MediaWiki talk:Captcha-addurl-whitelist

Proposed websites for white list

Please add any suggestions for websites that could be "white listed" i.e. exempted from the CAPTCHA to make it easier for editors to add references to MEpedia entries. An administrator will review suggestions and add to the whitelist as appropriate. Good candidates for addition to the whitelist will be high-quality, consistently reliable sources; the CAPTCHA is in place to prevent "ref spam" of unreliable sources.

Any explanation / comment must go after a # symbol, e.g. ac.uk # All UK universities

Updated Apr 16 2019

Can you add the extra ones requested User:Kmdenmark - notjusttired (talk) 16:27, 16 April 2019 (EDT)

Done! Kmdenmark (talk) 17:21, 16 April 2019 (EDT)

A big, big thank you to you both. It makes editing infinitely easier! Canele (talk) 17:26, 16 April 2019 (EDT)

User:Notjusttired I think the comments/descriptions must be after a single # rather than using round brackets, eg

ac.uk # all UK universities

Sorry for introducing this error. I've changed the instructions above. At the moment, I'm not seeing the Captcha message when signed out so I think this might be why. notjusttired (talk) 13:28, 17 April 2019 (EDT)

My fault too. What do you think, possibly the explanations could just be deleted when the URLs are pasted in, to save the formatting issue? I mainly was adding a comment when I thought it might not be obvious why something I was suggesting would be a valid addition to the whitelist, but once an admin has reviewed to add it to the page, that's probably sufficient screening and the explanations for inclusion may not be necessary to mention on the whitelist itself? Canele (talk) 18:17, 17 April 2019 (EDT)

Thanks for your work User:Canele - I'm happy for explanations to be left out but having just tested it the Captcha is back for anonymous users. It's still not right when signed in, for instance I just got a Captcha adding only this link for pubmed. Could you try using a regular expression just for pubmed, and see if that helps? I think it should be possible to test via Talk pages. Example given in ConfirmEdit documentation

I removed the explanations following the urls. Let me know if it helped. Kmdenmark (talk) 11:49, 23 April 2019 (EDT)

Yes it worked User:Kmdenmark - anonymous edits now get the Captcha again notjusttired (talk) 17:38, 23 April 2019 (EDT)

Great!Kmdenmark (talk) 17:48, 23 April 2019 (EDT)

ncbi\.nlm\.nih\.gov

notjusttired (talk) 17:17, 22 April 2019 (EDT)

notjusttired (talk) 17:17, 22 April 2019 (EDT)

Apologies in advance if I've done this wrong, this is likely above my tech skill level but is this what you wanted me to try? (I imagine you'll need to look in the source code to see if I formatted it right) test. It did triggered the CAPTCHA when I tried to save.

One thing I can tell you, I was getting the CAPTCHA for PubMed even before the comments were introduced, which is why I proposed it again even tho it turned out it was already on the list. Unfortunately I can't recall whether it had been successfully CAPTCHA exempt at one point and then stopped being, or if it never worked (I didn't keep close track of what edits prompted the CAPTCHA, mainly just felt happy any didn't!) My non-tech-person idea would be to strip all the comments out and then test to see what bugs remain? But no doubt others are more qualified to tackle than I. Canele (talk) 19:41, 22 April 2019 (EDT)

Could you add just ncbi\.nlm\.nih\.gov as a separate line? It's a form of computer code known as a 'regular expression' (regex) and just means the same as ncbi.nlm.nih.gov ? If that still doesn't work then it would be something for developers to look into. notjusttired (talk) 09:03, 23 April 2019 (EDT)

Added ncbi\.nlm\.nih\.gov per request. Let me know if it needs tweaking. Kmdenmark (talk) 17:48, 23 April 2019 (EDT)

Blacklist?

If these is a blacklist please add to avoid incorrect image links

• images.google.com
• images.google.co.uk

notjusttired (talk) 07:19, 21 January 2019 (EST)

I had a similar thought, User:Notjusttired. As far as this white list is concerned, I think it only works to turn off the CAPTCHA, so I think it would be a different operation to block sites altogether (if that's possible), but, it had likewise occurred to me there may be some links we should intentionally exclude from the white list, i.e. leave the CAPTCHA in place for. I'd also suggest

• Wikipedia
• WebMD

As with images.google.com, there will be instances where it makes sense to add these (BTW: this search tool is great for finding the handful of Google Images that are free use: https://search.creativecommons.org/ ), but on the whole they are not reliable sources of scientific information (Wikipedia because anyone can edit at any time; WebMD because of pay for play issues--see CBS article, Vox article) so I feel like we may not want to make it easier to add those. Canele (talk) 13:49, 21 January 2019 (EST)

I've recently realized that certain full text links always break over time since the link comes from a Content Delivery Network (CDN) - I would like to blacklist the websites that do this and then it doesn't accidentally happen again. Example: A case of pervasive refusal syndrome: A diagnostic conundrum - link I bookmarked (now broken) vs link working today - this article is also on researchgate which won't change the links - so s3.amazonaws.com could be blacklisted to avoid a link that will change later (although even using an Abstract would be better than s full text link that would break later). Other suggestions to blacklist would be shortened links that redirect eg bit.ly or shortened Amazon links (which could contain affiliate codes too, earning money to unknown people). Thoughts?

Secondly, blocking links to images.google.com and www.google.com or goo.gl shortlinks doesn't prevent images bring found there - it just means the person must properly click through to the proper link instead of accidentally linking to a search results page. E.g. for a WikiMedia thyroid image 3 different links could be used - Google results link - should be blacklisted or better Google result image - or permanent link to WikiMedia - same picture

Would a blacklist message including a link be possible, to explain why? notjusttired (talk) 17:58, 22 April 2019 (EDT)

Seems like a very reasonable proposal, I'm just not aware of a means to implement it. Can say Wikipedia doesn't have any such mechanism but that doesn't mean it doesn't exist in MediaWiki software somewhere. Canele (talk) 21:45, 22 April 2019 (EDT)

I thought the Blacklist option came with ConfirmEdit (which manages the Whitelist) - but it turns out it doesn't and another extension should be needed. I don't think it's worth it so going to leave this alone. notjusttired (talk) 09:13, 23 April 2019 (EDT)

Regex for Whitelist

content hidden

It looks like Regex is needed rather than just the domain names, hopefully this can be tested to check the format is right for a few addresses. notjusttired (talk) 16:30, 20 March 2019 (EDT) /^(.)*\.nih\.gov$ /^(.)*\.cdc\.gov$ /^(.)*\.ac\.uk$ /^(.)*\.cdc\.edu$ /^(.)*\.phoenixrising\.me$ /^(.)*\.s4me\.info$ /^(.)*\.dx\.doi\.org$ /^(.)*\.springer\.com$ /^(.)*\.wiley\.com$ Hey User:Notjusttired, curious where you learned this about Regex? From the example originally listed on the Suggestions page, it looks like URLs suffice: https://www.wikilawschool.net/wiki/MediaWiki:Captcha-addurl-whitelist but that’s really all I know about the subject so I’d be grateful to hear any more info you have! Canele (talk) 17:46, 20 March 2019 (EDT) [User:Canele] I found it described as Regex in the Confirm Edit extension configuration info. The open to turn off the Captcha for certain groups of users might be better? Perhaps based on a number of previous edits rather than account age? Notjusttired 20:24, 20 March 2019‎ I see--well this is probably above my technical wherewithal (I'm not seeing that string but I definitely could be overlooking it/something that should make obvious to me that's what's needed). In any case I agree a skipcatpcha user group, as suggested by that page, would be great; and even better if it could be set to kick in automatically at a certain number of edits, so we wouldn't have to find someone with admin rights each time a new contributor reaches the threshold. The main difficulty I see is that getting it set up does look a little bit involved and we haven't even been able to find an admin to help with this. (Also, in a totally ideal world, I'd love it if the completely reliable sites were whitelisted even for brand-new contributors. Really don't want to do anything to make it harder for someone to help!) Do you have any suggestions, Njt, about who we could contact for help with any of the above (off-wiki, if necessary, since I have the impression folks may not be seeing this)? Canele (talk) 23:58, 20 March 2019 (EDT) (I should add, I'm mindful people may not be on here because they just don't have the bandwidth, which is totally fair; don't want to assign anyone work. Rather, it's that I'd be glad to take on the task of fiddling around until I figure out the whitelist, for instance, but I don't have the right perms for that and I don't know who I should ask about helping.) Canele (talk) 00:02, 21 March 2019 (EDT) A piece of good news: if you look at the diffs from the example site that link gives, it looks as though it only requires the URLs, no extra formatting. You can see someone adding a URL here, for instance. (If you go through the history a bit, you can see in fact just how it was formatted; that here for instance, the https:// prefix had to be removed for the site to be whitelisted correctly.) If same is true here, that's not so bad. Just need the perm to access the page... Canele (talk) 00:07, 21 March 2019 (EDT)

Resolved below. Canele (talk) 15:07, 12 April 2019 (EDT)

Admin perms

Hey User:Kmdenmark, I saw you were recently added to the admins group. May I trouble you to solve a mystery for me: when you navigate to the entry associated with this discussion page (MediaWiki:Captcha-addurl-whitelist), do you have the option to edit it?
If yes, I'd be eternally grateful if you had time to add some of proposed URLs as you see fit (I believe you can just copy them in, just remove the bullet point), but I understand if you don't have time for it. Even just confirming whether it just requires admin access to edit or if that doesn't suffice would be so helpful, if you can. Thank you! Canele (talk) 22:01, 31 March 2019 (EDT)

Yes User:Canele, I am given the option to edit it. However, a warning comes up: "Warning: You are editing a page that is used to provide interface text for the software. Changes to this page will affect the appearance of the user interface for other users on this wiki."

I can try to add the proposed URLs and if it messes up other things, I can just delete my changes. Let me know in detail what you want done and we can go from there. Kmdenmark (talk) 14:39, 1 April 2019 (EDT)

Oh that's great to know, thank you so much Kmdenmark! I think it'd be great to try, say, just adding cdc.gov and nih.gov and then we can try making an edit with a ref to one of those sites, to test to see if it messes any thing up, first of all, and then secondly if it actually does work to disable the CAPTCHA. (And then if it does works, it'd be fantastic if you could add as many of the suggested sites as you consider reliable--my main goal here is to make it as easy as possible for people to add solid scientific references to MEpedia!)

Here is an example of how to format it (sourced via the Dev Queue at MEpedia:Suggested tasks): https://www.wikilawschool.net/wiki/MediaWiki:Captcha-addurl-whitelist

Thank you so much!! Canele (talk) 16:00, 1 April 2019 (EDT)

I copied and pasted the message from the other wiki and added the two suggested sites. Crossing my fingers that all is well. Kmdenmark (talk) 16:20, 1 April 2019 (EDT)

Oh my gosh, it worked: no CAPTCHA when I added this CDC ref! (And I didn't even think about how fitting it is that I happened to be working on a page on accessibility!)

Wonderful, thank you so much! Yes, any sites you consider trustworthy and therefore fit to add would be terrific. And I can work on tidying up the suggestions list (removing sites as they're added) so it's at least a little less of a hassle for you! Thank you!! Canele (talk) 16:32, 1 April 2019 (EDT)

Oops forgot to ping you, Kmdenmark! Canele (talk) 16:32, 1 April 2019 (EDT)

Canele, I transferred all the sites listed on the talk page. Let me know if you find any problems. Kmdenmark (talk) 17:11, 1 April 2019 (EDT)

A million thanks! Oh my gosh I'm so excited for how much easier editing will be. Thank you again!! Canele (talk) 17:19, 1 April 2019 (EDT)