Jump to content

Translation UX/Design feedback 4

From mediawiki.org

The goal of this round was to test:

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

Test Observations

[edit]

Test with User #55 on 7 Jan

[edit]

The user is a Hebrew speaker.

Test 1: Homepage

[edit]
  • 00:55: The first element to call the user attention is the sign-up form.
  • 01:00: The user understands how to disable the default selection for languages and how to add more.
  • 02:10: When the user is asked to correct a translation, the search bar is immediately identified for the purpose.
  • 02:50: The user finds it clear how to become a translator.
  • 03:28: When asked about the list of projects, the user comment they look like placeholders.
  • 03:40: The user discovers the "translate" and "proofread" actions but not immediately, and finds the progress bars are "helpful information".
  • 04:20: The user understands the statistics in the tiles.
  • 05:00: The user comments that she is familiar with translatewiki but not an advanced user.

Test 2: Translation workflow

[edit]
  • 07:30: The user has difficulties to identify where can find material for translation.
  • 08:26: The user access the language statistics page to find projects to be translated.
  • 11:04: The filters provided for the message table are understood (even the change of fuzzy for "updated")
  • 11:38: The loading indicator placeholder made the user think that she had to wait.
  • 12:32: The user finds a suggestion immediately and considers it to be a "perfect match" for the text to translate.
    • The layout seems clear and easy to process by the user.
    • The outdated warning was displayed by mistake.
  • 12:42: The user cannot save which creates some confusion about the workflow: "Did I saved? It's not clear"...
  • 14:47: ..and asks whether clicking twice is needed each time.
  • 15:40: The user comments that the layout is clear in general. The user comments that initially though the text to be translated was the message above. This is probably an issue due to the combination of texts in different directions (the user makes use of a Hebrew UI and the text to translate is in English) so the message to translate is not aligned with the translation.
  • 16:10: The expand button is not clear for the user. Once it was tried it was misinterpreted as "previous".
  • 16:43: The user has a doubt for the translation of a specific word and comments that concordance search is needed to understand how to translate this word in a specific context.
    • Access to glossary and translation for specific words/fragments is suggested by the user as a possible solution.
  • 19:22: The user wonders why suggestions with a higher percentage match apear below others with a worse percentage.
  • 19:21: Since the "outdated" warning is shown for all messages by mistake, the user wonders why the "show differences" option.
  • 20:08: The user skips a message because of lack of context information for the message.
  • 20:38: The user saves and goes to the next message, which is the expected behavior for the user.
  • 21:16: The user finds a word for which she was not sure how to translate, checks a dictionary and updates the former translation.
  • 25:32: The user finds the UI "pretty similar to the existing interface".
  • 26:10: When asked about the possibility of asking a question, the user does not consider that as a frequent option. Nevertheless, the user is able to access to the support forum from the editor.
  • 28:50: The user is able to identify which project she is translating and is able to navigate to a different project.
  • 30:00: The user comments that normally translates some specific project because in this way you can get familiar with the terminology.
  • 32:00: The user is able to explore the hierarchy of message groups.
  • 33:55: The user asks about the meaning of the blue line to confirm that it means "translated" since she found a group of messages she did not expected to be completely translated to Hebrew.
  • 34:54: When the user is asked about the "load all messages", the user understands the behavior.
  • 36:04: The user comments that she was aware that the project list was searchable but she did not used
  • 36:50: the user comments that the UI is "cleaner and neat" but was expecting some additional functionality but functionality-wise is quite similar to the current UI.
    • Considers "concordance search" a must.

Test with user 56 on 9 Jan 2013

[edit]

The user was only able to provide voice comments in Russian and some comments in english by text (so time indications with the video recording are lacking):

Homepage

[edit]
  • "The main page is beautiful, but not very useful to me, because I have all the relevant links bookmarked."
  • "It's not clear what is the search box on the main page for. What is searched there? There are no useful results in Cyrillic, and a lot of results in Latin, but I don't know what they are."

Translation workflow

[edit]
  • The user expected the editor to become closed when completed the translation for the last message in the list.
  • "It's unclear what to click in the group selector box."
  • "The breadcrumbs at the top are very good." (The user called it "history".)
  • "The separation between the rows in the translation editor is not clear. I didn't understand at first that these are separate messages, I thought that these two rows are one long message."
  • "The meaning of the colors in the completion status bars is not so clear."
  • "Documentation and suggestions on the side box are not comfortable. Long messages won't fit, especially in translatable pages."
  • The user expected optional messages not to appear by default. The user commented that they appear always despite the checkbox status (which indicates that was able to find the option under the "..." filter).
  • The user considered the project selector "cool" but would like to see more statistics about the translation status.
  • The user commented that he usually works with a view that combines untranslated and messages requiring update (that is whet "untranslated" filter provides)
  • The user would be interested in a better support for search in the message list: filter a list for those elements matching a string, and being able to translate (and save) the results of a global search.

Test with User #67 on 13 Jan

[edit]

Homepage

[edit]
  • 1:00: The sign-up form seems to be properly understood.
  • 08:30: The user notes that the language ("Afrikaans") is not the correct one.
  • 11:00: The user uses the "more projects" option to see more other projects in the platform.

Translate workflow

[edit]
  • 12:12: The user uses the project selector to navigate.
  • 13:43: The user starts to translate.
  • 14:06: A message description is shown at the "help area"
  • 14:38: The user skips to next by opening the row below the current one (not using skip button).
  • 15:53: The user does not always follow a sequential workflow.
  • 21:06: "I'm impressed by the options the tool is giving to choose. Actually it is as the current version of translatewiki but there is more comfortable. And it was clear for communicating".
  • 24:00: When asked, the user seemed to understand the way "Skip to next" works. The user was not aware of skipping so it seems not a problem in the labeling.
  • 31:26: The user tries to open the selector by clicking on the current project which leads to an empty selector.
  • 35:15: The user is not able to find the "Visual Editor" message group.
  • 36:26: After clarifying VE is a MW extension, the user is able to navigate through the Project selector to find the project.
  • 39:12: The user copies source text and copies it to the text area to replace translatable text leaving wikitext unaffected.
  • 41:01: A warning appears in the help area in red text, indicating that the namespace name should be kept.
  • 41:40: The user copies then namespace to be kept the same and pasts it to the translation area.
  • 42:00: The user makes use of a suggestion and a warning appears due to the lack of a parameter $1 in the translation. The user modifies the translation to include it (by copying and pasting it).
  • 42:40: The user makes use of copy and paste to add a parameter to the translation text.
  • 49:00: The user confirms he didn't noticed the search at the project selector.
  • 51:30: The user comments that the warnings are noticed and helpful.

Test with user #R1

[edit]

A Bengali speaker.

Translate workflow

[edit]
  • 00:41: The user finds his native language immediately under the common languages region.
  • 01:25: The user is able to navigate using the Project Selector through "Translatable pages > How to start"
  • 02:28: The user has doubts about how to translate "Translate:How to start", in order to distinguish is which part should be kept/translated.
  • 03:32: The user asks for support on copying non-translatable parts or the message source to make it easier to translate.
  • 04:20: It is not clear to the user when magic words end and translatable text starts when text such as "{ { sitenam }}!]]" appears
  • 05:29: The user does not notice that she has no permissions for translating. Asks for the "Save" button but does notice the indicator of the current status. Suggests that making the text area disabled would better communicate the current state.
  • 22:42: The user selects sub-groups using the project selector.
  • 22:52: The user starts translating.
  • 25:00: The user finds suggestions and discovers how to use them after some seconds.
  • 25:50: The user tries to use keyboard shortcuts but they don't work because they conflict with the (iBus) input method system used at the system level.
  • 28:00: The user finds strange one of the aids provided which says "Click on the link 'expand' to view messages that contain the same text" but no 'expand' link is found".
  • 29:30: The user assumes that the previous message indicates a way to access the context of each given suggestion.
  • 32:35: The user comments that the list of messages is clear because status is indicated by text, but that may cause problems when messages grow.
  • 33:55: The user considers that the two-column layout is beneficial for short messages since original message and the textarea are next to each other.
  • 35:40: The user moves to the "How to start" translatable page.
  • 36:16: The user discovers the warning system and finds it neat. Nevertheless, early feedback by means of color could help to distinguish which parts of the message are "magic words".
  • 38:40: For long messages, translators usually copy the whole original text to be used as a template. In this way they can keep the non-translatable parts verbatim. The user considers that copying into the message box is an important missing feature.
  • 39:57: When the user is asked about the "expand" icon, the user finds it helpful but was not immediate to discover probably because it was too close to the "x" button.
  • 41:30: The group selector initial state is considered confusing sinc eit shows an empty page.
  • 42:40: The user considers the selector "neat" but the user would prefer to see sub-groups in the same list or comparison purposes.

43:00: It is confusing for the user how to go up. 44:41: The user is able to select all sub-groups using the seletor, but once selected, the user would find helpful to determine when each group ends/starts. 45:33: The user considers that continuous loading of messages may be a problem for people with low network speed.

  • 47:00: After trying long messages, the user still considers the layout also to work in this context. The user considers that when long suggestions are provided, the layout may be more problematic.
  • 48:00: The user feels that there is a need to obtain the context of suggestions.
  • 50:15: The filters for the table was not noticed by the user.

Homepage

[edit]
  • 11:45: The user asks whether the "blue boxes" are personal statistics. It seems not obvious which statistics are for the user and which ones are general.
  • 13:10: It was not clear how to get statistics for each languages.The user things that a way to access more detailed figures about their statistics.
  • 14:32: The search bar is not clear in which search space the search will be produced.
  • 16:16: The user asks about the possibility of sending notifications.
  • 20:28: The user identifies the way to search for a translation but would like to have more fine control on the search space used for the search.