Entering your search term(s)
- Any number of search terms can be
used.
- If you enter more than one search
term choose the “AND” option to search for cards which include all of
these terms, or “OR” to search for cards which include any of these terms.
- Phrase
searching is not available, so it is not necessary to place quotation
marks around phrases.
- Truncation
searching is not available, so you can’t use a wildcard symbol like * or ?
to find a term and its variant endings.
Too many hits, or hits are not accurate enough?
The search results are limited to 10,000 hits. If your
search has resulted in more than 10,000 hits, this will be indicated.
To make your search more accurate or precise,
you could:
- Use more than one search term
with the “AND” option
For
instance, a search for "Wight" produces lots of results, many
not relevant, but if you search "Isle Wight" the results are
more accurate.
- Adjust the fuzziness of your
search upwards (see below)
- Use the Browse option instead
Too few hits, or not finding something expected?
To
find more hits, you could:
- Use more than one search term
with the “OR” option
You
can use this as a way of searching for synonyms or variant spelling.
- Adjust the fuzziness of your
search downwards (see below)
This
may make it to possible to find a difficult to read card that the Optical
Character Recognition (OCR) had problems with
Adjusting the fuzziness (similarity) of your search
The precision or fuzziness of a search can be adjusted.
Currently it is set to a default value of 90% where 50% would give you the
widest set of results and 100% would give you the narrowest.
The default value of 90% can be overridden.
Because of the way the fuzzy works using a higher value may exclude items which
have the correct term on the card but the OCR has interpreted it incorrectly.
For example:
"Engel"
Similarity 100% - is the most precise search.
"Engel" Similarity 90% - is the current setting and is the same as
entering “Engel” only.
"Engel" Similarity 70% - is less precise and may find extra
relevant cards which have had your terms mis-read by the optical character
recognition.
(Lower numbers can be used but are likely to give
large numbers of cards that are not relevant.)
About Optical Character Recognition (OCR)
The full text search has been enabled by a process called
optical character recognition which looks at the groups of characters on the
card image and converts the data into something that can be searched.
The fuzzy search means that even if the
characters have not been correctly recognised the system can interpret the data
and give you some results.
(If a whole book is being digitised the optical character recognition can be
adjusted to allow for inconsistencies in the typeface e.g. an indistinct
character, because it will occur throughout the document. Because our cards
were typed and produced on many machines and by several methods this
"learning" part of the OCR process cannot be used.)
The
optical character recognition does not always recognise a word or term but the
fuzzy search will display all near matches.