June birthdays by day: notable people organized by date

Two people can read the same six characters and get two different days. That is not an exotic failure. It is the ordinary condition of a numeric date moving between countries, and it lands hardest on a page that claims to know which day something happened.

What to take away

• A numeric date with a day at or below twelve has two readings and no way to tell them apart from the digits.
• Half of the year is therefore ambiguous in a form that most of the world writes without thinking.
• The fix is a written convention plus a stored format label, not a guess about where the source came from.

Where the ambiguity lives

Write a day and a month as numbers separated by slashes and you have made a claim in a convention you did not state. Day first is standard across much of the world. Month first is standard in some of it. The conventions differ by country and always have, so a bare numeric date carries no evidence about which one applies.

The ambiguity is not spread evenly. When the day number is above twelve there is only one reading that makes sense, and the date resolves itself. When the day is twelve or below, both readings are valid days. That is a large share of every month, and it means a substantial fraction of every numeric date in circulation could be either of two days without looking wrong.

A stamp like that prints whatever convention its owner set it to and says nothing about which one. So does a spreadsheet cell, an old index card and a scanned form.

Why guessing does not work

The tempting fix is to infer the convention from context: the country of the source, the language of the page, the other dates in the file. Each of those is wrong often enough to be useless.

The guess	Why it fails
Infer from the source country	Documents cross borders, and international bodies mix conventions inside one file
Infer from the language	The same language is written with different date conventions in different places
Infer from other rows	A file assembled from several sources has no single convention
Infer from which reading is a real date	Both readings are real dates for most of the month
Assume the exporting system normalized it	Exporters frequently reformat into the local convention without saying so

None of these are bad ideas because they are lazy. They fail because the information required is genuinely not present in the digits.

What to do instead

The answer is dull. Write dates in a form where the fields cannot be confused, and record which format a value arrived in.

ISO 8601 exists for this. It puts the largest unit first, uses four digits for the year and fixed widths for the rest, and produces a string that sorts correctly as text. It is not more readable to most people, which is why it belongs in storage and interchange rather than on the page. Convert to a friendly form at the moment of display, from a value that was never ambiguous.

Alongside the value, keep a note of what the source looked like. An entry that says the source wrote a short numeric date, convention unknown, is more honest and more useful than one that shows a confident day. It also tells a later editor exactly what to check.

Reading a day page in this month

• A day shown with no indication of where it came from is a day someone chose. That choice may have been made by a parser.
• Where two sources disagree by an amount that swaps day and month, the disagreement is almost certainly a format problem rather than a factual one.
• An entry whose day and month are both twelve or below deserves more skepticism than one where the digits resolve themselves.
• The month bucket is safer than the day, because both readings of an ambiguous date rarely fall in the same month, so a page organized by month inherits less of this problem than one organized by day. The June index works through what a month can carry.
• The general principle, that an entry should record its source and its confidence rather than a bare value, is the one the January index is built on.

Common questions

Can software detect the convention automatically?

Sometimes, for a whole file, if it contains at least one date with a day above twelve and the file really does use one convention throughout. Neither condition is guaranteed, and a detector that is right most of the time produces errors that are impossible to find afterward.

Is a written month name enough?

It removes this particular ambiguity, which is why it is worth doing. It introduces its own problems with language and abbreviation, and it does not help a machine sort. Store the unambiguous numeric form and print the month name.

What about a date that is only a month and a year?

That is a partial date, not a broken one, and it should be stored as what it is. The January by-day pages set out how a partial value should be recorded so that a rendering step cannot invent the missing field.

Does any of this change what a birthday index should publish?

It strengthens the case for publishing less. A month survives every failure described here far better than a day does, and for a public index the month is enough to place an entry. The February index explains why the finer value is a key rather than a fact.