Reveal the matches

A hidden word and a partly revealed word are different pieces of information. Start by keeping both. This first program inspects a fixed guess; it does not read the keyboard yet.

Use Sinclair BASIC on a stock 48K Spectrum. Begin with a new program. Meet BASIC gives a slower introduction to entering and editing listings if you need it. BASIC line numbers determine execution order; the gaps leave room for later additions.

  10 BORDER 0: PAPER 0: INK 7: BRIGHT 0: CLS
 200 LET w$="BOTTLE": LET g$="T"
 250 LET d$="": LET found=0
 260 FOR i=1 TO LEN w$: LET d$=d$+"_": NEXT i
 400 FOR i=1 TO LEN w$
 410 IF w$(i)=g$ THEN LET d$(i TO i)=g$: LET found=found+1
 420 NEXT i
 500 PRINT "Word: ";d$
 510 PRINT "Matches: ";found
 520 STOP

RUN should display Word: __TT__ and Matches: 2, followed by the STOP report. The report is deliberate: there is nothing else to do yet. The word BOTTLE stays in w$, while d$ contains only what the guess revealed.

Keep the answer separate

A variable ending in $ holds text, called a string. Quotation marks surround literal text; they are not part of the stored word. LET assigns a value. At line 250, "" makes an empty string. Adding "_" joins another underscore to its end.

LEN returns the number of characters. The FOR/NEXT loop at 260 builds six underscores because BOTTLE has six letters. Its counter i visits 1 through 6. Colons separate several statements on one numbered line.

The second loop examines the answer. w$(i) reads one character; string positions begin at one. IF compares it with g$, the guessed letter. When they match, LET d$(i TO i)=g$ replaces that one-character slice of the revealed string. The rest stays intact. THEN runs the remaining statements on that line only when the comparison is true.

The loop keeps going after a match. Both Ts are therefore revealed, and found increases twice. A search that stopped at the first match would hide information the player had earned.

Replace the "T" in line 200 with "L", then RUN. Try "Z": all six underscores remain and the count is zero. Restore "T" afterwards. Each RUN rebuilds d$ from empty, so these experiments do not preserve an earlier reveal.