Showing posts with label Language support. Show all posts
Showing posts with label Language support. Show all posts

Friday, July 02, 2021

Calling a spade a spade and the "friendly space policy"

In my previous blog post I established that facts supported by science trump personal opinions, even community consensus. It follows that pointing out a biased opinion can be hugely unpopular and considered offensive. The "friendly space policy" is applied on the spot and at best can be later appealed at the "trust and safety". This is after the fact, it takes a huge time effort and it is why I did not bother. One other reason is that I told the respected Wikipedian that what he said was biased and I did not make excuses for saying so after prodding by a trusted officer of the Wikimedia Foundation. I am sanctioned and can no longer use certain functions.

I run for a seat on the board of the Wikimedia Foundation and even though it is a popularity contest, I am not in it to be Mr Wonderful; I will not kiss babies. My platform is to improve our service particularly to the projects other than English Wikipedia. To be successful, I have established that this is doable and I have to undermine accepted opinions that prevent us from improving on our service.

I truly respect the Wikipedian who made the biased remarks but I will not apologise for pointing his bias out. 

When I am to establish my aims at board level, there must be acceptance that we do not serve everyone with the sum of all the knowledge we have when there is this persistent bias towards our English speaking public. The recent developments at the Croatian Wikipedia show that "community consensus" can be overridden when this is necessary to establish our accepted global Wikimedia policies. It follows that the board should be mindful of the evolving science about the "gender gap" and insist English Wikipedia to consider its recommendations (aka clean up its act). It follows that even the most respected English Wikipedian and in the light of the "friendly space policy" can be called out for such bias.

As to the organisation of the WMF; its director reports to the board and is to inform about the development for other languages and the effects on the traffic in other languages. 

Thanks, GerardM

Thursday, June 04, 2020

@Wikimedia and languages - @WikiCommons search, the most relevant development since @Wikidata

The Wikimedia Foundation is important for the support of languages on the Internet. The localisation of its software is done at translatewiki.net, it is done in over 300 languages.

The milestones for multilingual support are:
These milestones have been very much technology driven. For me the one reason why Wikidata became the success it is, is because it was from the start linked to every subject covered by Wikipedia and the solution was so overwhelmingly superior that nobody could reasonably object.

To make a success of this latest milestone, institutional support is needed. It is for the Wikimedia Foundation, its movement to reduce its bias for English and make room for improved language support.

My way of phrasing this as an essential objective: "All of is available to every single person on the planet". As we adopt this as our objective, it is first and foremost about making Special:MediaSearch useful in any and all of our languages and make it available from any and all of our Wikipedias.

As we adopt this, it is essential that priority is given to multilingual search over special interests including GLAM, Open Data, SPARQL and what have you. Priority when we are to open up in multiple languages first. Special interest only gain relevance when it is made obvious how it helps it helps open up Commons in Swahili, Hindi, German or Vietnamese.

Special:MediaSearch is possible because of everything that went before.. Its functionality is part of MediaWiki and localised at translatewiki.net. The existing search engine is now linked to the labels for items in Wikidata and it was made public after Hay Kranen brought us his proof of concept. It became available warts and all and while finding منصور اعجاز  in Punjabi is huge, it is not great when you do not find cats because a user is called Kočka..

The challenge to us as an organisation, a movement are we willing to work on our existing bias, open up Commons in all the languages we are said to support and accept that our hobby horses will get attention not in the next but in a future iteration.
Thanks,
       GerardM

Saturday, September 07, 2019

Language barriers to @Wikidata

Wikidata is intended to serve all the languages of all the Wikipedias for starters. It does in one very important way; all the interwiki links or the links between articles on the same subject are maintained in Wikidata.

For most other purposes Wikidata serves the "big" languages best, particularly English. This is awkward because particularly people reading other languages stand to gain most from Wikidata. The question is: how do we chip away on this language barrier.

Giving Wikidata data an application is the best way to entice people to give Wikidata a second look.. Here are two:
  • Commons is being wikidatified and it now supports a "depicts" statement. As more labels become available in a language, finding pictures in "your" language becomes easy and obvious. It just needs an application
  • Many subjects are likely to be of interest in a language. Why not have projects like the Africa project with information about Africa shared and updated by the Listeria bot? Add labels and it becomes easier to use, link to Reasonator for understanding and add articles for a Wikipedia to gain content.
Key is the application of our data. Wikidata includes a lot, the objective is to find the labels and we will when the results are immediately applicable. It will also help when we consider the marketing opportunities that help foster our goals.

Thanks,
      GerardM

Sunday, July 22, 2018

#AfricaGap - Sean Jacobs at #Wikimania

To be blunt, what Mr Jacobs is talking about is one or more step removed of the Wikimedia reality. His story is important and indicates that a specific type of source exists and is available for study. Mr Jacobs informs on the importance of Twitter for the Zulu language.

Mr Jacobs is an academic and the reality of Zulu Wikipedia is that only a few days ago we celebrated article number 1000 for the Zulu Wikipedia. What the Zulu Wikipedia needs is high school students writing in Zulu. Writing about what is important to them, what is important to their curriculum and to their world.

Just consider what one high school could do. Now consider what ten high schools could do. Compare that with one academic or what all the Zulu students currently in university could do.

Yes, history has been written so far and it does report in a biased way. When the Zulu language is to gain a foothold in the Wikimedia world, we need many people being involved in writing first the most basic information. Once there is a basis, the sources Mr Jacobs mentions become relevant in a Zulu Wikipedia.
Thanks,
      GerardM

Saturday, July 21, 2018

#AfricaGap - A #Wikidata based watch list about a Africa reality II

When there are many Listeria lists that you follow, when you care about the development about the subject, it is wonderful to see so much activity related to Africa.  As more people care to work on African politicians or "administrative territorial entities", the Listeria lists that also exist on Wikipedias in African languages will be updated as well.

When the Listeria lists become part of the main body of a Wikipedia, the politicians and entities will be found. When the info boxes as presented at the Celtic knot conference follow, slowly but surely quality content in quantity about Africa will no longer be a mirage.
Thanks,
      GerardM

Wednesday, July 18, 2018

#AfricaGap - Guinea; standing on the shoulders of giants

This map comes courtesy of the UN to Commons. It was downloaded in 2007 by Jeroen, the language on the map is French and Wikidata has much of its data in English. The names in French are mostly the same but that is for someone else to consider.

Many of the articles on "administrative territorial entities" are written by a small group of people. I want to single out Shevon Silva, the user page expresses the amount of work that went into adding stubs for so many African territories. The important thing about data is; once it is there you can change it in any way necessary.

When data gets entered into Wikidata, certain Wikipedia things are not possible; a "human settlement" is not a "administrative territorial entity". Such conflations need to be undone in Wikidata. Obviously the human settlement is located only in that administrative territorial entity and others only by inference. Attributes like "inception date" and links to other human settlements that are part of a sub-perfecture are for someone else to add/get right. Another consideration are historic administrative territorial entities particularly those of historic countries.

At this time it is important to celebrate what we have, morph it into a format that can be used on any and all of our projects. Once it is available in all the Wikipedias, it will generate more and more links and this will put Africa on the map.
Thanks,
      GerardM


Saturday, July 14, 2018

#AfricaGap - Support for "minority" languages

Support for "minority" languages was the subject of the Celtic knot conference. I have watched some of the presentations and find that there is a lot more to supporting minority languages from a Wikidata point of view than just adding missing labels. A vital strength of any Wikipedia is found in its relations between articles and that subjects of interest may be found.

Minority languages are a misnomer, what we mean is that the Wikipedias are small. They have a lack of articles, stucture is missing and subjects of interest are not found. Subjects have the same relations in any language and consequently lists expressing these relations can be shared using Wikidata in any language including "minority" languages. Missing labels need not be an issue; this is expressed nicely in this list of subdivisions of Egypt; the labels for most of them are only available in the Arabic script. A nice invite for people to add labels in the Latin and any other script.

The Welsh Wikipedia makes use of "Listeria" list in its main space and as a consequence, all items in these list can be found. They are available in a context, associated information may be available and they link to articles in other languages. The Welsh Wikipedia did implement the "Article Placeholder" and in this way they provide even more information for the ffspecific subjects.

When you consider Africa and information about Africa, there is no Wikipedia that provides adequate information. The data is incomplete, unstructured and often out of date. It is easy enough to improve on the quality of the data in Wikidata and when the information is updated in many Listeria lists on many Wikipedias, the impact is great.

The lack of coverage of subjects about Africa is huge. Less than 1% of humans is from Africa, we do not have up to date information about "administrative territorial entities" like provinces and districts. In my AfricaGap project only a limited range of subjects get some attention at this time. Obviously there is more that could be done. African cinema is one subject that is of interest to a group of Wikimedians. When they write their articles it will eventually translate to Wikidata and information about movies, actors and directors may be shown in Listeria lists in all the African language Wikipedias. This may generate interest from an African public for our projects.

There is only one purpose for Wikidata, Wikipedia and it is to find a public, a use case for the data, the articles, the information we provide. The one challenge we face is in both the quantity and quality of our articles and data.
Thanks,
     GerardM

Thursday, July 12, 2018

#AfricaGap - Considerations on the "Article Placeholder"

Having listened to a Youtube presentation on Article Placeholder, I am seriously disappointed. There are a few statements in there that show a lack of understanding on the functionality of the Reasonator. It is dismissed for all the wrong reasons and as a result there are a lot of missed opportunities.

What is missed is that Reasonator, as it is, provides superior representation in any language. It is a tool that helps with missing labels from within the tool. Missing descriptions in Reasonator do not need to be a problem; there are automated functionality that has shown its merits in many languages. Do compare the representation of Wikidata data and the structured representation will be seen to be more rich with the inclusion of maps, images and data linked to the subject in question.

What is particularly galling is that Reasonator is dismissed because "it is an external tool". Before work on the Article Placeholder started, it would have been easy enough to adopt functionality as provided by this external tool and it would not have been an external tool, an obvious argument AFTER the fact.

Where Reasonator provides texts, it is done based on little scripts. This is seen as problematic as is seen as a drain on the community. Templates on the other hand may be a part of the Article Placeholder and they have the same problem.

For me the bottom line is not so much about the Article Placeholder but the lack of usability of Wikidata. It is only because of Reasonator that it is easy and obvious to work on the subjects I work on. I have not spend hours learning how to query, Reasonator provides me instantly with the results in any context like the missing "Districts of Djibouti".
Thanks,
        GerardM


Thursday, July 07, 2016

#Wikimedia chapters

As a movement, much of the local effort is channeled through chapters. A lot of important work is done because they provide continuance to the work that we do. It enables us to foster relations and organise the more complicated activities.

The map shows beautifully where national chapters exist. In the USA chapters do exist but the map does not acknowledge this.

Typically I am content with the information that is shown through the Reasonator, it however has its limits. This is where the lists maintained through ListeriaBot become relevant. This list of chapters shows additional columns like "country" and start "date". It enables sorting and this is relevant functionality. The same list may exist on a Wiki supporting a different language, for instance Dutch. The cool thing is that the same bot will at some stage update all these "similar" lists. It is just a matter of completing the data for better use.
Thanks,
      GerardM

Sunday, July 03, 2016

#Wikimania - James Heilman on #quality and #language support




James Heilman presented at Wikimania 2016. I have not been to Wikimania and even the people who went to Wikimania may not have listened in. James had a lot to say about quality and the power of having information in "other languages". The talk is powerful and the arguments are compelling. At this time only 48 views of the talk on YouTube,
Thanks,
     GerardM

Monday, June 20, 2016

#Wikidata - #Pakistan Peoples Party politicians

There has been an announcement that lists may be generated on a Wikipedia using Wikidata data. For the Urdu Wikipedia, a list of politicians of the Pakistan Peoples Party could be interesting. This functionality is not available yet, but Reasonator does show us what would be on such a list.

As you can see many of them do not yet have an article in Urdu or alternatively a label. Once a label has been added, it will show up in the list. This may also help other languages from Pakistan like Sidhi because the label will fall back to Urdu and not English.
Thanks,
      GerardM

Friday, April 08, 2016

#Wikidata - labelling and the ArticlePlaceholder

The ArticlePlaceholder is a Wikimedia extension. It has huge potential to rapidly increase the available information in any and all Wikipedias because Wikidata has more data than any Wikipedia has articles. The big initial issue: How to translate the labels into the language of a Wikipedia.

There are many approaches and the initial one is to concentrate on "red links" first. With "red links" linked to Wikidata items, the ArticlePlaceholder may express the data as information for that Wikipedia. That is one big incentive to write the missing article :). Missing labels associated with properties may be added to localise all the information available on an item

It then becomes interesting; What to do next? There are the categories, the lists associated with articles. It would work, it would rapidly expand the number of items linked through ArticlePlaceholders. It would also rather quickly expand the number of items that seek a translation of their label. These can be found and sorted in order of prevalence.

Another approach would be to add labels as many labels as possible in Wikidata. When this results in terms that can be found, ArticlePlaceholders may be created when they are requested. When we keep track of the requests for ArticlePlaceholders, we can even prioritise the writing of missing articles.

Adding labels can be done in multiple ways; we can use dictionaries, we can transliterate. We can even use bots to do all this automagically for us. The biggest difference will be made when we seek and find collaboration. For most languages there are schools where students need to write papers. Particularly in the smaller projects this may make a huge difference. As they research their topic, they can localise all the missing labels starting from a "Concept cloud" for a subject.

Once this work gets underway, small Wikipedias will rapidly increase the number of subjects that are being served. The big question is not if it will be worthwhile but how it will affect Wikipedia article writing.
Thanks,
       GerardM

Friday, August 28, 2015

#Wikidata - Joseph Reagle not an #author?

The English Wikipedia article says: "Joseph Michael Reagle Jr. is an American academic and author focused on technology and Wikipedia". It seems obvious that the occupation "author" fits Mr Reagle.

Not so I am told, the word "auteur" is a generic term in French, so it is at best an anglicism. This gets us in a tricky position because it is suggested that if this appears in infoboxes which automatically import stuff from wikidata, it will create an absolute mess in the French wikipedia, with everybody being credited as an "auteur" which does not make sense at all.

When you analyse "author" in Wikidata, it is a subclass of "creator". Creator seems to me to be what the French understand for "auteur". Consequently, the labels used in French do not match what is meant by author in English.

Arguably, when items are labelled in a way where the meaning in one language is not the same as in other languages,  This has major consequences for the integrity of Wikidata.
Thanks,
     GerardM

NB Mr Reagle wrote a few books, that makes him more than an "essayist".

Monday, July 13, 2015

#Wikidata - more great statistics

When you want to know what Wikidata data is about, you find it in the statements "instance of" and "subclass of". The illustration is part of a set of new statistics by Zolo. It concentrates on these instances and statistics and a big surprise is that there are still so many more items without one (32.6%). The statistics by Magnus show that only 19.07% of all items have no statement.

The great thing of the new statistics page is that it shows pie charts for many Wikipedias. English Wikipedia for instance for instance does not know about 1.5 million people Wikidata knows about. When you analyse this, you get into the domain of set theory. These people can already be subdivided in the following way:
  • people only en.wp knows about
  • people en.wp knows about as well as others
  • people en.wp does not know about but another wp does
  • people no Wikipedia knows about
  • items Wikidata does not recognise as people
People and unidentified items are half the pie for Wikidata. In all cases they are pretty well connected to articles in whatever project and provide a perfect opportunity to provide information. I blogged about it before, wikilinks and red links are obvious targets for links to other information not only for people but also for people.

Many people have #Babel information on their user page. They thereby indicate their proficiency in a language. When a Wikipedia does not have an article, why not provide an article in a known language? When there is no article why not provide at least the information that is available? It will help an editor write a new article, it will provide a reader with at least some information.

The point is that we should share in the sum of all available knowledge because we can.
Thanks,
     GerardM



Monday, June 22, 2015

#Wikipedia - business as usual is not an option

"
The New York Times featured an article titled: "Can Wikipedia Survive?" by Andrew Lih. It is a good article in that it describes the current state of Wikipedia. It raises several important points and the main one is that because of smartphones and tablets many people do not contribute as much as they used to.

Articles like that describe the status quo. The issues that seem important based on the old understanding of what Wikipedia is about. The older understanding was: "To share the sum of all knowledge" the current situation is that there are 280+ Wikipedias and each is on its own to do its own thing using the MediaWiki software. There are other projects that contribute to the old motto but they were outside of the scope of the article.

When you reflect on the original objective, all Wikipedias fail. Every Wikipedia has its own content and is distinct in what it has to say. Consequently they do not share the sum of all knowledge; they are not even aware of the knowledge that is available elsewhere.

When you analyse Wikipedia, it has several components; there are accumulations of text and there are accumulations of data. There have been experiments that show clearly that it is not always necessary for a person to write the text. The experience from several Wikipedias is that bot generated content leads to more readers and more editors. This is quite counter intuitive but hey, why dispute the facts? If there is one draw back, it is with updating said texts when need be.

We know that once enough data is available for a subject information may be gleaned from raw data. Wikidata provides raw data and Reasonator among others transforms it into information. This may finally be accepted when the overly long awaited Wikidata Query engine will become available.

What this may do is several things.
  • people will want to add items and statements to Wikidata
  • results will pop up everywhere once the facts are in
  • more software will be written to produce texts based on Wikidata data
  • articles generated in this way may be cached without saving the text
  • generated articles will change once more facts are known
This is not rocket science. It has been done before. The only question is does sufficient motivation to accept changes to Wikipedia exist. As Andrew Lih so eloquently asks "Can Wikipedia Surive?" the answer seems obvious. Wikipedia has to change in order to survive. This approach will help us improve content in any language.
Thanks,
     GerardM

Friday, April 24, 2015

#Wikidata -#UN Prize in the Field of Human Rights

The United Nations recognises organisations and people who have made a difference. The United Nations Prize in the Field of Human Rights shows many people and organisations that were awarded in this way.

At this time this list is a "work in progress" at Wikidata. The people known to the English Wikipedia are on it. It is now a matter of filling in the "red links" and removing a few erroneous entries. The only way to see this list is by using the Reasonator.

What is really interesting is that most of the missing people already exist on Wikidata. They just do not exist on the English Wikipedia, Adam Lopatka is one example. Completing a list like this is therefore an exercise of patiently checking every entry for existing items. For others, like the Vicaría de la Solidaridad the link has not been made.

The beauty of working on a special interest like human rights in Wikidata is, that you can highlight the people that matter. This does not mean that they are easily found. Amir is working on a bot that will do transliteration. That will be a boon particularly to find the people who are only known in a different script. Mehranguiz Manoutchehrian and Petr Emelyanovich Nedbailo are two examples that may have an article after all. The first is from Iran, the second from the USSR.

It is wonderful to see how the pieces are coming together for Wikidata.
Thanks,
     GerardM

Thursday, March 12, 2015

Auto-#translations - let's talk "man and horse"

Niklas wrote in his blog about using a grammatical framework for generating texts. He also mentions Reasonator and indicates that its support for language is limited.

When you look at the code for generated texts, it is indeed a programmers job and not so much something a translator does. What it does is parses existing data and generates strings of text. The generated text for Marilyn Monroe for instance has it that she is an "actor".

"Lets talk man and horse" is a Dutch proverb and it is exactly what we will not have in generated texts. It is indeed grammar that we need to concentrate on. I am sure that many people will be EAGER to work on generated text using a grammatical framework.

Once it is shown to work in Reasonator, the next step is obvious; generating the texts for consumption in Wikipedias that want to share in the sum of available knowledge . This is best done by caching results because of the flexibility it offers. When this is seen as problematic; it is even easier to generate articles using bots.

Niklas, what category of topics tickles your fancy?
Thanks,
       GerardM

Tuesday, February 10, 2015

#Wikidata - Tokyo University of the Arts alumni

Mr Kenji Ekuan died. He was a Japanese designer who studied at the Tokyo University of the Arts. As can be expected, he was not the only one who studied there. There are categories in several Wikipedias informing us about them. The category on the Japanese Wikipedia has includes 642 alumni and for many of them there is no label in other languages.

It is no surprise either that these people refer to many items that have a label in Japanese and not in the languages people are familiar with. The automated description for Mr Munemoto Yanagi in Dutch for instance is "kunsthistoricus (*1917); Mainichi-Kulturpreis; kind van 柳宗悦 en 柳兼子 ♂". As more labels become available in Dutch, this information becomes easier to comprehend.

With every label that is added, all the associated descriptions are improved. Every item will be easier understood in Reasonator as well. Adding labels in Reasonator will provide you with instant gratification. Every statement of that item will show the new label.
Thanks,
     GerardM

Sunday, January 04, 2015

#Wikidata is more like a #ball and not so much like a #pie

The value of the data in Wikidata is in its connections. From the start there are the inter language links to Wikipedia and as more statements are added, they are adding value in their own way.

Many statements are repetitive; many people are born or die on the same day. Many people go to the same school or take part in the same event. Items like these can be seen as a cluster, a ball. All of them together are an even bigger ball.

One early choice in Wikidata was to link items to external sources. This has many potentials for the future. We did experiment with comparing data. We may represent external data and not store it in Wikidata.

Through these external links, Wikidata is very much in a galaxy of data. There are clusters here and there of similar data. The value of that data is very much in our ability to represent it. From a language point of view all this data is not diverse. It is therefore our challenge to bring more language people to Wikidata not only for them to add labels but to do this in a SMART way.
Thanks,
     GerardM

Sunday, December 07, 2014

#Wikipedia- The numbers are what ?

So the numbers are flat.. You want initiatives that help provide us with relevance.. Ok, how about this scenario:

A reader queries Wikipedia for a subject and does not find it. Many more people query for this subject and it becomes the most wanted subject without an article. An editor writes the article and it proves popular. It becomes the most read new article in the next month.

In this way:
  • we give our search statistics a purpose
  • we indicate what subjects our readers want articles about
  • we celebrate the most read new articles and their editors
  • we advertise that we ask our editors to write articles people are looking for
  • we can do this for every Wikipedia in every language
Yes, we can provide search results from Wikidata as well. When people make use of this, it counts as a not found instance. In the mean time we did provide information that is available to us.
Thanks,
      GerardM