TL;DR: Smart quotes, non-breaking spaces, rich-text bullets, and PDF line breaks corrupt paste input. Paste through plain-text Notepad or use paste-as-text shortcuts before encoding.
Smart quotes from Word
Curly “hello” may not map—tool skips unknown characters silently or drops words. Replace with straight ASCII quotes or remove.
Non-breaking spaces
Look identical to spaces but break tokenizers. Find-replace NBSP (U+00A0) with regular space in editor showing invisible chars.
Rich text from websites
Copying article paragraph includes hidden span tags if pasted into rich editor then copied again. Strip to plain text once.
PDF column paste
Line wraps mid-word: transfor-
mation. Join hyphenated breaks before encode.
Trailing whitespace per line
Batch paste from Excel adds invisible tab—encode odd empty groups. Trim cells first.
Mobile paste from SMS
Auto-correct replaces I with l in some fonts—verify letter I vs number 1 in decode input.
Fix workflow
Windows: paste into Notepad, re-copy. Mac: Paste and Match Style or Shift+Option+Cmd+V into TextEdit plain mode.
Verify after cleanup
Character count in Notepad vs Word should match visible letters. Then run letterstonumbers.com.
Prevent in worksheet design
Distribute student sheets as plain PDF text layer or .txt download—not screenshot images requiring OCR.
Browser extension conflicts
Grammar plugins altering clipboard—disable temporarily if output differs across browsers with same input.
Fix workflow plain text
Windows Notepad or Mac TextEdit plain mode strip before encode—smart quotes and NBSP eliminated.
Excel batch trim cells
Trailing tabs cause empty encode groups—TRIM column before batch paste into converter.
Mobile SMS autocorrect
Verify letter I vs digit 1 when copying decode input from text messages.
Browser extension conflicts
Disable clipboard managers temporarily if paste shows plain despite correct copy.
Worksheet design prevention
Distribute .txt downloads not screenshot images requiring OCR that drops spaces.
Curly quotes wrapped around FIND THE KEY
A worksheet exported from a word processor often turns FIND THE KEY into a sentence wrapped in curly quotes. The quotes are not letters. A converter may skip them, or it may refuse the line, or it may leave a blank token. The letters you want are FIND, THE, and KEY. FIND is 6 9 14 4. THE is 20 8 5. KEY is 11 5 25. Paste the sentence into a plain-text editor first and replace the curly marks with nothing. Then encode. If you encode the rich-text version, you are testing the editor, not the student.
Show the two pastes on a projector: one with curl quotes visible as slanted marks, one with the marks gone. The number groups should be identical after the cleanup and messy before it. Students who copy from the slide will copy whichever version you left selected. Leave the plain version selected. The curly version is the demonstration, not the source.
A non-breaking space hiding inside HELLO WORLD
HELLO WORLD copied from a web page sometimes contains a non-breaking space. It looks like a normal gap and it does not break the line, and some tools treat it as a character they cannot map. The encode you want is 8 5 12 12 15 for HELLO and 23 15 18 12 4 for WORLD. If the output is one glued group or an error in the middle, paste into a plain editor and retype the space with the keyboard. Do not keep deleting invisible characters by guessing.
A practical test: move the cursor through the gap with the arrow keys. A normal space takes one press. A weird gap sometimes takes one press and still fails to encode. When that happens, delete the whole gap and type a fresh space. Then encode again. The two groups should return. If they do, the alphabet was never the problem. The page was.
Notepad hop before a science list is encoded
Take the eight-word science list you already trust and pass it through a plain-text hop even though it looks clean. The hop is: paste into Notepad or an equivalent plain editor, select all, copy, paste into the converter. OSMOSIS should still lead the list and PIGMENT should still end it. This hop is boring on a good day and it is the whole repair on a bad day. Make it the default, not the emergency step, so students do not have to decide whether their paste is “bad enough” to deserve cleaning.
If the hop changes the output, keep the plain version as the file you submit. Do not keep both. Two files become two answers. Name the file with the suffix “plain” if you need a reminder, and delete the rich-text copy after you confirm the first and last words. The confirmation is PHLOEM’s numbers, 16 8 12 15 5 13, which you can check in under a minute because you have checked them before.
Selected-text check after the curly quotes come off
After you strip the quotes around FIND THE KEY, select the plain sentence and look at the character count your editor shows, if it has one. The three words plus two spaces are 12 characters. FIND THE KEY. If the count is 14, two quote characters survived. If the count is 13, one survived. Delete until the count is 12, then encode. The groups are 6 9 14 4, then 20 8 5, then 11 5 25. A count is less poetic than a careful eye and much harder to fool.
Students on phones may not have a character count. They can type the sentence themselves under the pasted line and encode the typed line. Typing is the phone version of the count. It is slower, and it removes the invisible characters the paste brought in. Keep the typed line. Discard the pasted one once the groups match the three known rows.
