Am 14.09.2016 um 10:51 schrieb Léa Lacroix:
/- What else can provide wikidata to wiktionary?/ Machine-readable data will allow users to create new tools, useful for editors, based on the communities' needs. By helping the different communities (Wiktionaries and Wikidata) working together on the same project, we expect a growth of the number of people editing the lexicographical data, providing more review and a better quality of the data. Finally, when centralized and structured, the data will be easily reusable by third parties, other websites or applications... and give a better visibility of the volunteers' work.
Here are some examples of things that will become possible with the new structure:
* the fact that the English word "sleeper" may refer to a railway tie, and in which regions this is the case, only has to be entered once, not separately in each Wiktionary.
* the fact that "Stuhl" is the German translation of (a specific sense of) the English word "chair" only has to be entered once, not separately in each Wiktionary.
* by connecting lexeme-sense to concepts (items), it will become possible to automatically search for potential synonyms and translations to other languages.
* by providing a statement defining the morphological class of a lexeme, it becomes possible to automatically generate derived forms for display and search
* different representations (spellings, scripts) of a lexeme can be covered by a single entry, information about word senses does not have to be repeated.
* the search interface will know about languages and word types, so you can search specifically for "french verb dormir" (or perhaps more technical "lang:fr a:Q24905 dormir")
* Similarly, you can search for or filter by epoch, region, linguistic convention or methodology, etc.
- Will editing wiktionary change?
Yes, changes will happen, but we're working on making editing Wiktionary easier. Soon as we can provide some mockups, we will share them with you for collecting feedbacks.
The question is if you consider editing wikitext with complex nested templates "easy" or not. With wikidata, editing would be form-based, with input fields and suggestions. This makes it a lot easier especially for new editors. And even for experienced editors, I think it's more convenient for editing individual bits of information.
The form-based approach is less convenient when you want to enter a lot of information at once. The solution is to identify the use cases for this, and provide a specialized interface for that use case. This does not have to depend on Wikibase developers, it can also be done by wiki users using gadgets, Labs-based tools, or even bots.
Because Wikidata is a multilingual project, we already have to deal with the language issue, and we hope that with the increase of the numbers of editors coming from Wikidata and Wiktionaries, it will become easier to find people with at least one common language to communicate between the different projects.
Interestingly, we found that on wikidata there is rarely a conflict about whether a statement about an item should say X or Y, e.g. whether Chelsea Manning's gender should be given as "transgender female" or just "female" or even "male". The conflict does not arise because you can and should simply add all three, and use qualifiers and source references to specify who claimed which of these, and for which period of time.
Long discussions do take place about the overall organization of information on wikidata, about which properties to have and how to use them, about whether substances like "ethanol" should be considered subclasses or instance of classes like "alcohol".
I agree however that cross-lingual discussions are indeed an issue, and finding techniques and strategies to help with communication between the speakers of different languages will be a challenge. But isn't the Wiktionary community perfectly equipped for just that challenge? Isn't it just the crowd you would ask if you had to solve a problem like this? I would (along perhaps with the folks from translatewiki.net).