This is an archive of past discussions from Wikinews:Water cooler/technical/archives/2013. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current page.
Notifications inform you of new activity that affects you -- and let you take quick action.
(This message is in English, please translate as needed)
Greetings!
Notifications will inform users about new activity that affects them on this wiki in a unified way: for example, this new tool will let you know when you have new talk page messages, edit reverts, mentions or links -- and is designed to augment (rather than replace) the watchlist. The Wikimedia Foundation's editor engagement team developed this tool (code-named 'Echo') earlier this year, to help users contribute more productively to MediaWiki projects.
We're now getting ready to bring Notifications to almost all other Wikimedia sites, and are aiming for a 22 October deployment, as outlined in this release plan. It is important that notifications is translated for all of the languages we serve.
There are three major points of translation needed to be either done or checked:
Notifications help on mediawiki.org. This page can be hosted after translation on mediawiki.org or we can localize it to this Wikipedia. You do not have to have an account to translate on mediawiki, but single-user login will create it for you there if you follow the link.
Comment from deliverer: the WMF can't tell the difference between Wikinews and Wikipedia apparently, but it doesn't need to be translated to English anyway. ;) PiRSquared17 (talk) 20:30, 4 October 2013 (UTC)[reply]
...but something is broken on the "make lead" page. This seems to happen on-again-off-again. I like to mess with the lead parameters (e.g. first sentence, first 2 sentences etc. etc.) but when I try to 'refresh' via the button....nothing refreshes. Gremlins in the gears?? --Bddpaux (talk) 02:05, 29 October 2013 (UTC)[reply]
I just pulled up WN:ML, picked a couple of the articles it offered, and the 'Refresh Preview' correctly updated the content.
That doesn't mean there's nothing wrong with it, just that it's probably a little subtle in its defectiveness.
First things to check:
Clear browser cache; a lot of css/js stuff has been tweaked recently, make sure you're using current copies.
When it's reproducible for you, note details (which article? Copy markup out the WN:NL box).
It could-well be mismatched, or otherwise mangled, markup; something in the article markup not being correctly read to build the offered lead; but, to pinpoint it will likely need some examples. --Brian McNeil / talk09:23, 29 October 2013 (UTC)[reply]
OK, it just happened when I was trying to set up the Congresswoman lead. I got this message that popped up in Chrome:
"A local JS function experienced an error on an API request. Please leave a note at User_talk:Bawolff.
Details: Error: Could not retrieve page "Undefined" as it does not exist. <undefined:undefined>
browser=Mozilla/5.0 (Windows NT 6.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/30.0.1599.101
Safari/537.36"
It usually, the moment WN:ML loads, throws the suggested lead, photo etc. etc. at the bottom of the page and I can usually play around with "use 1st sentence" or "use 1st 2 sentences" or whatever....but when I'd try that and hit "refresh", nothing would happen. Don't know if this helps you Java-gurus or not! Although technically, I never did any hotkey stuff to clear my browser's cache...so that could've been the issue.--Bddpaux (talk) 04:59, 9 November 2013 (UTC)[reply]
For the record, I had a similar issue yesterday when I tried to publish an article. It came up with a null title. And I had to manually put a title in after that. I lost the extra data around that to see how it was situated in the broader front page context. --LauraHale (talk) 06:53, 9 November 2013 (UTC)[reply]
The null-title thing (actually, it comes out as "undefined") has been happening to me consistently for the past several days, I think. --Pi zero (talk) 12:27, 9 November 2013 (UTC)[reply]
New Search Coming - Please Help Shape It
At some point in the new few months WMF will be replacing the backend that powers search on all of the wikis we host/power/serve/whatever verb is most appropriate. We have very good reasons to do this that I won't bore you with but will link in case you want reading material. Anyway the old search system had some kind of special boosting for recent articles just for wikinews. Do you still want that feature? Do you want it but with some slight modification? I'm asking here because this is the largest and most active wikinews by about an order of magniture. And I speak English. Thanks. NEverett (WMF) (talk) 20:56, 6 November 2013 (UTC)[reply]
Ooh. I didn't know that we had this. That's definitely useful. Even more useful would be if we could have faceted search, search ordering and having the ability to display metadata on the search page (a bit like Google's Snippets). Other news sites tend to have the ability to choose to order by relevance or by date. Thanks for letting us know. —Tom Morris (talk) 10:10, 7 November 2013 (UTC)[reply]
Regarding faceting: it is on my list of things I want to do once we've got everyone running happilly on the new system. Are you thinking of a date based facet? On article creation date or article modification date?NEverett (WMF) (talk) 15:07, 7 November 2013 (UTC)[reply]
Regarding choice of relevency vs date: can you link me to an instance of this that you particularly like? I can't quite wrap my head around how a purely date based search would work well. In my experience ignoring the "goodness" of the match and sorting on something else tends to make the results garbage.NEverett (WMF) (talk) 15:07, 7 November 2013 (UTC)[reply]
As I recall, we've wished it were possible to restrict a search to pages belonging, or not belonging, to a category. There's something almost like that in the query language, I believe, but it actually depends on the category being specified directly on the page rather than transcluded. --Pi zero (talk) 17:08, 7 November 2013 (UTC)[reply]
CirrusSearch get the "belongs to category" for transcluded categories "for free" because it indexes fully rendered pages rather the wikitext that makes up the pages and supports all the old query syntax for which I've been able to find documentation. I've filed a bug for negating the category filter. That sounds useful and simple enough to implement. NEverett (WMF) (talk) 18:02, 8 November 2013 (UTC)[reply]
Wikidata
I figure I should start an exploratory discussion here about using Wikidata for our category system.
Wikidata is already being used by Wikipedia, Wikimedia Commons and Wikivoyage. One way that we can use this is to keep our inter-project links from category pages up-to-date.
Consider a page on Wikinews like Category:United Kingdom. The template on that page contains links to sister sites including Wikipedia, Commons and Wikivoyage. Using Wikidata, we could have it so that inter-project links to those projects are kept up-to-date. We can also use our existing links between our categories and projects like Wikipedia to fill connections with Commons and Wikivoyage. Let's say we had a category called 'Category:X' and it linked to the Wikipedia article 'X'. Then let's say someone at Commons created a category for pictures of X, creating 'Category:X' on Commons. They could then edit Wikidata to update to link Commons to the existing Wikidata object. Then Wikinews' Category:X page gets automatically updated to have a link to Commons.
Consider also another use case for Wikidata. It contains property links between objects. For instance, on the category "United States", we might pull in the property from Wikidata of who the current President of the United States is. I'm hoping that when Obama's term is over, we'll be able to rustle up an article on whoever replaces him. But presumably whoever it is, we'll have articles about them already. Let's say that Hilary Clinton becomes the next President. There is a category for her (I think). If we've got all the Wikidata infrastructure in place, the 'United States' category could 'know' who the President is by querying Wikidata.
The downside here is we are delegating editorial control over our category system to Wikidata to some extent. But let's be real: we are here to write news articles, not necessarily to maintain a category tree. And we don't really have the time to maintain it (hell, we struggle to write the news every day). Delegating some metadata management to Wikidata is quite possibly a huge win for us. I hope Wikidata people can come and give useful advice too on what we might do with Wikidata. —Tom Morris (talk) 22:25, 14 November 2013 (UTC)[reply]
I'm glad that someone has given thought to how Wikidata could work on/with Wikinews. I do see linking doing interwiki links for Wikinews categores as something that would be useful. It's certainly something that will happen. However I've never really seen how it could work with the Wikinews mainspace. Yes, you can link the Wikinews article for an event to the Wikipiedia article on that event, but many Wikinews stories don't match up very well with pages in other projects, and truthfully I have no idea whether or not there are English Wikinews article and Spanish Wikinews articles on the same story (which is the other way interwiki links would be useful). If Wikinews is going to do interwiki links on Wikidata, a lot of thought needs to go into how that would work. Sven Manguard (talk) 22:40, 14 November 2013 (UTC)[reply]
Well, the entire idea of the interproject links via wikidata is mildly controversial. To use Tom Morris's example, there is not a 1-to-1 mapping of the concept of "United Kingdom" across the Wikipedias; there is Great Britain, the British Isles, Brittania, Pax Brittania, and on and on. But that's okay, we know the system is inexact and in any specific case it will always be somewhat 'wrong', yet it can still be useful for helping make sense of needlestacks.
I would be mildly interested in hearing how Tom would implement his forms, since it sounds like it would be adapting some Wikipedia syntax to use on en.WN. - Amgine | t22:56, 14 November 2013 (UTC)[reply]
It's long been my sense that Wikinews main space is very unlikely to be a useful subject for Wikidata. Using it for category space is possibly not as hopeless, but I'm aware it can be quite challenging, as once you get beyond "obvious" things like country categories (which are already somewhat tricky, as Amgine notes) the different Wikinewses are apt to use very different conceptual taxonomies. I recall my bemusement, some time back, when I realized we have categories for free speech and freedom of the press, while most other-language Wikinewses I looked at have instead a category for censorship. --Pi zero (talk) 23:13, 14 November 2013 (UTC)[reply]
Mainspace interwikis of synthesis articles between different-language Wikinewses are usually not, in my experience, mere translations of each other. Often they have slightly different foci. It can then become a quite subjective decision whether or not their content is close enough to warrant an interwiki. --Pi zero (talk) 23:46, 14 November 2013 (UTC)[reply]
Moving inter-wiki for mainspace to Wikidata isn't something I think could work. But using it for inter-project links and other metadata needs is something worth exploring, I think. —Tom Morris (talk) 00:10, 15 November 2013 (UTC)[reply]
I think categories would be a useful way to implement Wikidata use. Outside of that, I cannot see the use for article writing. I talked to Gerard from Wikinews. I explained how things work on Wikinews. It was a Wikipedia centric point of view, that had little thought in implementation for actual news writing. Wikidata does not have fields for references, and one of its inherent advantages is to live update information. This does not work for a news site where things need to be locked in time. It does not work for verifiability. Beyond that and at the end of the day, there is a very real concern I have that integrating this into news writing creates major barriers for participating. I would like to see some redesign from the Wikidata people for things like Wikinews:Writing an article where we can see a lower barrier to entry for writing articles and this complimenting this, for making information verifiable. Until we can overcome those hurdles, this is a fundamental non-starter. Gerard suggested that for the Sochi Paralympics, a Wikinewsie could manually enter the results from Sochi onto Wikidata, then they can create a table on Wikinews that pulls from that inputted data, then cite Wikidata for the verifiability for the Wikinews review process. As a Wikinewsie, this terrifies me because that sounds so complex, and having to write on Wikidata to write a Wikinews article? And then as a reviewer, how do you verify on Wikinews articles the Wikidata information? We'd need a special template that says "This article is verifiable based on original reporting data submitted on Wikidata." And yeah, that does not work. I have no idea how this is implemented. --LauraHale (talk) 13:45, 15 November 2013 (UTC)[reply]
Categories were supposed to be all that was under discussion here. If there is a downside to introducing use of Wikidata for categories, it probably lies in the extremely subtle difference between making a task easier to perform by empowering local users to do it, versus taking it away from local users so that someone else can do it more easily. And what makes it all really subtle is that the difference between the two doesn't show up when you're talking about our category space. However, if given the choice between making category interwikis easier to maintain by empowering local users, versus by handing over control to a separate sister, it would be healthier for Wikinews to empower local users, because that's the direction we (and all sisters) need to move in if wikimedia is to survive and prosper in the long term. This is, I fear, a subtle design flaw in the current concept of Wikidata, relative to the wikimedian movement as a whole: it gently pushes things in the opposite direction from where the long-term best interests of the wikimedian movement lie. Wikidata seeks to take certain functions away from the other sisters and isolate those functions in the Wikidata sister — which would be great if the functions to be taken away from other sisters were truly orthogonal to the things that make those other sisters different from each other. Unfortunately, afaics that is not the case atm: the functions being taken away seem to be chosen based on understanding only of Wikipedia, without deep understanding of what makes other sisters different. There is a potential for this approach to favor centralizing functionality that, for the long-term best interests of the wikimedian movement, should be distributed amongst the sisters. --Pi zero (talk) 15:06, 15 November 2013 (UTC)[reply]
There is a trade-off—and every other project including the Wikipedia projects has had to weigh that up. I think that the important bit of Wikinews is the writing the news bit, not necessarily inter-project links on category pages. —Tom Morris (talk) 20:06, 15 November 2013 (UTC)[reply]
Indeed. I'd put it that category space on en.wn is not a content space. (At present our only content space on en.wn is mainspace; I'm familiar with one example of a project with multiple content spaces — en.wb, where mainspace, Wikijunior space, and Cookbook space are all content spaces.) --Pi zero (talk) 20:42, 15 November 2013 (UTC)[reply]
The problem is until a clear case is explained, or something is demonstrated, it is hard to have an idea. What does having Wikidata mean? It compliments writing articles? I was told that we could basically write articles by generating data tables that automatically update pulling Wikidata information that live updates articles from Wikidata, and that this information would comply with all local policies. I strongly suspect lots and lots of bullshit. Maybe I am wrong, but I need to see how that happens. How is this done? Some one write an article pulling in Wikidata so we can see this works. I would support backend installation of Lua modules to make it possible to test, but a few examples are necessary to see if this is something we should be scaling up. --LauraHale (talk) 08:36, 16 November 2013 (UTC)[reply]
I'm sitting at Gerard's presentation about Wikidata. I asked about verifiability repeatedly in terms of Wikidata and its use for projects requiring verifiability. The answer was having information verifiable is a non-priority. They do not care about doing this. No one does it because most people are too lazy to do this. Given this issue, I would oppose on principle any inclusion of Wikidata usage in news articles where verifiability is required as part of the review process. --LauraHale (talk) 13:35, 16 November 2013 (UTC)[reply]
I point blanked asked Gerard about the Freedom of Speech versus Censorship issue and how that would be handled. The answer was these things are black and white conceptual linking with only one real way to do it. Thus, it is clear to me that a move to Wikidata for interwiki linking would involve local projects having to give up their freedom to conceptualize the world in their own language like this. The answer was yes, Wikidata would ultimately control that conceptualization. --LauraHale (talk) 14:29, 16 November 2013 (UTC)[reply]
Which makes it clear Wikidata has no potential, in its current form, for positive contribution to this project; it can do damage, but is not set up to help us. As I noted earlier, Wikidata represents a factorization of Wikipedia content; a flawed concept, from the perspective of the sisterhood as a whole. Serious consideration might be given to how the functionality of Wikidata could be refactored so that control is more distributed amongst the sisters. I'm distantly reminded of the idea, bandied about here from time to time, of some sort of (carefully limited) mechanism for transcluding a template from a different project. --Pi zero (talk) 15:06, 16 November 2013 (UTC)[reply]
That functionality (transcluding templates) exists in the software, some preliminary work has been done to centralize complex template systems on Commons. Afaiaa it will never be enabled due to the well-known risk of cross-site/cross-project vandalism. - Amgine | t15:11, 16 November 2013 (UTC)[reply]
Yeah, the tricky part is always how to design a feature with natural limits to avoid creating exploits (something I've put massive thought into for my interactive tools). I fundamentally approve of the native wiki template mechanism for this reason: it allows any user to create arbitrary templates, without requiring them to learn a separate programming language, while at the same time not giving them the ability to do unbounded iteration. That encourages distributed expertise, whereas the (technically useful) Lua extension encourages centralized expertise. --Pi zero (talk) 15:22, 16 November 2013 (UTC)[reply]
Hi Laura and everyone else. I'm the product manager at Wikimedia Germany for Wikidata. I want to clarify a few things here before you all get a very wrong impression of Wikidata. I don't know what Gerard said but let me assure you that verifyability and sources for information are utterly important for Wikidata. Yes we currently don't have enough sources yet but we are working on that. It is my highest priority for all Wikidata work. My hope is that Wikidata will actually be much harder to vandalize than any of the other sister projects. There are two reasons for that: 1) Changes on Wikidata show up in the sister projects. If you edit the data for Berlin on Wikidata for example it will show up on the watchlists and in recent changes on all the Wikipedias, Wikivoyages and Commons that have a page about Berlin. A lot more eyes will see this than currently. 2) Sources in Wikidata are structured data. This will make it a lot easier to write tools that verify that for example a source actually says what is stored in Wikidata.
The other thing I want to stress is that we are slowly rolling out to all sister projects. It will still take a while until Wikinews' turn. We will start only with the interwiki links first and only later enable data access. During all this you will have time to decide to which extend you will want to make use of the data in Wikidata. As people previously said it'll probably not make too much sense in the news articles you are writing. But what might make sense is giving readers a small box with information from Wikidata to put the news item into context. But again: this will be up for you all to decide.
And one more addition: Wikidata is built with the very intention to represent a very broad diversity of knowledge. It is possible to have a lot of differing opinions on Wikidata for the same thing. It is not about the one single truth but about collecting what different sources say is true in one place. Editors are free to chose the data they think is "true". --Lydia Pintscher (WMDE) (talk) 15:59, 16 November 2013 (UTC)[reply]
Thank you @Lydia Pintscher (WMDE): for the response. It does reassure me. Many Wikinews articles in other languages using sourcing similar to Wikinews, and like Wikipedia, have verifiability as a central pillar of the project. Wikidata appears to have currently no way to verify the information. Including Lua scripting into an article that automatically pulls that data without sources fails verifiability. Many Wikinews projects are also locked in time, so calling to updated Wikidata articles becomes a problem. These are major concerns. There is a very real issue that editors are being set up to fail by encouraging them to use a tool that will not pass publishing standards. The other issue is the complexity of Lua as a code adds another barrier to participation. Small projects like Wikinews, across all languages, have problems attracting new contributors. Asking them to have coding expertise in Lua is a huge hurdle. (Look how many Lua coders we have on project now who are clamoring to use Lua to write articles. None really.) VE is a step I am looking forward to because it makes fewer barriers for new contributors... this, not so much because articles are not needing continual updates.
On some level the problem is probably this when it comes to the category structure thing discussed above: Wikinews projects treat categories like article tags. The solution on some level is for us to adopt a Wikimedia extension to do away categories and portals altogether and replace it with a tagging system, and from that tagging system, build on Wikidata so we're not locked into that. It just would require a lot of investment on the technical backend. Is there some way you could help us figure out how to switch from categories to tagging that would allow a less controversial potential integration of our data with Wikidata? --LauraHale (talk) 16:28, 16 November 2013 (UTC)[reply]
No problem. The point you raise about articles being locked in time is of course very valid and I agree that it doesn't make sense to use data from Wikidata dynamically there that can potentially change. It is entirely ok for enwikinews for example to say that Wikidata data should not be used in the main namespace or something like that. No-one will be mad with you about that from the Wikidata side ;-) As for adding sources: This doesn't currently happen automatically but it should be possible to do this with Lua if wanted. As for complexity: Yes this is indeed an issue. Wikidata is still very much in development though and my second priority (after sourcing) is making it easier to use and understand. This includes for example tying in with the VisualEditor. This will however not happen in the very near future because there are still a few more fundamental things that the developers need to work on over the next weeks. As for new contributors: My hope ( and I think it is justified) is that Wikidata will help give a lot more visibility and thereby contributors to the smaller sister projects in the long run by for example having them all linked together in one place. To get back to the Berlin example from earlier: You have the Wikipedia articles, the Wikivoyage articles, images from Commons, news from Wikinews and so on all in one place and easily accessible. We're not there yet of course... We've already attracted more than 5000 new Wikimedians in the past year with Wikidata. This will become a lot more and Wikinews has a chance of becoming visible to them on a lot of pages in Wikidata. As for tagging: I am happy to have a chat on IRC or via Hangout with you or someone else very familiar with Wikinews (ideally with more than just the English Wikinews). --Lydia Pintscher (WMDE) (talk) 16:44, 16 November 2013 (UTC)[reply]
Thank you Lydia. :) There are some places where I could see the value of Wikidata in article writing. It might be nice for covering major news events if we could get locked in time but dynamically created tables to write articles like France leads medal count after third competition day of 2013 IPC Alpine World Championships, which was what someone suggested was currently entirely possible with Wikidata as it stands to do on Wikinews if we would just implement it. The article space is pretty much something I can see as a no go zone except for potentially information like the example. That said, I contribute occasionally to Spanish Wikinews, and have a number of contacts outside of English Wikinews. I have researched policies extensively for Local Wikinews policies. I would be more than happy to chat when I return home following the conference. The tagging thing is a big potential issue. Categories are treated like tags here, and the more categories in an article the better. We would not put an article into Chicago, but into Chicago, Illinois, United States and North America. Each category has (or should have) a dynamic list of the 5 to 10 most recently published articles, and all other articles published using that category in the standard category. Most categories have links to sister projects like Wikivoyage, Commons and Wikipedia. The goal of categories is to make content easier to find for readers, especially given that some category publishing is very erratic in timeframe. Having categories, rather than straight up tagging, is a barrier for a number of new contributors, especially student writers who are used to tagging on things like blogs and other platforms they use. Tagging would also take away some of the conceptual modelling problems the project has and the information hierarchy issues that Wikidata based required usage fears bring, especially when it calls for standardization practices for this across languages. :) --LauraHale (talk) 17:02, 16 November 2013 (UTC)[reply]
Wikidata requires contributors (to Wikidata content) to edit on wikidata, thus competes with sister projects for contributors. E.g. your 5000 new contributors.
Wikidata forms an abstraction layer, and thus obscures the source - if the source is a sister project it's contribution will be obscured rather than highlighted.
Despite these points, which in my opinion are each negative, I believe wikidata is both inevitable and likely will be a benefit across WMF projects for centralizing content control. Given that, in my opinion en.WN should focus on how best to use it to improve the reader's experience. - Amgine | t17:07, 16 November 2013 (UTC)[reply]
Just a small note that Wikidata use on-project does not require Lua. You can use a parserfunction ({{#property:president}}) which provides access. That said, Lua is really the most complete way to access Wikidata information. Legoktm (talk) 21:09, 16 November 2013 (UTC)[reply]
You assume the contributor base is a fixed-size pie. I don't agree. We're growing the pie all together. Yes obviously there is some competition but in the long-run we'll hopefully all have more. The small Wikipedias for example are already now profiting a lot from Wikidata and I wouldn't be surprised if we see growth there in the near future. --Lydia Pintscher (WMDE) (talk) 17:13, 16 November 2013 (UTC)[reply]
First, I hear the 'fixed sized pie' argument often. Yes, in fact, the number of humans on the planet *is* finite, the number of people on the internet is smaller yet, but the more important number is the active contributors to WMF: it's declining. Until these known facts change I will work from the assumption that the pie is finite.
However, to get a better handle on this particular discussion, I did a quick and crude study. I feel confident, since I was conservative there, in saying that yes, Wikidata poaches contributors. I don't think this is an entirely bad thing, but it should not be dismissed as it is real, and measurable. A more-complex study could be useful, but this was enough to confirm what has been assumed to be the case by some. - Amgine | t10:02, 17 November 2013 (UTC)[reply]
I don't think your study is accurate though for various reasons. For example an edit on Wikidata is not the same as an edit on Wikipedia. You can't just compare these by number. Also a large part of these edits are intended to benefit their home-wiki. In addition Wikidata offers quite a different kind of tasks than the other sister project. Those _do_ attract different kinds of people. I'm not saying that there are no people who do less work on their Wikipedia than they would without Wikipedia. But that doesn't mean less work gets done for that Wikipedia. It is not that simple a number game. I would very much like us to look at this more closely and get better numbers since this is obviously something I care about. --Lydia Pintscher (WMDE) (talk) 10:17, 17 November 2013 (UTC)[reply]
Portals are no longer maintained here, though they are updated a bit more on other languages. They are actually being depreciated here as DPL integration has been improved and the ability to do better and longer lists of most recently published articles by topic has occurred. If there was a low project maintenance way of doing tagging that replaced categories (which portals sought to do is my understanding), that would be ideal. The category as tagging issue comes up a fair amount in reviews because many contributors just do not add them. --LauraHale (talk) 17:40, 16 November 2013 (UTC)[reply]
What are the parent categories ?
What are the broadest topical categories for news? This would exclude regional categories, but just the very basic 'sections' of news. I would like to see only 6, but possibly as many as 10.
These seem like common core news sections:
Disaster/Accident
Politics
Crime/Law
Culture
Science/Technology
Business/Economy
Common optional coverage topics which contain non-news-event material:
Some digging finds no distinct category for Financials, it's all lumped under Economy and business. I could see PayPal and've gone looking for a few others to lump into financials (RBOS, Barclay's, Bank of America, Citibank).
I would expect these to be, potentially, totally different on different-language Wikinewses; it's very abstract, and abstracts is where the different Wikinewses are most likely to use different taxons, as I noted earlier. (I might add that it'd be dreadful for individual Wikinewses to not have that degree of freedom.) The overall shape of our category hierarchy always needs work, and every great once in a while I get around to improving it a bit (usually in small ways that don't require a major consensus-gathering project; there's lots of little stuff that can be done, in between major community discussions).
In the checklist at WN:Tips on reviewing articles, in repeating our general advice that every article should have at least one geographical and one topic category, I provide for geocats an expandable category tree, and for topics a link to {{Topics}}. The latter, though a bit of a kludge, is of some interest. Another view of the matter may be glimpsed by viewing an expandable category tree for news articles by section, sections being (as brianmc has remarked) to topics what continents are to geocats.
One of the times I worked on cleaning up our overall category hierarchy was just after I wired the region category tree into the Tips page, because viewing that tree made it immediately visible that we had too many cats at the top level of that tree. I'm tempted to do a bit of cleanup now that I look at the section category tree. The category "women" has always worried me in how it ought to relate to the rest of the hierarchy. We'll probably never have a category "men". Is there some natural larger grouping of topics that "women" ought to belong to? I'd also really like to have a clear definition of what does and doesn't belong in category "women". --Pi zero (talk) 14:38, 16 November 2013 (UTC)[reply]
<chuckles> @Brian: I'm trying to engage a *discussion* of what English language contributors and readers see as the broadest topics which cover as near to all news articles as possible. I mean, I know what I think they should be, and what the current usage is on en.WN, but do people feel that these really are the best?
@Pi zero: my personal experience trying to understand French news sites finds the top six are still very much a useful guideline. Much more is classified as cultural which might be classified as political in English. And some sites seem to add 'Conflict/War' as a category. I hope to engage in this conversation with native speakers to get their opinion. I hope Wikinewsies with multilingual skills will also engage other communities of newsies, and then a discussion can be begun in the Wikinewsie Group to harmonize our root categories as much as possible in order to improve our search outcomes.
Ultimately, the goal is to improve our methods of detecting common stories across the languages, and to use that information to add interwiki links and encourage collaboration. - Amgine | t15:01, 16 November 2013 (UTC)[reply]
Improving methods of detecting common stories sounds cool. I find it depressing to realize how very unsuited Wikidata itself is (in its current form, of course; one can always hope); but, yeah, once one gets past the black-and-white match/non-match Laura mentioned, deep understanding of how the different taxonomies relate to each other could be seriously useful. --Pi zero (talk) 15:12, 16 November 2013 (UTC)[reply]
Technical problem on french Wikinews
Hello everyone,
Since the lastest update of Mediawiki we have many problems on french wikinews and we think our fr:Mediawiki:Common.js is the cause but we do not know how correct it. For example on french Wikinews fr:Modèle:Boîte déroulante or fr:Modèle:Ticker are broken.
I asked Bawolff to look at this, and he discovered an error in the common.js which was preventing it from loading any of the js. (Of course, this was *after* he modified the collapsible navbox template to use the mw-collapsible classes.) Mattho69 implemented the fix Bawolff suggested. So, hopefully everything is now working! - Amgine | t10:25, 16 November 2013 (UTC)[reply]
