User talk:Johnuniq

I'll reply to messages here, unless requested otherwise.

Index of stuff

• Comments re free speech
• Jagged progress • more • genre articles
• Links to highbeam.com from dumps of external links • earlier • more Credo • JSTOR results • JSTOR + Cochrane January 2015
• Gregorian serial date template
• Analysis of EL dump using Amazon S3 very interesting methods from GabrielF
• Module:Number at Hindi Wikipedia request from Siddhartha Ghai

Strange edits on ccTLDs?

Can you take a look at the ccTLD pages. There have been some rather strange additions/edits to these pages (cu/ke/al/br/in and a pile of others) with fake stats being added by IP addresses. The .cu one was a bit mad with a claimed registrations number of about 600K when there are only 6K according to the cu NIC site. The .cn ccTLD didn't have 49 million domains and .co didn't have 69 million either. Seems to be a persistent pattern of iffy edits. Thanks. Jmccormac (talk) 20:38, 18 July 2017 (UTC)

@Jmccormac: Ouch, a lot of pages are listed at {{ccTLD}}, and a lot of nonsense. I don't want to watchlist them all, but if I have some time I'll review "Related changes" (Special:RecentChangesLinked/Template:ccTLD). Please let me know if there are other pages being hit. Presumably the motivation is some excitement about making money by exploiting domain name registrations. Thanks for noticing the problem. Johnuniq (talk) 05:01, 19 July 2017 (UTC)

It looks like persistent vandalism on ccTLD pages from an IP rather than any attempt to fluff up numbers. There are fake stats being added and edits with actual sourced stats being reverted along with nonsense claims mentioning 2018 etc. I just reverted the co and br pages to their last known good versions. The .cn, .al, .ro and .in pages have all been hit today again. It seems to be an ongoing issue. Jmccormac (talk) 14:40, 19 July 2017 (UTC)

@Jmccormac: I see what you mean. I can't tell whether it is an attempt to be helpful, or malicious/playful/trolling. Please see my sandbox (permalink) where I collected some information. Feel free to add anything pertinent. I was thinking of posting a report at WP:AN to get opinions. What do you think? Johnuniq (talk) 04:58, 20 July 2017 (UTC)

By the way, the reason I used a sandbox is because I am looking for extra evidence. If you can point to unequivocally bad edits, please do so. Johnuniq (talk) 06:46, 20 July 2017 (UTC)

The reverts of good edits are continuing and fake stats are still being added and stats from the registries are being removed. It is a widespread issue (just checked .br and the registry domain count with link has been replaced with fake counts.) The start date on some of ccTLD pages has an "available since" date added that is prior to the launch date. There is also a history of fake stats being added to smaller African ccTLD pages like .sn with 69 million domain names. (https://en.wikipedia.org/w/index.php?title=.sn&type=revision&diff=789424502&oldid=787261790). On the .gn page, there's the fake stats and the claim that the ccTLD has been active since 1990 when the IANA (https://www.iana.org/domains/root/db/gn.html) registration date for the ccTLD is 1994-08-09. There are similar incorrect launch/registration date edits to other ccTLD pages. The IANA database is the official TLD database and details the launch dates, sponsors, contact details and nameservers. The Ugandan ccTLD page has had fake stats added: (https://en.wikipedia.org/w/index.php?title=.ug&diff=prev&oldid=789078389 ). The Nigerian ccTLD page also has fake stats and start date information. (https://en.wikipedia.org/w/index.php?title=.ng&diff=prev&oldid=789254775) The Albanian ccTLD page has similar iffy edits and additions: (https://en.wikipedia.org/w/index.php?title=.al&diff=prev&oldid=789504051) The .ph page also has had fake stats added: (https://en.wikipedia.org/w/index.php?title=.ph&diff=next&oldid=790390662 ) The .gm ccTLD page has had fake stats (https://en.wikipedia.org/w/index.php?title=.gm&diff=prev&oldid=789182180) and false start date information added. (https://en.wikipedia.org/w/index.php?title=.gm&diff=prev&oldid=789636457) It looks like a longterm pattern of fake stats and start dates being added to what are primarily ccTLD pages. The .ke page has also had a fake 900,000 domain count added but the infobox has the accurate figure from the registry. Some of these smaller ccTLD registries don't publish domain counts. Jmccormac (talk) 16:05, 20 July 2017 (UTC)

@Jmccormac: I have a good understanding of DNS and related issues, but am very much out-of-touch with the details that the IP is engaging. People at a noticeboard where I might report this will have even less knowledge. That means easy-to-understand evidence has to be prepared. Evidence that uninvolved editors can check.

It seems unlikely that .sn would have 69 million registered domains, but weird stuff happens with domain registration and for all I know there are a handful of "valid" domains, and 69 million junk domains that have been temporarily registered. It is very unsatisfactory to add numbers like that with no source, but how is it known to be fake?

Consider diff at .br. Someone who is not familiar with the details (such as myself) will look at that diff and see perfectly reasonable additions of citation needed and other minor adjustments. Please spell out exactly which text is blatantly wrong, and how the fact that it is wrong can be checked. I can see that adding "registereddomains=950,000" with citation needed is unhelpful and could easily be trolling, but how is it known to be wrong? Adding a URL with a nonexistent domain (http://www.nic.net.br/) is very unhelpful. However, is that a URL which might work for some TLDs? If it is, the addition is just unhelpful, not sanctionable. The text added is inappropriate, but it does not appear to be vandalism.

The "S.O.C.O." edit summaries appear to be related to S.O.C.O. (Scene of the Crime Operatives). There is no doubt that the edit summaries are garbage/trolling, but onlookers at a noticeboard would regard them as merely a quirk of the editor, and not sanctionable.

@Kvng: Would you mind reviewing a few edits by 81.135.204.201 and giving an opinion on this mess? Johnuniq (talk) 10:15, 21 July 2017 (UTC)

The two I've looked at are a mixed bag. this one dithers around quite a bit and has not particularly helpful edit comments but seems to land well. this one looks like a good faith effort also dithering and unhelpful edit comments but is not a net improvement (and was reverted). I'm not sure there's anything nefarious going on here but it looks like there will be more effort required to review these edits than is justified by the potential improvements but that's a general issue we have with an encyclopedia anyone can edit. ~Kvng (talk) 14:50, 21 July 2017 (UTC)

Have been going back through the dubious additions and fake stats. Some of the stats are available from the registries and I've added those with links and links to subdomain naming policies. @Kvng I added the domain name count from the Chinese registry count. I added the actual domain name counts from the .br registry site with the refregistereddomains like to the .br registry stats page but the IP editor removed them and replaced the fake stats. There is a history of the IP editor adding fake stats along with a CN tag. The IP editor also reverted @Yamaguchi先生's edits/reversions. Some of these fake stats are just completely weird like the 69 million domains on various ccTLD pages. The .com tld only has about 127.7 million domains so any large addition or increase on a ccTLD would not have gone unnoticed. There are also repeating numbers in these fake stats as if they are cut and pasted. Some of these ccTLD registries are very small operations and are often run out of local university Computer Science departments. They don't go adding tens of millions of domain names when their domain name counts are often below 50K. That would increase their ccTLD zonefile from a relatively small zone of a few hundred KB to one approaching about 7GB or so. The .ly registry and .ro registry publish their domain counts on the index pages of their registry site and I've added these. The .ng edit introduced a PPC parked website, the nic one, as a registry site. It also had the 69 million domains fake stats added. I've cleaned it up and added the official registry count and reference. Some of these small ccTLDs have enough problems without iffy Wikipedia stuff adding to them. It might be necessary to step back through these edits. Jmccormac (talk) 21:13, 21 July 2017 (UTC)

Looks like the IP editor is back reverting good edits, adding fake stats, messing with start dates. Jmccormac (talk) 12:29, 22 July 2017 (UTC)

Thanks for your work. I did some reverting, but the issue will need to be escalated. I would like some way of demonstrating to uninvolved editors that at least some of the numbers are fake. I need a diff and and a brief explanation that a particular number is wrong because ..., with one or two links that would allow the fact that it is wrong to be checked. If that is not possible, what is the best that can be done? Johnuniq (talk) 10:34, 23 July 2017 (UTC)

@Jmccormac: I expanded my sandbox. Please have a look and see if there are any obvious mistakes, or whether anything more can easily be added. Johnuniq (talk) 10:53, 24 July 2017 (UTC)

There's another revert on .ke this morning which completly contradicts the domain count in the infobox. ( https://en.wikipedia.org/w/index.php?title=.ke&oldid=prev&diff=792081666 ). There are some other iffy edits on other ccTLD pages like .ci where the same fake stats claims (https://en.wikipedia.org/w/index.php?title=.ci&diff=next&oldid=789075447) are applied. According to the .ci registry page .ci registry it only has 9,158 domain names but the iffy edit has it at 9,500,000. There are some sources on the web for gTLD counts ( http://www.registrarstats.com/TLDDomainCounts.aspx ). The Canadian ccTLD registry has its 2016 annual report with the top ccTLDs and their domain counts ( https://cira.ca/factbook/domain-industry-data-and-canadian-Internet-trends/domain-name-industry ). It shows the .co count at 2,018,450. Some of the other ccTLD registries publish periodic reports which also have similar stats on the ccTLDs. Took a quick look at the sandbox and it seems OK. Will go through it in detail later today. Jmccormac (talk) 15:50, 24 July 2017 (UTC)

Just checked the .ng and it has been hit again. The IANA start date for .ng is 1995 ( https://www.iana.org/domains/root/db/ng.html ) but the nonsensical edits have 1990 and 1994, the fake 69 million domains stat and a lot of fake subdomains. Jmccormac (talk) 16:32, 24 July 2017 (UTC)

Thanks, that .ci information is very helpful and I added it to the sandbox. In about 24 hours I will post at ANI and I will ping you here to let you know. You might then add a brief comment at ANI to the effect that you are familiar with domain name management and am sure the edits are nonsense. Johnuniq (talk) 11:29, 25 July 2017 (UTC)

Thanks. I spent a few hours going back through the edits on the smaller ccTLD articles. It seems that the .uk and .dk articles were not hit but they are some of the larger European ccTLDs. The ones being hit were small ccTLDs that often had no published domain count or accessible registry site. Some of the 69 million domain edits were in complete disagreement with the registry counts in the infoboxes. The ccTLDs, apart from the repurposed ones like .co and .tk, tend to have a fraction of the population of the country and the registrations are dependent on the state of the country's internet infrastructure and economy. The ccTLDs with over a million domains tend to be in larger and more developed economies. Jmccormac (talk) 15:47, 25 July 2017 (UTC)

There is a reference in an ICANN study on the African DNS market ( https://www.icann.org/en/system/files/files/africa-dns-market-study-final-06jun17-en.pdf ) where it refers to the size of the total African ccTLD market on page 117. It is referring to the Domaintools site (http://research.domaintools.com/statistics/tld-counts/ ) and estimates the total number of domain names in these ccTLDs to be about 2.9 million. Jmccormac (talk) 17:19, 25 July 2017 (UTC)

Excellent, thanks. I added that to the report which is now at User:Johnuniq/ccTLD. Johnuniq (talk) 22:54, 25 July 2017 (UTC)

@Jmccormac + Kvng: Please see the report at ANI. Johnuniq (talk) 05:53, 26 July 2017 (UTC)

Thanks. Commented on the ANI. The IP addresses seem to be dynamically assigned and will change every few weeks or when the router is rebooted. Semi-protection of the ccTLD articles might be the best option. Jmccormac (talk) 13:43, 26 July 2017 (UTC)

A barnstar for you!

	The Original Barnstar
	Thanks for warning me about nonsense ~Kvng (talk) 04:24, 21 July 2017 (UTC)

It was very easy, but thanks! I think you know that we are active in networking articles (me not much), and I must have been watching your talk from something that arose a couple of years ago. It's amazing how far nonsense can spread. Johnuniq (talk) 04:27, 21 July 2017 (UTC)

It is also very easy to give barnstars to competent editors such as yourself ;) ~Kvng (talk) 04:47, 21 July 2017 (UTC)

Transclusion problem

Hi John, I was wondering if you would look at a problem I'm having on User:SlimVirgin with the transclusion of User:SlimVirgin/Articles (collapsed within the "Articles" banner). As you'll see, when you click on "show", the top of /Articles is missing from the top but it appears at the end. Without going to a lot of trouble, do you have any idea what might be causing that? SarahSV ^(talk) 21:02, 21 July 2017 (UTC)

Hi Sarah: I started playing with this at User:Johnuniq/sandbox2 which transcludes Template:Convert/sandbox2 (an almost never-used template). Using a template means I can test it more easily. Interestingly, the problem goes away if I do a global search and replace, changing {{TFAIMAGE}} to {{green}}. I'll be elsewhere for a while but will ponder this more, and may have to resort to WP:VPT. Johnuniq (talk) 02:35, 22 July 2017 (UTC)

John, thank you! I hate to cause you so much trouble. Perhaps I should just get rid of {{TFAIMAGE}}? SarahSV ^(talk) 02:55, 22 July 2017 (UTC)

No, not yet. Let me have a look. I'll do that soon. Johnuniq (talk) 03:40, 22 July 2017 (UTC)

There's no rush for this, by the way. It's been like that for a while, so please take a look at your own convenience. I'll be very grateful for any help at any time. SarahSV ^(talk) 03:55, 22 July 2017 (UTC)

@Sarah: Weird! Have a look at it now. I temporarily removed the "::" indent before TFAIMAGE. That messes up the appearance, but it seems to avoid the problem. I'll continue looking (techo people love techo problems) but if you are around you might say if the transclusion works on your system now. Johnuniq (talk) 04:30, 22 July 2017 (UTC)

That's so bizarre. Yes, page is showing properly now. Thank you! SarahSV ^(talk) 05:03, 22 July 2017 (UTC)

My last experiment broke it again. I'll leave it for a while but will return. Johnuniq (talk) 05:57, 22 July 2017 (UTC)

@Sarah: I'm planning to post at WP:VPT asking for help regarding the simplified test at User:Johnuniq/sandbox2. Do you mind the attention that would bring? I won't do it at the moment as I'll move on to other things. Johnuniq (talk) 10:12, 22 July 2017 (UTC)

Thanks for checking first. I think I'd prefer not to have that much attention on the page. I'll probably either remove it from my user page or remove "::". Thanks again for looking into it. I wish I could remember what I did last time to get rid of it. SarahSV ^(talk) 00:08, 23 July 2017 (UTC)

You could fix the problem by editing Template:TFAIMAGE and deleting the newline at the end of the long line that starts with "}}". That would move </div><noinclude> from a separate line to the end of the long line, with no space between the two lines. That should not cause any problems, but explaining why you had done that edit to whoever is watching the template may take a fair bit of effort.

Meanwhile, I have edited User:SlimVirgin/Articles to expand the TFAIMAGE template so it is not used. That works around the problem, and I guess you would be happy to work with that. I do not think "id="mp-tfa-img"" which the template includes is needed on your page so I omitted it.

It's a bit irritating how the TOC includes "About me" + "My work" + "Contact". I don't see a clean way of avoiding that.

I believe this is now settled (no VPT). Let me know if there is something else, and congratulations on that fantastic list! Johnuniq (talk) 05:14, 23 July 2017 (UTC)

John, I see you've fixed it. Thank you so much for working your magic! I'll edit that template to avoid anyone else experiencing this and will have fun trying to explain why. SarahSV ^(talk) 02:11, 24 July 2017 (UTC)

Gamergate controversy

Please note that I am not very familiar with the Gamergate controversy other than what I read here and there. The controversy isn't very well known in my area also the article(s) I posted is not about gamergate but a detailed criticism of the video series in question. They may be related but the not a reason by the user to remove criticism from someone (In this case who also happens to be a feminist) that is not part of the Gamergate movement. I have removed anyparts that apeared to be OT in anyway but summerizing over 5 articles is hard so I only posted about the articles and a few relevant parts from the 1st article. UmdP 11:01, 22 July 2017 (UTC)

@UMDP: This relates to my post at your talk. The facts overwhelm opinions in a situation like this. Tropes vs. Women in Video Games is obviously part of Gamergate controversy (search the latter article for the title of the former), and as I mentioned, Talk:Tropes vs. Women in Video Games has a box at the top saying "This article is subject to discretionary sanctions"—the "subject" links to the GamerGate case. It was clear that you are not familiar with the background because edit warring on an article under discretionary sanctions is not done. If you want another opinion about what I posted, you could try asking at WP:HELPDESK. That is more a how-to page, but a quick question will probably get a direct answer. For example, you might ask whether WP:NOTCENSORED applies to the edits in question (as I said, no it doesn't). Johnuniq (talk) 10:16, 23 July 2017 (UTC)

Deleted Controversial Content from Casanova

The "controversial" content is about homosexuality. I thought it would less likely be deleted if I titled it controversial instead of homosexual. Garfield7380 (talk) 04:33, 29 July 2017 (UTC)

The content was deleted from another article by a homophobic editor. It may be of interest for some readers to know that Casanova speaks of homosexuality in his own memoir. Garfield7380 (talk) 04:43, 29 July 2017 (UTC)

This concerns Histoire de ma vie and should be discussed at the article's talk page (Talk:Histoire de ma vie). Recent edits diff added a "Controversy" section with text extracted from a primary source by an editor. It does not matter what the section title is (although "Controversy" is not a good choice—who said it is a controversy?). Suppose the book contained a couple of paragraphs on how lovely roses are—should the article have a section on roses? The answer is no, not unless a reliable secondary source has described the author's views on roses, or otherwise commented on the significance of the text. That is basic policy—WP:SYNTH, as mentioned in my edit summary. If there is to be a reply, please copy the wikitext from here (I fixed the formatting of the original post), and make a new section at the talk page of the article. By the way, there is no point linking Histoire de ma vie when writing on the page with that title. Johnuniq (talk) 07:19, 29 July 2017 (UTC)

Template:Extract

Hi! I would like to use the name Template:Extract for another template with a very different purpose (see this essay I'm developing). I checked Special:WhatLinksHere/Template:Extract and it looks like there are zero articles currently using it. However I notice that the template calls the Module:Age, and I'm not too sure on how Lua modules work, so I'm afraid I may break something if I just move the current Extract template to some other title and then code the template I need under the title "Template:Extract". I noticed you're the most active user on the Module:Age, so I thought you could help me out. Can we somehow free the "Template:Extract" title so I can re-use it? If not, I may come up with another name for the template, but "Extract" is really the most appropriate. Thanks ! --Felipe (talk) 14:30, 3 August 2017 (UTC)

In principle, no problem. However, I would need a couple of days to collect my thoughts and work out a new name, and handle the small number of places where the template is used. I put a lot of effort into Module:Date and Module:Age but was then too exhausted to get much work out of them. They are more consistent and don't have the occasional glitches exhibited by the many templates that implement data functions.

I just had a look at Wikipedia:Cascading content and I would prefer to see some support for the proposal before bothering to investigate what moving {{extract}} would entail. Has there been a discussion about it, or has anyone suggested it would be desirable? I often see {{main}} used in summary sections and I do not recall seeing a case where the summary was the same as the lead of the main article. The summaries I recall are usually quite a bit shorter than the main lead. See Anarchism for an extreme example, including cases of multiple main articles. Also, editors often rebel against attempts to require a rigid structure. Are you sure this is going to work? Let's not debate the merits here, but I would appreciate it if you would let me know when a discussion occurs so I can follow it. Johnuniq (talk) 22:26, 3 August 2017 (UTC)

@Johnuniq: Thanks for your thoughtful reply! I've been testing the "cascading content" technique in the Spanish Wikipedia and just now decided to try to spread it here. In the Spanish Wikipedia it got consistent support, but not much has been done yet, partly because it's rather new. Anyway, I understand your argument and I agree: I will finish the essay and start a discussion about the technique using some other template name (maybe Template:Introduction) and if it gathers some support I'll contact you again. Cheers! --Felipe (talk) 00:15, 4 August 2017 (UTC)

Editor of the Week

		Editor of the Week
		Your ongoing efforts to improve the encyclopedia have not gone unnoticed: You have been selected as Editor of the Week in recognition of your continuous collaboration. Thank you for the great contributions! (courtesy of the Wikipedia Editor Retention Project)

User:Buster7 submitted the following nomination for Editor of the Week:

I nominate Johnuniq, a continuously editing collaborator on a wide variety of articles, (active since 2009) to be an Editor of the Week. He always uses the edit summary to explain his actions and is extremely helpful in technical matters even extending to Interwiki cooperation. He supports those who do significant content development and realizes that it's getting along with other editors and engaging with their comments that is the hard part of Wikipedia editing. He is active at WP:ANI and at the Village Pump.

You can copy the following text to your user page to display a user box proclaiming your selection as Editor of the Week:

{{User:UBX/EoTWBox}}

Thanks again for your efforts! Lepricavark (talk) 20:42, 6 August 2017 (UTC)