Community noticeboard

Archives:
 * Archive 1 (23 July 2017 - 25 November 2017)
 * Archive 2 (1 December 2017 - )

Install GlobalUserPage
I propose that we install the GlobalUserPage extension. It is used by the WMF and would be very useful on miraheze. Those who would not want a GlobalUserPage can use or __NOGLOBAL__ on their user page. MacFan4000 (Talk Contribs) 17:52, 3 February 2018 (UTC)

Votes

 * 1)  as proposer. MacFan4000 (Talk Contribs) 19:44, 3 February 2018 (UTC)
 * 2)  as long as abusefilters are setup to prevent global userpages to be used as an advertising place. Zppix (Meta | CVT Member | talk to me) 17:56, 3 February 2018 (UTC)
 * 3)  (as it is currently) My issue with this is the fact that you need to "opt-out" if you don't want it, not that you need to "opt-in" if you do. I would definitely support this if it's disabled by default, as then users decide whether they want a global userpage or not. Reception123  (talk) ('C' ) 18:03, 3 February 2018 (UTC)
 * 4)  See my previous comments on T1681. —AlvaroMolina (✉ -  ✔ ) 18:06, 3 February 2018 (UTC)
 * 5)  Esta función me parece perfecta. La he probado en Wikimedia y la veo útil. Así no tienes que actualizar cada wiki con nueva información o modificar el existente. Solo tienes que ir a Meta y actualizarlo. Wiki1776 (talk) 18:13, 3 February 2018 (UTC)
 * 6)  There is only one of me, and Mirahezians wanting to write to me ought not have to research which wiki has the busier talk page (though there is a hint in my signature).  If users want project-specific user pages, this is easily done and the global user page is an ideal directory of them.  Reception123 is right to be concerned about users getting opted-in involuntarily, but in this case, reacting to it is simple.  I support it with either opt-in or opt-out.   18:47 3-Feb-2018
 * 7)  Would change vote to support if it were opt-in and not opt-out. CoolieCoolster (talk) 19:44, 3 February 2018 (UTC)
 * 8)  I have no problem with this and the caveat isn't an issue for me either as I trust Miraheze admins.  20:28, 3 February 2018 (UTC)
 * 9) I also think this should be opt-in (  opposed to current NOGLOBAL opt-out). &mdash;  revi  03:07, 4 February 2018 (UTC)
 * 10) For the betterment of the community, I believe that it should be opt-in not opt-out. I thoroughly believe only people interested in using such a feature will actually use it, so there is no need to make it opt-out. I fully support this motion if it is opt-in. &#32;  Miraheze Logo.svg CnocBride | Talk | Contribs  12:20, 9 February 2018 (UTC)

Comments/Questions

 * Making the feature opt-in can be achieved by setting wgDefaultUserOptions['globaluserpage'] = false; However, this requires the addition of the global preferences extension as well. -- Void  Whispers 20:47, 14 February 2018 (UTC)
 * Per https://phabricator.wikimedia.org/T184643 GlobalPreferences passed WMF security review. MacFan4000 (Talk Contribs) 22:02, 14 February 2018 (UTC)
 * If I remember correctly, wasn't there a performance issue related to GlobalPreferences, pointed out by Labster? Reception123 (talk) ('C' ) 19:43, 16 February 2018 (UTC)
 * I would imagine that was GlobalContribs, if only for the reason that WMF is adding GlobalPreferences, while a phab task on GlobalContribs hinted that the performance impact from GlobalContribs would make it impossible to use with more than several hundred wikis. -- Void  Whispers 20:04, 16 February 2018 (UTC)

Extension reviewing
Extension reviews seem to be the biggest backlog as far as Phabricator requests are concerned. Perhaps Miraheze should hire someone to help review them? CoolieCoolster (talk) 00:50, 8 February 2018 (UTC)
 * Yes, extension reviews are a big issue currently as the two reviewers are both busy. Unfortunately, I do not think we have enough budget to actually hire someone to do these reviews. Reception123 (talk) ('C' ) 06:04, 8 February 2018 (UTC)
 * How much time does it take to review an extension? Depending on how much it would cost, perhaps I could hire someone to review several extensions for Miraheze. CoolieCoolster (talk) 11:57, 8 February 2018 (UTC)
 * I'm not personally sure about how much time it would take to review an extension, but I imagine it does take a while. A reviewer must make sure that there are no security vulnerabilities in the code, and thoroughly analyse it. Reception123 (talk) ('C' ) 20:01, 8 February 2018 (UTC)
 * Alright. I will ask a PHP developer how long they think it would take to properly review an extension and will compile a list of the unreviewed extensions on Phabricator, and will then see whether or not it is economically worth paying a PHP developer to review the extensions. CoolieCoolster (talk) 20:56, 8 February 2018 (UTC)

$30 an hour
Here is the list of extensions under review: https://podpedia.org/wiki/User:CoolieCoolster/Miraheze_List_of_Extensions_Under_Review There are quite a few of them, and the PVP Dev charges $30 an hour and said that it would take them about an hour on average to review each extension so we need to pick and choose which ones should be reviewed. I can probably pay for three or four extensions to be reviewed. The extension being made for my wiki will be done in about a week and a half and I want it reviewed ASAP so I can use it, so two or three other extensions should be picked for the dev to review. It seems like some of the Phabricator requests for extensions are for unmaintained wikis, so perhaps some requests should be declined. CoolieCoolster (talk) 18:43, 9 February 2018 (UTC)
 * 30$ is quite a high price tag. I would not be in support of this unless the review of these extensions are absolutely necessary to the smooth running of a community. &#32; Miraheze Logo.svg CnocBride | Talk | Contribs  18:52, 9 February 2018 (UTC)
 * In this case it would just be me paying for it, although perhaps I should look for someone who is willing to do it for cheaper. Cheaper might also mean not as good of a review though, since the person $30/hour person is a MediaWiki expert. CoolieCoolster (talk) 19:11, 9 February 2018 (UTC)
 * I don't think $30 per hour (£22 for those in the UK) isn't overly expensive for something specialist such as PHP developer fees. Of course it would cost a small fortune to get all the extensions/skins reviewed so realistically I can't see that happening but I guess if people are willing to pay or make additional donations to get their requested extensions reviewed sooner it's not a bad idea. I would pay to get my requested skin (Pivot) checked if it meant having it sooner, but sometimes you just have to be a little patient for a service that's free. 21:59, 9 February 2018 (UTC)
 * No, $30/hour is a bargain for consulting by a MediaWiki expert. However, where we left it last episode (Archive 2 § 27) is that certain extensions are not vital for the community (which can certainly "run smoothly" without them) but only for Coolie to induce authors of podcast wikis to come to his Podpedia.  There are many questions languishing in Phabricator and Coolie is welcome to pay to expedite any of them, but this is not a compelling case to use Treasury money.  It would be nice if Stewards review Coolie's choice of developer before he invests any money.
 * In detail: As well as paying a MediaWiki expert to review a technical question (namely, does a given MediaWiki extension jeopardize the health and security of the system?), it is also easy to pay a fake expert to deliver such an opinion as favors the person paying for it.  I read above that we have two Mirahezians who could do such an evaluation and they are busy.  The rest of us presumably do not know how to evaluate an extension.  The consultant will have to do more than report that "This extension is safe"; he would have to enumerate possible threats to Miraheze and explain to those without his skills why the extension does not deliver any of the threats.  If there are technical credentials a professional could provide to certify expertise with MediaWiki, which would people accept?  An agency with a reputation to preserve would be less likely to bang out an opinion to satisfy the payor, but it will charge much more than $30 an hour.   03:33 10-Feb-2018
 * Small correction: Extension decision is up to Labster and his delegates (currently Samwilson). &mdash; revi  05:39, 10 February 2018 (UTC)
 * I am willing to donate to Miraheze or pay Labster or Samwilson for faster extension reviewing if that is at all possible. In a week and a half my extension will be ready and I will need it for my wiki to grow. CoolieCoolster (talk) 06:12, 10 February 2018 (UTC)

Load times, previews and saves very hit and miss
Is anyone else experience issues with load times? For about the last week I have had multiple problems with page load times, previews and saves not working and periodically, nothing loading at all. This is getting somewhat difficult to work on my wiki at the moment as everything is taking so much longer. Am I an isolated case or has anyone else been experiencing similar issues? Currently none of my pages are loading; they're just stuck on a blank screens, yet I can post here on Meta. Can someone enlighten me as to the problem and if there will be a resolution soon? Many thanks. 11:51, 11 February 2018 (UTC)
 * I blame most slowness on my own affection for Walmart-cobranded no-contract Internet service; but yes, in the last week, I have been staring at "Waiting for meta.miraheze.org" for up to ten seconds. No complete inability to load, though, and it seems all right just now.   13:14 11-Feb-2018
 * Yes here too, over a week now getting 504 Gateway Time-out. There ought to be something on @miraheze but there is no info there. --Rob Kam (talk) 13:22, 11 February 2018 (UTC)
 * Rob Kam, Borbeman, Spike, a mi siempre me ha salido el 504 Gateway Time-out en todas las wikis que he creado. Nunca he podido saber el porqué, aparentemente es la conexión pero... Me funcionan otras páginas como Wikipedia, Wikia, y otros. Recuerdo que lo he comentado en el IRC pero no he recibido respuesta o han sido "vagas". Espero que se pueda solucionar para poder realizar mis proyectos sin problemas. Wiki1776 (talk) 14:28, 11 February 2018 (UTC)
 * It's good to know it's not just me but I'm sorry to hear others have experienced the same or similar problems. I know there have been no issues with my ISP to date as I have been streaming films and episodes on Netflix with no issues whatsoever, and other websites load properly and speedily. Just like  and, I also get 504s, but when pages just stops to load entirely (loads about ⅓ of the way then stops dead) this is really frustrating. The twitter feed / facebook page would certainly benefit users if they were updated more regularly even if to say they acknowledge they are aware of service issues and identify them if known. Having checked again, none of the pages I have just refreshed are loading at all, not a single one, and I had planned a fair bit of work for today. Staff are usually on the ball at sorting issues like this out but I hope it doesn't last another week.  15:37, 11 February 2018 (UTC)
 * Update: Pages seem to be loading properly again at the moment. All those that previous froze have now refresh (some pages took well over an hour). If staff have fixed it, thanks. 17:14, 11 February 2018 (UTC)
 * PS — Sometimes, even after a page renders, the little dial keeps spinning (as Miraheze fishes for style and JavaScript to apply to it). Sometimes my browser gives up.   19:34 11-Feb-2018

Wikispaces is shutting down
Wikispaces, a major wiki farm, will begin shutting down in a few months, and its last wikis will shut down on January 31 of next year. Perhaps this is an opportunity for Miraheze to convince customers of the soon-to-be-gone Wikispaces to move to Miraheze? It could be an opportunity to gain more funding since universities and organizations that have paid Wikispaces in the past for their wikis could donate that money to Miraheze instead. Here is a link with information about their shut down: http://blog.wikispaces.com/ CoolieCoolster (talk) 21:03, 14 February 2018 (UTC)
 * I volunteer to be the contact person for any college or university interested in making the switch. I looked seriously at Wikispaces as a possible alternative in case Miraheze ever closed down.  I discovered that Miraheze is in many ways superior because it's syntax so closely mirrors the Wikimedia Foundation.  I am also looking to help out with any grants, especially those involving Quizbank and designing courses compatible with https://openstax.org/ or http://astro.unl.edu/.  Also, there is something called CPR that needs to be made compatible with a wiki.  It was developed as a free resource by UCLA, and then sold to a for-profit organization.  There is no reason why a system like this shouldn't be open source.  For more info on CPR, visit http://cpr.molsci.ucla.edu/Home.aspx.  My CV is at http://www.wright.edu/~guy.vandegrift/shortCV/GVCV.htm.  I am close enough to retirement age that having tenure is not a big deal and would gladly work full-time on soft money to get this effort going. --Guy vandegrift (talk) 15:16, 15 February 2018 (UTC)
 * Here is a list of educational Wikispaces wikis: http://educationalwikis.wikispaces.com/Examples+of+educational+wikis CoolieCoolster (talk) 19:08, 15 February 2018 (UTC)

A quibble
A quibble: When a university develops a resource for free distribution, then sells it to a for-profit organization (whose own vision of the future might be: profit), the question is not, "Shouldn't it be open-source?" but "Is it open-source?" 16:01 15-Feb-2018
 * I am not certain about this, but I don't think the university "sold" it because for several years they hosted a free version that a colleague of mine used for an online economics course. He was glad the university offered the service because the other one cost something like $500.  My understanding is that the university took down the free version only because they replaced the computer.  CPR is vaguely analogous to the systems of "likes" or "shares" used in social media, except that there is quality control.  A student is unable to submit their essay until the student has correctly assessed other essays that have already been "graded".  And, each student's grade is based partly on how accurately they "grade" recently submitted essays that have not yet been graded.  To keep people from copying each other's essays, all essays must be about a recent news article.  It's a bit more work than the crazy "likes" and "shares" that facebook uses, by my colleague assured me that CPR was a labor-saving device for him.  My argument is that if a robot can drive a car in city traffic, we should be able to teach college courses with a lot less effort (before college they are still children and need real people to encourage them.)--Guy vandegrift (talk) 17:16, 15 February 2018 (UTC)
 * Without a doubt, college could occur "with a lot less effort" and a lot less student debt. There is no magic about the lecture-hall setting; colleges excel not at teaching but at certifying that the student has learned.  As for saving effort, there are institutional forces to the contrary, from teachers' unions to our sudden monopoly on student lending.
 * But your speculation becomes: How to enhance the Wikilove extension ("thanks" for posts) to improve the quality of the votes. Without baking it into the wiki, Uncyclopedia has a system by which authors review each other's articles.  Users are encouraged to submit reviews and not just request them, and one's own reputation is at stake if he is thoughtlessly negative.   18:53 15-Feb-2018
 * (PS—Actually, one's reputation is not at stake but only the reputation of one's user name. And there is no comparable assurance against cribbing someone else's review, except that a typical review request leads to only one review—though the reviewee can "request a second opinion."   19:01 15-Feb-2018 )
 * You are right about the inertia in the system regarding "setting education free". But it doesn't hurt to push in the right direction.  Regarding the "peer review" in CPR, each instructor has complete control over the peer review in the present configuration of CPR.  The instructor selects the papers that the students must successfully review before the article can be submitted, and to all but the instructor, the peer reviews are completely anonymous; neither reviewer nor reviewee know each other's true identity or even username.

Formats for moving an account here
In the past, when wiki.wiki was shutting down, we were open for possible incoming migration, AFAIK there were none. If someone would like to be imported, sure, we can handle it. (PS: We only accept XML backups. (I don't know if they accept this format as export option.) HTML, as explained in their migration doc is not a standard form for MW import. Our server is 100% Unix (in fact, Linux) so you should choose Unix format, too.) I have sort of draft (without knowing this) in Moving wiki to Miraheze, improvements welcome. &mdash; revi  16:49, 15 February 2018 (UTC)
 * I agree with revi. This is a good opportunity, and if anyone wants to migrate from Wikispaces, we would be happy to host them :) We strongly prefer XML dumps, but if the only possibility is SQL dumps, that could also be done, if necessary (and if the version is 1.30). But HTML dumps or anything (as revi said above) are not accepted for technical reasons., thank you for volunteering to be a contact person. Hopefully we will get wikis willing to migrate. Reception123 (talk) ('C' ) 18:08, 15 February 2018 (UTC)
 * As always, the key is advertising, and this time I don't mean a big-budget professional agency. We ought to have a liaison to Wikispaces.  They might be reluctant to give Miraheze an official endorsement, but they are as motivated to give their stranded users information about one fine replacement service, as we are that they do so.  Their closure announcement, linked to above, has the flavor "Have a nice life!" but there is no reason it couldn't include something of the form, "Here is one place you might try!"   18:57 15-Feb-2018

Quick skimming smells like they’re not MediaWiki; I’m not sure. Anyway, if their export tool cannot generate xml file that Special:Import and importDump.php understand, we can’t help them with their ‘backups’. &mdash; revi</tt>  21:26, 15 February 2018 (UTC)
 * Well, I remember when MWZip (another hosting service) shut down, they mentioned Miraheze and another wiki hosting service in their announcement. Either Wikispaces could do that, or they could simply link to a list with all wiki hosting services (such as the one on MediaWiki.org) Reception123 (talk) (<font color="#FF0000">'C' ) 15:02, 16 February 2018 (UTC)

Backup
Hi,

Is it possible for me to make my own backups of my wiki? And in that case - how?

/Madelene
 * It is possible to have backups, but these must be done by sysadmins (either manually once, or automatically at set times). Please read Backups and make a request on Phabricator. Reception123 (talk) (<font color="#FF0000">'C' ) 14:27, 16 February 2018 (UTC)

The function "main" does not exist
Recently I asked for an import of files to my wiki which seemed to break several things. I fixed most of the issues, however it says that the function "main" does not exist, which after some research makes me believe that my wiki's MediaWiki.php file got edited during the import, although I'm not sure. If anyone knows how I can fix this, please let me know! CoolieCoolster (talk) 19:12, 19 February 2018 (UTC)
 * CoolieCoolster ¿conservas el mensaje completo? ha estado ocurriendo algunos problemas de importación, ver solicitud T2693 en Phabricator. Wiki1776 (talk) 19:30, 19 February 2018 (UTC)

Manage a wiki - from private to open
When I started thegreatwar wiki it was private. When I realised it had a lot of potential I decided to make it open by unchecking the "private" option in "Manage a wiki." That was quite a while ago but I have noticed that the site is not being indexed by google and it doesn't appear in Miraheze's wikistats on wmflabs either. My question is if a wiki was original set to private but then changes to open, will search engines etc. eventually find it and start indexing or does the private option prevent this happening (even when the setting is changed at a later date)? 11:22, 21 February 2018 (UTC)