Sunday, April 28, 2013

Spouse does not translate

I love #Wikidata, it gets so many things right. At the same time it fails by design. Consider; there is this article about a woman. In English she may have a spouse. In translation there is a problem; because the spouse has a spouse and it is a different word depending on the gender of the person involved.

This is the kind of problem that has been solved for the MediaWiki software. It means that you consider the sex of the person involved. Applying this principle on Wikidata is possible because we do register if a person is male or female.

It seems obvious; sex matters. In many languages you address people based on their sex. As long as Wikidata is not able to address this issue, it is broken. When this is intentional, it is broken by design.
Thanks,
      GerardM

Thursday, April 25, 2013

#OmegaWiki works and so can #Wikidata

OmegaWiki provides functionality that is on the agenda for Wikidata. The OmegaWiki community has ALWAYS wanted to be a Wikimedia project.

What it already provides is:
  • links to Wikipedia articles in other languages when the article does NOT exist in the preferred language
  • links to a commons category associated with a concept
  • a picture painting the thousand words associated with the concept
Yes, it is quite shocking; it also provides multi lingual dictionary support. OmegaWiki has even been described in a book on the development of lexicography.

Both OmegaWiki and Wikidata are "right in front of us" and, the functionality described above is imho "right for us". The challenge we face is to do something that is "shockingly rare"; meet halfway.

The road towards such a meeting could be:
  • Adopt OmegaWiki by the WMF in a labs kinda construction
  • Create the logical requirements of OmegaWiki data in Wikidata
  • Convert the OmegaWiki content to Wikidata technology
  • See how it fits in the big picture
  • Make it fit in the big picture
We do not need much talk and, the amount of work needed is relatively minor. The benefits are legion because even when all the OmegaWiki data is not used in the end, the lessons learned will be.
Thanks,
      GerardM

Machine translation for #Wikipedia

The Wikimedia Foundation suggests that machine translation is the kind of infrastructure that makes sense to it. Given what it aims to do: making knowledge available to everyone, this makes perfect sense. A lot of translation has already been going on in order to fill many gaps in the many Wikipedias and machine translations were often an important part of this.

One of the arguments why the WMF could enter the fray is that it has something to add. It does have monetary reserves but more importantly it has several resources that may make a difference. The biggest two are Wikipedia itself and the other are its awesome communities.

When translating a Wikipedia article, the concepts that are specific to a subject are likely to be found in that article. Similarly when such concepts have their own article, they will contain a similar set of concepts. Combine this with a multi-lingual dictionary build with Wikidata technology along OmegaWiki lines and it will be relatively easy to find the corresponding expressions in articles in different languages on the same subject.

The point here is that the meanings of words do not exist in a vacuum.

When such concepts have been identified and linked to Wikipedia articles and dictionary meanings it becomes possible to help people understand a text in a different language by providing native language support.

<GRIN> I know Erik Moeller has a lot of experience in this field and I know mutual friends are quite interested to help </GRIN>

Relevant is that we do not have to invent something new; it has been part and parcel of things we have done before. The difference is that we gained in experience and, technology has evolved as well.
Thanks,
     GerardM

Sunday, April 14, 2013

#Wikivoyage #statistics revisited


I am happy to report that many of the issues with several statistics have been resolved. The most visible improvement is that the page views for Wikivoyage are now available and updated on a daily basis. When you check out these statistics you notice that Wikivoyage got the most attention when it was started. Now in the fourth month we can notice some stabilisation.

The latest Wikivoyage projects in Hebrew and Ukrainian are now included as well and it will be interesting to follow how they will do in the future.

I am really happy that many issues have been resolved. If anything, statistics are trusted because of the consistency and quality of the presentation. When you follow developments regularly you get a feeling for the underlying data. Waiting for the resolution of what seems like cosmetic issues destroys that feeling.
Thanks,
     GerardM


Wednesday, April 10, 2013

Sign languages are important everywhere

When in 2005 the Austrian Sign language was constitutionally acknowledged, there was a curious second sentence added: "The Austrian Sign language is acknowledged. All details are determined by law". This sentence was a puzzle from the begin, because either it was trivial (then now one would add it because it should be found after every single point of the constitution) or there was some hidden intent behind it.

Now there is proof for the hidden intent by letters coming from "horrible jurists" of two ministries: They tell us that since there is no respective law, Austrian Sign Language cannot be established as a mother tongue in schools for deaf pupils. The jurists use the tardiness of the education ministry (this would have been obliged to develop realization laws for deaf people) as an argument that the Austrian government now dismisses the language rights of deaf people in the education process.

I read this news and, I am sure that these jurists have no clue why it is so important for children to learn to read and write in their mother tongue.. It is beneficial for all of their academic career. This is known to be true in the USA, Tunisia and Saudi Arabia ... it is also true for Austrians.

Thanks,
      Gerard