Once each purple line of the letter has been transcribed three times, it turns gray. The letter transcription is then complete and is considered “retired.” This occurs once the user answers “yes” to the question "Have all the volunteer-made underline marks on every page turned gray?" that appears to the right of the transcription window. [Figure 1]
After the letter is retired, it is then stored on the backend component of Zooniverse—ALICE. Here the transcription lines are compiled as data for correction before export. Each line is assigned a color which, when clicked on, displays the contributed transcriptions. There is the potential for user error if the user incorrectly marks the letter as complete, evident from a few letters which appear in ALICE despite having few lines transcribed. These I am tracking in an “Unretire” document to send to the Zooniverse team as there is no way to do this on our end.
Although some letters remain untranscribed, there is a significant sample set available for correction within ALICE. Cleaning up and refining this data is the next step of project progress that I am tasked with. Each line—differentiated by color—stores the contributed transcriptions, followed by an aggregate which combines all three into a potential option. Below this is a blank line to make any corrections in the user-contributed lines. Each contributed transcription has a checkbox which, when selected, indicates the final correction. [Figure 2]
To streamline the correction process, I usually copy the aggregate into the box and make edits from there. To mitigate factors which might discourage users from sustained interaction with the project, we did not require the editorial corrections usually made by the project team as they break up the “flow” of transcribing. Editorial corrections primarily entail following words and characters with modern spellings or expansions in brackets. The latter instance is relatively frequent as letter writers employed abbreviations and abbreviated characters for the sake of both time writing the letter and space available on the paper. One of the most frequent abbreviated characters to appear throughout the letters is the thorn—a character commonly used at the time, but one that no longer appears in modern usage. Written as a “y,” it represents the letters “th.” As such, that which appears in the letter as “ye” is transcribed and followed by a “[the]” correction; “yt” is corrected with “[that].” [Figure 3]
While not requiring users to make editorial corrections eases the transcription process for public audiences, it does complicate the processing of crowdsourced transcriptions as it essentially necessitates intervention with each line to make these corrections. These corrections are further complicated when made using the ALICE platform. Figure 4 shows how visually complicated the ALICE interface can be.
When necessary to delete redundant or insufficient lines, each subsequent line is negatively impacted.
Figure 5 shows a line which must be deleted. Figure 6 shows how the following line now displays stored transcriptions from the previous, now deleted, line (now the dashed orange line). Waiting until the end to delete lines is not always ideal. Transcriptions with multiple redundant lines muddy the waters and make it difficult to know which is the line intended to be edited, those which are correct, and those yet uncorrected. Outlining the problematic aspects of the correction process can highlight barriers to usability and present issues for eventual correction by the Zooniverse development team. Sharing usability experiences with transcription platforms and processes helps foster more empowered contributors and engaged audiences.
Comments