Transkribus

We are still transcribing abbreviations as they appear in the text, but knowing what abbreviations stand for can be helpful and provide context clues. Here are some common ones people have found:

• do = ditto
• n.d. = no date
• people's initials (usually in VIA column)
• amal = amalgamated

If there are some common ones you come across and/or are some that don't seem intuitive, feel free to let us know at digitize@ucalgary.ca and we'll add it to this list.

Yes. In the Transkribus desktop application you can use the virtual keyboard to add special characters or use alt code (see the checkmarks FAQ for a tutorial on adding special characters to the virtual keyboard). In the web version, you can use alt code (for example, è is alt0232).

Yes, you can transcribe the covers. You can create a table around any text and then run layout analysis. You can ignore the typed text found in the inside front cover about the book's physical specs (page and column count, etc.).

Yes, transcribe the text and then mark it with a strikethrough. If a whole entry has been crossed out, add a strikethrough to all text included in this entry.

Yes. For example, if the layout analysis includes some checkmarks from the columns at the end of a page we are skipping, you can delete those baselines. Hold the ctrl key and select all the baselines you wish to delete, and then delete them with the "remove a shape" button found in the vertical toolbar.

Sometimes there are extra notes or stamps on the left side of the page that correspond to the rows. To include these notes, when creating your table include this section as a separate column (it will just not have a column heading). This will create cleaner data when exporting it later on, as opposed to the notes being in their own text boxes or a separate smaller table.

It depends! The order of the table cells of whether the transcription order follows a single row at a time or sometimes switches back and forth between different rows does not make a difference, as when we export the data into spreadsheets, the transcription will appear in the right order as a table.

The order of baselines within a single cell does matter, as that order is how the text appears. For example, if there is a note that is written on 3 separate lines and so is made of 3 baselines, make sure the order of the baselines follows that of the original note.

Overall, make sure that the baseline is in its correct corresponding cell (for example, that the date appears in the date column cell, etc.). 

At the end of Job A (adding tables and baselines), the page status is in progress.

At the end of Job B (transcribing), the page status is ready for review (web version) or done (desktop version). 

At the end of Job C (final review of the transcription), the page status is done (web version) or final (desktop version).

After you change the page status, make sure to hit save. You will have to refresh for the status change to appear.

Punctuation that is important to the accuracy of the information we are transcribing should be marked up in the baselines and transcribed. For example, the dashes in the file numbers are important and should be included. 

However, if, for example, there is a period at the end of a sentence that is not crucial to the data, don’t worry about fixing baselines to cover that period. If the baseline does already cover it though, still transcribe it.

Sometimes the layout analysis will not add baselines to text written in pencil, as often the pencil is more faded. Add a baseline to include this text as often it is important and should be transcribed.

There are some exceptions: see the pencilled numbers FAQ for one.

Some registers have a pencilled number between the Source and VIA columns. In later volumes these numbers have their own column (which we would then transcribe), but for these numbers that are not in their own column, we won’t worry about transcribing.

This tutorial shows how to select all the cells in a table as well as any other text regions on a page in Transkribus.

This tutorial shows a quick method to split baselines using the vertical line tool in Transkribus.

Yes, transcribe stamps. And if, for example, a stamp was not stamped perfectly and some letters are missing, just transcribe the complete word(s) if you know what it should be. Some common stamps read: transferred to library, oddments, posters, cards, or menus, for example.

Often times stamps are on the left on the other page, and each stamp corresponds to the row it's beside. Extend your table to include those as well as another column (it just won't have a column heading). That way the stamps are transcribed in the same table as the rest of the text, as opposed to doing separate text regions around each of them.

Sometimes a row of text will have, for example, a "ditto", a quotation mark, or a dash. that refers to the line above, instead of writing out the same repeated text. When you are transcribing, transcribe it as you see it, whether it's ditto, a dash, or quotation mark. Later, in Job D, we will clean up this data by replacing the ditto or quotation mark with the complete corresponding text of the line above.

You can go to https://ucalgary.yuja.com/V/PlayList?node=268209&a=1272099729&autoplay=1 to find all the Transkribus tutorials in one place.

Check your browser. Google Chrome can be problematic when playing these videos, so try another browser such as Microsoft Edge. If you continue to have technical difficulties, please contact digitize@ucalgary.ca.

Watch this tutorial to learn how to go back to different versions or steps you completed on a page in the desktop application of Transkribus.

A note to add: When you are in the version tab, if you double click a version, it will load the page back to that version. That way you can see what you did in each step.

UPDATED: Include that year in the table. The best way is to extend the table so that the year is included with the other column headings. Then the year is transcribed in the same cell as the column heading below it (for example: 1987 Date Ent'd). This is preferred over creating a whole separate column on top of the headers for just the date.

Overall, important information like a year (especially if it's not included elsewhere in the date column) should be put in the main table and not a separate text region, as this will ensure it doesn't get missed in the final export.