This #epic ticket captures the Minimally Viable Product (MVP) to validate the idea of using [MinT to support Wikipedia readers](https://www.mediawiki.org/wiki/MinT#MinT_for_wiki_readers) based on [recent research results](https://meta.wikimedia.org/wiki/Research:MinT_(Machine_in_Translation)_Research ).
The basic workflow involves the following steps:
- **Home.** The main starting point, from where the user will be guided to get a machine translated page.
- **Search for topic.** Supporting the user to search for a page to be automatically translated. Users can select the language in which they perform the search to find the page to be translated.
- **Confirm** Once a selection is made, users can make the final adjustment or checks before reading the machine translation. This is an alternative landing point for external tools that may select the page for the user, and an opportunity to surface local content on the selected topic.
- **Translation view.** An automatic translation of the page contents is shown to the user with guidance to navigate it (see link preview and translation options next)
- **Link preview.** Surfacing content from the linked articles which can be from the local wiki or also from a machine translation depending on the availability.
- **Translation options.** Additional options for the translation, including options to select languages, view contents, and contribute back.
{F42395571, size=full}
You can [ check this prototype](https://www.figma.com/proto/RTBmXJ4ZGmhPoNP1nlnm5C/MinT-for-Readers-MVP?page-id=3%3A379&type=design&node-id=3-380&viewport=452%2C400%2C0.37&t=HOyujJrCD1bpEO7j-8&scaling=scale-down&starting-point-node-id=3%3A380&hotspot-hints=0&hide-ui=1) or the video below for a better sense of how this workflow looks like:
{F42599000}
More detail on each step below:
____
# Home (T359401)
[View in Figma](https://www.figma.com/file/RTBmXJ4ZGmhPoNP1nlnm5C/MinT-for-Readers-MVP?type=design&node-id=195%3A5478&mode=design&t=SQ8jSuRFV6HTheHr-1)
A landing page for automatic translations. From there uses can start searching for a specific article to get a machine translation.
{F42466966, size=full}
More details at {T359401}
____
# Search for topic (T359494)
Facilitate finding articles to be translated from languages the user may not know. From multilingual search to searching on a specific language Wikipedia or using identifiers such as Wikidata-Ids or urls.
[View in Figma](https://www.figma.com/file/RTBmXJ4ZGmhPoNP1nlnm5C/MinT-for-Readers-MVP?type=design&node-id=207%3A6015&mode=design&t=2vlNHNWKlZGeNk8u-1)
|{F42467107, size=full}|{F42467109, size=full}|
More details at {T359494}
____
# Confirm (T359512)
Represents the selection for topic and languages to translate. Allowing final adjustments to the selected languages, including the exploration of the different language-specific versions of the contents, and surfacing the human created contents as an alternative when available. This serves as the landing step for external entry points.
[View in Figma](https://www.figma.com/file/RTBmXJ4ZGmhPoNP1nlnm5C/MinT-for-Readers-MVP?type=design&node-id=207%3A6345&mode=design&t=2vlNHNWKlZGeNk8u-1)
{F42467132, size=full}
More details at {T359512}
____
# Translation view (T359801)
The machine translated contents presented to the user. With indicators and options to identify the contents as machine generated and access the relevant translation-related options.
[View in Figma](https://www.figma.com/file/RTBmXJ4ZGmhPoNP1nlnm5C/MinT-for-Readers-MVP?type=design&node-id=207%3A6889&mode=design&t=2vlNHNWKlZGeNk8u-1)
|{F42543425, size=full}|{F42415217, size=full}|
More details at {T359801}
____
# Link preview (T359824)
[View in Figma](https://www.figma.com/file/RTBmXJ4ZGmhPoNP1nlnm5C/MinT-for-Readers-MVP?type=design&node-id=237%3A7396&mode=design&t=2vlNHNWKlZGeNk8u-1)
|{F42544901, size=full}|{F42544904, size=full}|
More details at {T359824}
____
# Translation options (T359829)
Options for users to adjust and navigate translation-related aspects.
[View in Figma](https://www.figma.com/file/RTBmXJ4ZGmhPoNP1nlnm5C/MinT-for-Readers-MVP?type=design&node-id=237%3A7864&mode=design&t=2vlNHNWKlZGeNk8u-1)
{F42467191, size=full}
More details at {T359829}
## Explore languages (T359863)
On a multilingual context such as Wikipedia, information about a topic can be very different from language to language. Users have a choice to select from which version to translate the contents. This step surfaces which contents can be accessed from each version to help a user explore the different contents available.
[View in Figma](https://www.figma.com/file/RTBmXJ4ZGmhPoNP1nlnm5C/MinT-for-Readers-MVP?type=design&node-id=237%3A8004&mode=design&t=2vlNHNWKlZGeNk8u-1)
{F42595938, size=full}
More details in {T359863}
____
# Entry points
## Access from language selector (T363183)
User use the language selector to switch across the languages they want to read content in. When they look for a language for which content is not available it may be useful to propose a machine translation as an alternative. This expands the current language support for the creation of new articles by translating with an additional option to access machine translation.
{F48594850, size=full}
More details in {T363183}
## Access from the footer of an article (T363338)
Even when an article exists in a given language, the content present may be very different than what is available in other languages. This entry point provides access to machine translation from the page footer. In this way, after users have read a short article, they can find a way to read more about the topic from other languages by using machine translation.
{F48447659, size=full}
More details in {T363338}
# Enablements
We plan to support a gradual deployment approach where we enable the tool gradually and get new feedback on each iteration to improve the tool. These are some relevant tickets in this space:
- {T363472}
- {T367852}
- {T363464}
- {T363465}
- {T367067}
- {T363470}
# Instrumentation
In order to check whether this MVP has the expected impact, user actions will be instrumented to analyze in which ways machine translation is accessed:
- {T363685}
- {T366149}