Compare Text

Who compares texts

Anyone holding two versions of something and needing to know exactly what changed.

• Writers and editors

  What a co-author or an editor changed in a returned draft, word by word.

• Contracts and policies

  Whether the clause sent back for signature is the clause that was agreed.

• Developers

  Two configuration files, two API responses or two log lines, characters marked, and the zero-width space that broke a build named.

• Data and spreadsheets

  Why two cells that look identical do not match in a lookup: a no-break space or a trailing space, named.

• Translators

  A source text against its updated version, to translate only what changed.

• Teachers and students

  Two versions of an essay or an answer, re-wrapped text compared as running text.

• Security checks

  A web address or a payee name against the real one, look-alike letters named.

• Subtitles and transcripts

  Two caption files, line by line, after a re-edit.

Why lines that look the same are marked as different

Every text compare marks a line that differs. The question it leaves open is the one people ask most often about the result: why is a line marked when the two versions look exactly the same? The answer is nearly always a character that prints as nothing, or as something you cannot tell apart from the character you expect. A no-break space looks like a space. A zero-width space looks like nothing at all. A Cyrillic а looks like a Latin a in most fonts. A comparison is exact, so to it these are different characters, and the line is different.

This compare does two things about that. In the comparison, a hard-to-see character inside a change is drawn as a small labelled box, NBSP or ZWSP, and a change made only of spaces is drawn as dots. And under Differences you cannot see, every line whose only differences are characters like these is listed, with each character named, its Unicode code point, and where in the line it is.

The example, worked. Press Try an example and two versions of a shop’s opening hours go in. Two lines are visible edits: 5pm became 6pm, and 2pm became 4pm. Two more are marked although they read the same. Line 5, Call us on 020 7946 0000., has a no-break space between 020 and 7946 in the changed text, the kind a word processor puts in a phone number to keep it on one line. Line 6, We’re closed on bank holidays., has a curly apostrophe in the original and a straight one in the changed text. The list names both: the original has a space (U+0020) where the changed text has a no-break space (U+00A0), at character 15.

Whether such a difference matters depends on the text. In a contract or an essay it usually does not, and the Ignore settings below take most of them out. In a password, a product code, a spreadsheet lookup, a line of code or a web address it is exactly the difference that breaks things, and it is the one the eye will never find.

What each Ignore setting treats as the same

By default the compare is exact: one character different is a different line. The five Ignore settings each say that a kind of difference does not count. They change only what is treated as the same; the text shown is always your own, unchanged.

On the example, the settings take out exactly the differences they name:

Ignore spacing removes the no-break space; the quotes setting removes the apostrophe; capital letters change nothing, because none differ. With all five ticked, the two lines left changed are the two real edits to the hours, and Sunday and the parking line are still removed and added.

Copying the differences, and opening files

Copy the differences puts the comparison on your clipboard as text, in a format other tools read. Compared by lines, it is a unified diff, the format of diff -u and git diff: a line starting - is in the original only, + in the changed text only, and a space means unchanged; each block starts with @@ and the line numbers it covers. The example copies as:

--- Original
+++ Changed
@@ -1,6 +1,6 @@
 Opening hours
-Monday to Friday: 9am to 5pm
+Monday to Friday: 9am to 6pm
-Saturday: 10am to 2pm
+Saturday: 10am to 4pm
-Sunday: closed
-Call us on 020 7946 0000.
+Call us on 020 7946 0000.
-We’re closed on bank holidays.
+We're closed on bank holidays.
+Parking is free after 6pm.

Compared as running text, it is git’s word-diff form instead: removed words in [- -] and added words in {+ +}, in place, as in 9am to [-5pm-]{+6pm+}. With an Ignore setting ticked, the copy shows the comparison as you see it, so it may not apply cleanly as a patch.

Opening files

Each box has its own Open a file button for plain-text files: .txt, .md, .csv, .json, .xml, .html, source code, logs and subtitles. The file is read on your device, as UTF-8 or, if it starts with the marks for it, UTF-16. A file opened here keeps its byte order mark, so a file saved with one and a file saved without compare as different on their first line, and the compare names the BOM, which is the usual reason a desktop diff marks the first line of two identical-looking CSV files.

Line endings are reported, not compared. Windows ends lines with two characters (CR LF), Mac and Linux with one (LF). A text box in a browser turns every line ending into LF as the text goes in, so the compare cannot see them; instead the note under the boxes says which endings the opened file used. Word, PDF and spreadsheet files are not read: copy their text in, or, for a Word file’s tracked changes, use the edit counter, which reads them.

More writing tools

Every writing tool on the site runs in your browser, needs no sign-up, and sends nothing anywhere.

Related guides

Longer answers to the questions behind the numbers.

• Reference

  Types of counters explained

  Synchronous vs asynchronous, mod-N, ring and Johnson, industrial and PLC counters, plus mechanical and software counters.

  11 min read
• Counting

  Why word counts differ between tools

  The same sentence is 26 words to one counter and 39 to another. Four counting rules side by side, and how to meet a hard limit whichever one a form uses.

  11 min read

All guides

People also search for

• compare text
• why identical-looking lines differ
• how the texts are compared
• how similarity is worked out
• edit counter
• word counter
• character counter
• why word counts differ