Stamp Collector

Stamp Collector is a philatelic inventory management program that allows users to track postage stamp collections by Scott catalog number and condition category. It uses a two-dimensional string array, `F$`, dimensioned at 600 rows by 15 characters, where each record packs a 5-character Scott number, a 1-character condition code, a 6-character quantity field, and a 1-character want-list flag. The program supports 18 stamp categories (singles, blocks, plate blocks, coils, imperforate, first day covers, and strips in mint or used conditions) encoded as alphanumeric codes 1–9 and A–H. Options include adjusting inventory quantities with positive or negative values, searching records by number and type, listing wanted stamps, and saving the entire array to tape.

---

Program Analysis

Program Structure

The program is organized as a menu-driven application anchored at line 70, which calls the main menu subroutine at 1000 and dispatches to one of five branches based on the value of O$. The structure is as follows:

• Lines 50–80: Initialization — sets SLOW mode, dimensions the master array, and enters the menu loop.
• Lines 1000–1099: Main menu display and input subroutine.
• Lines 1500–1599: Code-to-description translation subroutine (maps C$ to human-readable J$).
• Lines 2000–2022: Quit confirmation handler.
• Lines 2500–2999: Option 1 — Adjust inventory (add/remove a stamp record or update quantity).
• Lines 3000–3320: Option 2 — Search inventory by Scott number and/or category.
• Lines 3500–3710: Option 3 — List all stamps flagged on the want list.
• Lines 4000–4020: Option 4 — Save the inventory array to tape.
• Lines 5000–5999: Classification input subroutine (displays category table, reads code into C$).

Data Storage Layout

All data is held in the two-dimensional string array F$(600,15). Each row encodes one inventory record using a fixed-width field layout:

Columns	Width	Contents
1–5	5	Scott catalog number (left-justified, space-padded)
6	1	Condition/type code (1–9, A–H)
7–12	6	Quantity as a string (STR$ of numeric value)
13	1	Want-list flag (“Y” or “N”)
14–15	2	Unused

The 6-character key K$ formed by concatenating the padded Scott number and the type code (columns 1–6) serves as a composite primary key for searching and insertion.

Menu Input Idiom

The main menu uses a polling loop rather than INPUT for single-keypress capture. Line 1028 waits until INKEY$ is non-empty, then line 1029 captures the character into O$. Line 1030 validates the key and re-loops on invalid input. The same pattern is repeated in the classification subroutine at lines 5072–5075 and in the quit confirmation at lines 2016–2022. This approach avoids the cursor and input-buffer overhead of INPUT for menu navigation.

Inventory Adjustment Logic

Option 1 (lines 2500–2990) performs a linear scan of F$ from row 1 to 800 — notably exceeding the declared dimension of 600, which would cause a subscript error if the file is full and the loop reaches row 601. When a blank row is found (line 2556), a new record is inserted; when a matching key is found (line 2558), the existing record is updated. The quantity is read back with VAL F$(N,7 TO 12), adjusted by the signed input Q, and written back via STR$ (P+Q). Line 2811 guards against negative resulting quantities by looping back to re-enter the adjustment.

Search Logic

The search routine (lines 3000–3320) accepts an optional Scott number. If left blank, E$ becomes all spaces after padding, and the condition at line 3044 — IF K$(1 TO 5)=" " AND C$=F$(N,6 TO 6) — matches all records of the selected type across the entire file, effectively doing a type-only scan. The display subroutine at 3300 uses a counter G to suppress clearing the screen after the first match, allowing multiple results to accumulate on screen before prompting for the user to press Enter at end-of-file.

Category Code Translation

The subroutine at lines 1500–1599 maps the single-character C$ to a descriptive string J$ using 18 sequential IF tests. Codes 1–9 cover numeric ASCII characters and codes A–H cover uppercase letters, giving 18 distinct stamp categories. The validation in the classification subroutine (line 5082) checks C$<"1" OR C$>"H", which in ASCII order accepts all characters from “1” (ASCII 49) through “H” (ASCII 72), inadvertently including characters like the colon, letters A–G, and digits 1–9 — the intended valid range — but also characters such as “:”, “;”, “<“, “=”, “>”, “?”, “@” which fall between “9” and “A” in ASCII. These intermediate codes would produce the default J$="?????????" translation but are otherwise harmless.

Want List

Option 3 (lines 3500–3710) iterates all 600 rows and prints entries where column 13 equals “Y”. The want-list flag is set during inventory adjustment (lines 2840–2842) via a Y/N prompt. The listing routine at line 3600 reads the type code from F$(N,6 TO 6) back into C$ before calling the translation subroutine, correctly recovering the stored category description.

Tape Save

Option 4 prompts the user to ready the tape recorder and then executes SAVE "STAM%P" at line 4016. The %P notation represents an inverse-video “P” character in the filename.

Notable Bugs and Anomalies

• The inventory adjustment loop at line 2552 iterates FOR N=1 TO 800, but F$ is only dimensioned to 600 rows. Accessing F$(N,...) for N > 600 will cause a subscript error if the file fills up and no match or blank is found within the first 600 entries.
• The quit confirmation at lines 2016–2022 reads INKEY$ twice in sequence without looping between the “Y” and “N” checks, meaning the second INKEY$ call at line 2018 may capture a different (or empty) value than the one that satisfied line 2016. This can cause neither branch to fire, falling through to the unconditional GOTO 70 at line 2022.
• Line 2516 prints an error for Scott numbers longer than 5 characters but does not loop back to re-prompt; the program continues to the GOSUB 5000 at line 2520 regardless.
• The search loop at line 3038 iterates FOR N=1 TO 600, consistent with the array size, unlike the adjustment loop. However, blank-row detection at line 3040 terminates the scan early, so stamps entered after a gap of blank rows would never be found.