesweeper
    robot-simulator
    tournament
    word-search
    connect
    ledger
    poker
"

i=0

for slug in $exercises; do
  i=$(expr $i + 1)
  stub=$(echo $slug | tr - _)
  file="exercises/$slug/$stub.go"
  if [ -f $file ]; then
    lines=$(wc -l $file | cut -d\  -f1)
    printf "%2d %16s %3d\n" $i $slug $lines
  fi
done

[kytrinyx]

Interesting.

I think hamming could go earlier, and we could probably keep bob's stub file and delete all the subsequent ones unless there's a particular reason to have it.

[petertseng]

I examined word-count's stub file. It defines a type Frequency map[string]int that is subsequently used in the tests and expected as the return value of WordCount. For now I'm going to leave that particular stub file. The three-line stub files are not adding much value and I'm going to delete them.

[kytrinyx]

That sounds good.

[petertseng]

Great, now the stub files look like this:

 1      hello-world  23
 2             leap  12
 3            clock  26
 4       gigasecond  10
 5          hamming   7
 6        raindrops   7
 7              bob   3
 8         triangle  17
30       word-count   9
61    tree-building 109
71           ledger 227

This all looks reasonable to me.

[kytrinyx]

Yeah, this looks good to me too.