Abstract Wikipedia:About/tok
lipu Wikipesija pi toki ale li seme?
lipu Wikipesija pi toki ale li pali poka sin pi lipu Wikipesija.
lipu Wikipesija li wile ni: jan ale li ken jo e sona ale. lipu Wikipesija pi toki ale la jan pali li ken pali tawa lipu Wikipesija li ken weka e lipu pi sona ala.
nanpa pi lipu sona lon lipu Wikipesija en nanpa toki lon lipu Wikipesija li suli mute, taso ni: wan la toki ni li ken ala jo e sona ni. lipu Wikipesija pi toki Inli la nanpa pi lipu sona li 7 000 000; toki Lowasi la nanpa li 230 000; toki Sawili la nanpa li 100 000. sina ken ala toki e toki ni la sina ken ala jo e sona ni. tenpo poka la toki ante kepeken sona pi jan ala li kama pona, tan ni: toki ona li pona ala lon toki mute.
Abstract Wikipedia does that without relying on AI. Each step of the way remains under human control, and is accessible and editable by the volunteers. There are no hallucinations, no inexplicable and unfixable mistakes: If anything is wrong, fixing it is just one step away, and there is full transparency about how the text has been created. It is not just a probabilistic model choosing the capital of Kenya per chance every time. It is human-curated and owned knowledge.
lipu Wikipesija pi toki ale li pali tan seme?
Articles in Abstract Wikipedia are stored in a notation independent of human languages. The interface of Abstract Wikipedia allows volunteer contributors to create and maintain that language-independent notation. In a second step, this notation is being turned into text in human language. The steps required to make this happen are also created and maintained by a community of volunteers, on Wikifunctions. These steps can use the data on Wikidata, looking up population numbers, dates of birth, or irregular grammatical forms, which have been added to and are maintained on Wikidata by communities of volunteers.
See Help:How to create an article for a step-by-step guide.
Let's follow one example: Nigeria is the most populous country in Africa. In Abstract Wikipedia, this might be stored as:
Or, if we use English labels for these identifiers:
This is a call to a function on Wikifunctions. Based on the language, another function is being called. For English, we get the following function call:
And that results in the English text
If we had chosen German, another function would be called:
And that in turn would result in the following German text:
An example with more languages can be found here.
Abstract Wikipedia brings together numerous communities from the Wikimedia Movement and allows them to work together. Basic lexical data and structured data about many topics from Wikidata, functions that capture the necessary computations from Wikifunctions, and images and maps from Wikimedia Commons are all tied together in Abstract Wikipedia and made available for the more than 300 language editions of Wikipedia.
The goal of Abstract Wikipedia is to make more knowledge available to more people in more languages, and to allow volunteer editors to be more effective and work across languages, no matter what language they speak. Improvements to the knowledge will automatically propagate to the Wikipedias, errors will be spotted faster and fixed more effectively, and collaborations across languages become possible. We are tearing down the language barriers which are keeping knowledge apart.