Translation UX/Design feedback 5

The goal of this round was to test:

• The prototype for a translation homepage
• The current implementation of the translate editor

Test Observations

Test with User #66 on 14 Jan

Test with a Polish user speaking Polish, English, German, and Russian.

Translation Workflow

• 00:10: The user is asked to find a project and follows the usual workflow: go to the homepage, select a project and click on "translate this project" from the project description page.
• 00:56: The user is able to change the table filters to find "untranslated messages".
• 01:15: The user notices the dynamic loading of translations
• 01:24: The user selects "untranslated" filter but when loading dynamically, also translated messages have been added.
• 02:45: The user tries to confirm an outdated translation by clicking on the save button which is shown as disabled. The user deletes and adds some characters at the end of the translation to overcome the limitation.
• 03:15: The user comments that "it looks very good".
• 03:40: When asked about the project selector, the user discovers that "what is surprising and I like is that you can select sub-projects directly from breadcrumbs".
• 04:11: The user comments that it is nice that the editor moves to the next translation after saving.
• 06:12: The user comments that suggestions were clear and it is a needed feature.
• 06:46: When asked about the reason for skipping some The user understands the word but don't know which of the possible translations is better to use in the specific project.
• 07.55: When asked, the user finds the "expand" interesting to get a bigger editing field but the user find suggestions positive and would prefer not hiding them.
• 08:30: A suggestion was not completely correct (was in singular instead of plural). The user reads the translation but writes the text himself.
• 09:52: The user opens the last part of the breadcrumb getting an empty selector, he was expecting to access sub-groups.
• 11:33: The user is able to navigate and select translatable pages but all where translated into his native language.
• 12:45: The use finds cool the possibility of using shortcuts but considers the proposed key combination too complex.
• 12:20: The user notices that the filters are not working properly.
• 20:25: The user makes use of the ULS to change the language. From the project page the global ULS is used for doing the change.

Homepage

• 15:21: The user first impression was "Wow! This looks good! It looks like a modern UI. That's how modern websites mostly look".
• 15:50: The user asks for a way to translate a project from the project list. The user does not notice the link probably due to being hidden by too many hovering events.
• 16:45: The user comments that he would be using the "22 projects" tile as a way to search for projects. So he would the tiles to be clickable.
• 17:27: It is clear how to become a translator.
• 18:03: The user identifies the search bar function and comments that he would expect a list of matching strings.

Test with User #R2 on 14 Jan

Translation workflow

Test with a user from India.

• 00:45: The user is shown the initial list of recent messages and is able to move to a different group. The progress bars are properly interpreted as translation completeness.
• 02:05: The user finds strange the "hide translated" options since he is just exploring the untranslated messages.
• 02:30: The user starts translating after exploring the editor (using expand action and reading the shortcut information). Contributes some translations and moves using the shortcuts.
• 03:36: The user asks for a way to copy the source message when he founds a string with several "magic words" (the user tried to find it under the secondary menu option, not implemented yet). When asked the user comments that translators with a language with a different language structure can also benefit from being able to use a copy-source-and-replace-parts approach.
• 07:20: The user asks for glossary support.
• 08:25: The user considers the most important things for translations to be: source text, target textarea, glossary , translation memory (with shortcuts and percentage matching), shortcuts to add placeholders (with keyboard shortcuts), and concordance search.
• 11:30: The user explains the importance of glossaries to provide consistency in the translations (e.g., file has five different translations for "file").
• 14:00: The user explains the concept of concordance search (detect sentences that are frequently repeated and suggest translations to the user).
• 17:37: When asked to move to a different project, the user clicks on the "All" filter instead of the project selector.
• 18:50: The user has problems locating the "Visual Editor" project since he does not know the structure of projects, but finally uses search for locating it.
• 20:49: The user copies and pastes the source message and starts translating different part.
• 22:55: The user explains that copy from source feature is also relevant to avoid users to switch language frequently with their Input method system.
• 24:20: The user is presented a design to allow source text copying and finds that "it is good if it can be commanded through keyboard shortcuts".
• 26:27: The user understands and uses one of the suggestions for a translation.
• 27:32: The user asks for a way to get the source context for the suggested translations, to check the difference between the current source and the suggestion source.
• 30:20: The user is happy to discover that translation memory instantly included a translation he made earlier.
• 31:40: The user considers that the new design is "very clear".

Homepage

• 34:34: The user first impression about the homepage is that "it is looking good".
• 35:42: The search bar calls the attention of the user and thinks that he purpose could be to search for translations to correct.
• 37:34: The user comments that it is also clear how to become a new translator.

Search

• 38:50: The user is asked to use search to find a specific translation.
• 40:00: The user expected the results matching the literal string (a Hindi text) to be shown on top.
• 42:02: The user finds filters to be ok, but comments that the priority of the results does not look ok.

Test with User #65 on 20 Jan

Test with a Malayalam user from India.

Translation workflow

• 00:30: The user selects a group of messages to translate, and chooses "Article feedback" extension.
• 02:20: The user edits an outdated translations (refers to it as "modified message") and uses copy-paste-replace approach to provide the translation.
• 04:18: The user comments that in previous version of the editor, an option to copy the source message was available.
• 06:00: The user selects "Mediawiki > Extensions > Article Feedback Tool" but the resulting path shown is "All > Article Feedback" and the user is confused about how to go back to the "Mediawiki" main group. In addition, the user is presented an empty project selector when clicking the last element in the breadcrumbs.
• 11:44: The user thinks that the new editor is "not much better than the older one, the look&feel is very good but in function is not better."
• 12:30 The user considers the table form to better than the other version.

Homepage

• 18:50: The user asked about the language auto-detection mechanism "I think it is I based detection".
• 20:00 The user thinks that it should be possible to correct and contribute translations starting from the wiki where the text is available ("using an extension, gadget, or so").

Test with User #R3 on 21 Jan

Test with a user from India speaking Punjabi.

Translate UX

• 01:50: It is not clear for the user how to switch to a different project, and the user tries to switch from project using the toolbar.
• 04:33: The user changes to a different project ("etherpad lite") and starts translating.
• 04:48: When trying to save using the shortcuts, the Firefox developer console opens by mistake probably due to some similar keyboard shortcut.
• 05:30: The user notices the Input method support and is able to use it since he is familiar with the mapping.
• 06:20: The user makes use of the "hide translated" button to clear current translated.
• 06:28: The user comments that he likes the idea of having keyboard shortcuts.
• 07:55: The user finds an "outdated" translation and is able to access the diff to correct the translation.
• 11:21: The user is able to identify the tabular structure of messages, the filters that can be applied and the translation aids that are provided on the right of the editor.
• 11:50: The user things that there should be "some shortcut to copy the source message to the translation".
• 13:10: It was not immediately clear for the user how to add the proposed suggestion to a translation.
• 13:56: The user considers that the warnings provided are "really helpful" as it is the description of the parameters provided at the right side.
• 14:35: The user comments that he is familiar with the input method and that it is helpful.
• 15:47: The user comments that he skipped some messages because they were too long or complex (with many parameters) to accomplish as many translations as possible.
• 16:40: The user is able to use the project selector but comments that the first time it was not clear whether the link was clickable or not and suggests that a tooltip may help.
• 19:29: The user thinks that the tool provides comparable features to offline tools.

Homepage

• 22:06: The user thinks that it is clear where to go to become a translator.
• 22:26: The user understood searching as searching for projects. The user comments that the first idea is to search for a project, and then search for a specific translation from there.