Text correction guidelines

Purpose of text correction

The primary purpose of text correction is to improve the accuracy of search results in Trove’s newspapers and gazette zones. Text correction also allows for more accurate transcription downloads of the article text.

Text correction is not intended to aid in the accurate aesthetic reproduction of an article. As such the ability to update formatting styles such as italics, bold and SMALL CAPS has not been provided.

General guidelines

The transcription in the left-hand column should match the text as it was originally published. Where words are misspelt, or there are typesetting errors, the correct spelling should be added as a comment to the article.

On occasion it is not possible to insert the same characters used in the publication, these characters result in a term that is not searchable. The text may be updated to represent a character that can be searched when there are:

  1. Upside down letters, these should be transcribed as the correct characters. Add a comment notifying other users that the line/character is upside down when it is important to the flow of the text (when a letter has been inserted upside down within a word then this does not need to be mentioned, if the upside down text is the solution to a puzzle then this should be mentioned).
  2. Transcribe the ‘long s’ or ‘ſ’ (as appears in many early publications) as a lower-case ‘s’.
  3. Transcribe names starting with M’ as Mc (where appropriate), or add a comment with the corrected spelling.
  4. Insert obviously missing letters – for example if you have Melb urne the ‘o’ may be inserted.
  5. Minor errors, such as letters being transposed in non-name words, where the word appears with the correct spelling elsewhere in the article, may be corrected.

Any other situation – including, but not limited to, inserting full names instead of initials, correcting article content, or inserting [sic] –  should be added as a comment or a tag on the article, not through changes to the text. DO NOT correct spelling for words that are names (either personal or place), as these may reflect the spelling of the name at the time.

Example: If the word ‘Melbourne’ is spelled ‘Melbawn’ in an article, the transcription should contain ‘Melbawn’ and a comment saying ‘Melbawn = Melbourne’, or ‘Melbourne spelled as Melbawn’, or similar, should be added.

Do not delete text that exists in the article.

Punctuation and formatting

Punctuation and capitalisation should reflect what is published in the article. The following exceptions apply:

  1. Small Caps may be transcribed as either all uppercase or as mixed case.
  2. Hyphens indicating words split across lines may be left in (as per the original), keeping the word split across two lines; or removed with both parts of the word combined onto the end of the first line. In both of these cases the word will be correctly indexed for searching. Do not remove hyphens that occur elsewhere in the text (e.g. George-street keeps the hyphen)

Text location

When text correcting in Trove, the text should be maintained in the lines where they originally appear. When multiple lines have been highlighted as a single line these may be split into the appropriate number of lines using the +/- tool.

 If the text appears in the incorrect order in the correctable text, the original processing order needs to be remain, however use '||' at the beginning and end of any block of text that is in the incorrect reading order. This can also be used to indicate where a caption, or other text, appears within an article. Add a comment to the article to indicate where this has occurred.

Missing text

Text can be added or removed in the left-hand column by using the +/- tool. To add or remove lines, wave your cursor over the end of the line you are currently using and then wave down. Select + to add a line below or - to remove your current line.


Due to the nature of the OCR process, tables may appear in the transcription as either each line read left-to-right across the entire table or as columns one after another. Correct tables as they have been originally created – if the table has been transcribed as left-to-right across the lines use '|' to separate columns.

Incorrectly identified columns

Where columns have been incorrectly identified (i.e. text from multiple columns appear in a single line), transcribe the text in the correct reading order, using the ‘add line’ function to add the lines from columns after the first to the end of the first column's text.

Non-alphanumeric text

In some cases non-alphanumeric characters will appear in the transcription where the OCR engine has tried to process non-text (e.g. a picture) as text. These characters can either be ignored or removed as they do not impact on searching.

Text in incorrect article

If the boundaries of an article have been incorrectly identified, and text is included that belongs to another article in Trove it needs to remain in the current article. Add a comment to explain that the text should be separate.

Special Characters

A selected set of special characters can be added using the ‘Insert Symbol’ menu option. For characters that are not included in this list use the Character Map (Windows) or Character Palette (Mac) to copy and paste these characters into Trove.

Non-English language newspapers and articles

Some articles in Trove appear in languages other than English. Text correction of these articles follows the same rules for correction as articles in English with the following exceptions:

  1. Chinese language articles may appear in any of the following orientations: Left-to-right horizontally, right-to-left horizontally, right-to-left vertically. Correct these to the modern left-to-right horizontal reading order as this will give the best results for a Trove search
  2. Transliteration or Translations should be provided as Comments, not in the article text.

Abuse/Spam/malicious corrections, etc.

If you discover corrections that are clearly not part of the original text feel free to correct them back to the original text. If the corrections contain links, swearing, or occupy significant amounts of the article please report using the Contact us page and these will be rolled back (rolling back corrections removes them from the correction history as well as reverting the corrections to an earlier stage).

Text correction counts

Text correction counts are based on the amount of lines corrected and are updated immediately in your profile. Changes may take a little longer to appear in Trove searches.

Hall of fame

The Trove Text Correction Hall of fame is a running tally of corrections made by the Trove community. Corrections are added to a total once they are indexed in Trove. The Hall of fame leader board lists the top correctors of all time, from Australian Newspapers Beta to the present day. You can also filter the leader board by month or by publication type.