Local counts: longest sentence and longest grouped run

Sometimes the value to maximize is not printed on one card. You must first build a local count across several cards, then compare it with the best completed group.

Two variables with different lifetimes #

Count measures the current sentence, so reset it at each sentence boundary. Longest measures the best completed sentence, so keep it across boundaries. Count the full-stop word before testing the boundary, then compare and reset.

Count = 0
Longest = 0
For each word X in paragraph order {
    Count = Count + 1
    if (X.Word ends in a full stop) {
        if (Count > Longest) { Longest = Count }
        Count = 0
    }
}
If Count > 0: handle the unfinished last segment by the stated rule

Check the actual paragraph #

These lengths total 65 cards, and the longest is 25 words. The spoken demonstration calls the fourth count 20; the supplied table has 21. The final cane... token closes the last sentence once; its three printed dots are not three extra words or sentences.

Never forget the last group #

The supplied paragraph ends with punctuation, so its last group is finalized inside the loop. For general input without a final stop, choose whether an unfinished segment counts as a sentence, then implement that rule. For grouped records whose boundaries are signaled by a changed key, flush the last group after the loop: no later key change will do it for you.

Tutorial: students grouped by birth month #

Tutorial 2.2 groups equal birth months together and counts each contiguous run. On a new month, finalize the old month’s count and start the new one at 1 for the current card. At end of input finalize the last month too. Alphabetic month order is enough for grouping; it is not calendar order.

From the supplied Scores table, March, May and December each have 4 students, the highest birth-month frequency. A run method is correct only after equal months are together. If cards are shuffled, use a separate counter for every month instead.

Source: Lecture 2.3; Tutorial 2.2. Small teaching datasets and diagrams are prepared for these notes; lecture and tutorial totals are identified explicitly.