inputing fuzzy dates

I'm wondering what is the general thought for databases handling fuzzy dates -- dates that use "circa", "approximately", "yyyy?". Didn't have a lot of success finding general database design information on this topic.

Even though I now realize that Omeka's date boxes don't seem to have a max length, one concern would be eventually sorting by date. I don't want all the "circa" to be clumped together, when they should logically be dispersed by the numerals. Does anyone know how ContentDM or any museum software handle this problem?

As a relative newbie to PHP and database design, my solution was to add a custom field through the Omeka admin called "Date Certainty". Here, my input rules are that it should only be filled with the terms Circa, Approximately, Before, After, Probably. The date value (the numbers part) still goes in the DC Date field. Then on the public display, I've concatenated the two fields to display together -- "Circa yyyy-mm-dd" or potentially, "After April 7 1921".

I like the idea that they are logically split apart in the database, but I'm not sure this is the ideal way to handle it. I got the idea from the EAD XML spec which has a "date certainty" attribute. Am I making this more complicated than it needs to be?

Hi kkay,

One option to the date field is coverage. The .10 beta uses the general "coverage" Dublin Core field that can be interpreted for temporal or spatial coverage. Meaning, there is an open text box where you can enter "Circa 1940", or "After April 7, 1921", et al. The challenge with that will be instructing your end users that they may find date information in that field.

Hope that helps a little.
Sheila