Correcting text

Why the text needs to be corrected

Trove searches across the electronically translated text. Much of this text has not been translated accurately due to the poor quality of old newspapers. By correcting the electronically translated text, Trove is immediately able to retrieve better search results for users. Corrections should be made to match the text that was originally published (without formatting).

As a measure to reduce the impact of malicious changes to text, corrected text does not overwrite the original text of the article. Both the corrected text and the original text are indexed and searched.

How to correct text in the digitised newspapers

Correcting the electronically translated text refers to changing the text in the left hand column to match the text as it was originally printed. Text corrections are saved to the database and will improve the service for all users by increasing the accuracy of search results. Below are instructions for how to correct the text. If you're wanting inspiration for where to begin fixing text, there's been discussion in the forums about finding articles to fix

Only the text as it appears can be changed. Changes that do not improve the accuracy of the text for searching (e.g. specifying formatting) are not currently supported.

Getting Started

  • To open the edit mode, click Fix this Text. You will then be able to make your changes.
  • When you have finished click Save & Exit to save your changes and exit the edit mode.
  • Click Save if you wish to save your changes and continue editing. For long articles it is recommended that you save every 10-15 lines.
  • Click Cancel to discard changes made since the last Save and exit the edit mode.
  • Single letters, words, lines or entire articles can be corrected.

You can correct the text without logging in with a Trove username. You will need to fill in a short 'captcha' test before your corrections are saved. If you login to Trove your corrections will be recorded in your profile and you will appear in the text correctors Hall of Fame. You will not have to complete the 'captcha' test.

In older newspapers, the long ‘s’ (appears as an ‘f’ without the crossbar) may be replaced with a lower case ‘s’. 

Special Characters

To add special characters (e.g. £) to the electronically translated text use the special characters button.

  • For characters not in the list:
  • Windows users:
    • Open Character Map (Start -> Programs -> Accessories -> System Tools -> Character Map), navigate to the character, click on the character, click ‘select’, then copy and paste into the electronically translated text; or
    • If there is a keystroke displayed in the lower right corner of the character map, e.g. Alt+0163, hold down Alt and type the numerals on the numeric keypad to insert the characters into the text.

Macintosh Users:Place the cursor in the correct location, open ‘Special Characters’ (Edit -> Special Characters), navigate to the character, select the character, click insert. 

Keyboard shortcuts

To speed up the editing process, you can use keyboard shortcuts to navigate through the text and save your changes. Keyboard behaviour depends on your browser but in general:

  • TAB allows you to move to the next line while Shift-TAB allows you to move to the previous line.
  • Enter allows you to move to the next line.
  • Up and Down arrow keys also allow you to move from line to line.
  • Left and Right arrows allow you to move within a line or a text area. Use the Home and End keys to position the cursor at the beginning and end of the input field respectively.
  • F12 allows you to move from word to word.
  • Windows: Alt + s focuses the Save button, Alt + e focuses the Exit/Save & Exit button and Alt + x focuses the Cancel button. Hit Enter to perform the action. (These rely on browser “access key” functionality and may not work with all browsers.)
  • MacOS X: Ctrl + s is equivalent to Save, Ctrl + e to Exit/Save & Exit and Ctrl + x to Cancel.

Missing Text

If you come across an article with missing text, please add the missing text to the preceding line. As you type you may reach the text limit, click the add missing text link to allow you to add any missing text into the current text field (this feature is for registered users only).

Multi-page articles

You can only correct mistakes on the page the system is showing. If an article spans over 2 or more pages, save any changes on the current page then go to the appropriate page by clicking on the On page… link before correcting that page.

Searching on your changes

Although your changes will be immediately applied to the article, it may be several hours before the search index is updated. Hence your updates may not be immediately found by a search.

How to add new lines of text

It is not possible to insert additional lines of text. When logged in to your account and correcting the text, you can extend lines of text by adding the missing words to the end of the preceding line. If there is no preceding line, add the text to the start of the following line. Where possible make sure that the start of each line matches the start of the original line of text.

Deleting blank lines from newspaper articles

Currently there is no way to remove blank lines from articles.

How long it takes for corrections to be searchable

Due to the re-indexing process of the database, it may take up to 5 minutes before a search on the new, corrected text finds the corrected article.

Adding documents to Trove when the text can't be edited

Unfortunately it is not possible for us to import text supplied to us as a document into Trove. We are not able to use any documents you send us.

Why the electronically translated text is corrected on a line-by-line basis

Each line of the electronically translated text is linked to a location on the article image. Line-by-line correction maintains these links, allowing clicking of the text to move the article image to that location on the screen (synchronisation).

The difference between correcting the electronically translated text and adding comments

Correcting the electronically translated text makes changes to the electronically translated text in the database and improves accuracy of searching across terms for other people. The electronically translated text should exactly match what is printed on the newspaper page.

Comments allow users to provide further information about the events described in an article, or state if they believe there is journalistic error in the original article. Both comments and the electronically translated text are searched.

Electronically Translated Text

The Electronically Translated Text has been created through the process of Optical Character Recognition (OCR). This is an automated method of converting text contained in images, which cannot normally be searched, into text that is fully searchable. For more information on OCR and newspapers, see the Optical Character Recognition Overview on the Australian Newspapers Digitisation Program webpage.

Each line of the electronically translated text is linked to a location on the article image. Line-by-line correction maintains these links, allowing clicking of the text to move the article image to that location on the screen (synchronisation).

The difference between correcting the electronically translated text and adding comments

Correcting the electronically translated text makes changes to the electronically translated text in the database and improves accuracy of searching across terms for other people. The electronically translated text should exactly match what is printed on the newspaper page.

Comments allow users to provide further information about the events described in an article, or state if they believe there is journalistic error in the original article. Both comments and the electronically translated text are searched.

Spell check when correcting text

The text correction module does not have a built-in spell check facility. Spell checking in this module is performed by your web browser. If you are noticing it is offering American spelling options you will need to check which language your browser is using to check spelling. Also, you need to be aware that the Electronically Translated Text is corrected to match the text originally printed on the page, so you may find articles that contain 'American' spelling of some words.