Wikipedia talk:Community health initiative on English Wikipedia/Interaction Timeline/Archive

Welcome! We have several open discussion questions on the article page that we'd appreciate to hear your thoughts on so we can to start making some decisions for our next round of designs. — Trevor Bolliger, WMF Product Manager 🗨 18:52, 3 October 2017 (UTC)

Ping notifications

Under "Other than edits, what information should be included?", should ping notifications be part of this? Certainly, pings may be used to harass editors. However, this might not show up under The 'Interaction Timeline' concept because it "...would only display the edits made on pages where both users have edited."
 — Berean Hunter (talk) 21:55, 3 October 2017 (UTC)

@Berean Hunter: The simplest version of the timeline concept would only show edits on mutually edited pages, but if the tool is more powerful by adding in other actions (such as pings, thanks, emails) I think it should be considered. Adding in less public actions raises the question of privacy, in which case we will need to consider permissioning this tool (or certain functionality within) to certain usergroups. — Trevor Bolliger, WMF Product Manager 🗨 23:26, 3 October 2017 (UTC)

Other than emails, there's no private information here. There is a thanks log already, and the software prompts when thanking specify that the action is public. I'm not aware of a notifications log (there should be one, really) but any edit which generates a ping is visible when the ping is delivered. Permissions shouldn't be an issue here. If permissioning/private information is of concern, we already have arbs and checkusers trusted with that sort of thing. Ivanvector (^Talk/_Edits) 13:01, 11 October 2017 (UTC)

Harassment through notifications

Regarding "thanks" notifications, I've seen a few complaints lately where someone is spammed with constant "thanks" notifications (example 1 – permalink; example 2 – permalink). These are logged by MediaWiki, but it's difficult to tell the exact circumstances of what's going on. Another complaint I've seen is that someone uses "ping" notifications too often, perhaps as a method to annoy someone. If we knew how often someone had used the notification system, it would help. Also, there are cases where someone may have pinged another editor to solicit assistance in a dispute. This is frequently objected to by other editors as canvassing and/or bullying. If someone has a habit of canvassing the same people, who then gang up on the outnumbered editor, this can be an example of tag team edit warring. NinjaRobotPirate (talk) 22:04, 3 October 2017 (UTC)

@NinjaRobotPirate: Great points, thank you. WP:Harassment states that "Harassment is a pattern of repeated offensive behavior that appears to a reasonable observer to intentionally target a specific person or persons" and I believe that repeatedly notifying or calling in croonies fits this definition. All this evidence adds up but can be extremely time consuming to manually gather and collate. I think the tool should show as much information as is necessary to tell the full story. We can also build in filters/toggles to disable certain types of information from showing in the timeline, as not all cases are identical. — Trevor Bolliger, WMF Product Manager 🗨 23:26, 3 October 2017 (UTC)

A series of checkboxes where you could select what you wanted displayed would be very useful, especially if it could track interactions in logs, such as special:log/thanks and, possibly, special:log/block (for when someone says an admin is harassing them). NinjaRobotPirate (talk) 00:32, 4 October 2017 (UTC)

As I manually built some of these timelines, I came to a preliminary conclusion that this tool will only be useful if there are filtering options, so I expect we'll create a panel of options (potentially like on RecentChanges) to control what is and what is not displayed. — Trevor Bolliger, WMF Product Manager 🗨 19:06, 4 October 2017 (UTC)

Off topic: Sockpuppet tool needed

Although this is essentially off-topic - and can be moved with my permission to wherever it would be more appropriate -- it was brought to mind by the section on using the Timeline tool for searching out sockpuppets. I've been asking for a tool such as I'm about to describe for a while now, but it never seems to get any traction.

We have tools which will compare the contributions of multiple editors to see what their overlap is, but we also need a tool which will do the inverse: take multiple articles and spit out the editors those articles have in common. Most of the returns will not be sockpuppets, of course -- especially when the articles are in a single subject area but these can be pretty easily eliminated, and what's left are editors that it's worthwhile investigating.

So, my apologies for the off-topic post, but it would be great to have a tool such as that. Beyond My Ken (talk) 01:18, 4 October 2017 (UTC)

Give me a few days and I might be able to cobble something together. MER-C 09:57, 4 October 2017 (UTC)

I've also made a start on this, though it's not exactly clear when I might make a finish. GoldenRing (talk) 16:09, 4 October 2017 (UTC)

If anyone would like either of our developers to help with code review or input, please ping David Barratt or Dayllan Maza on IRC or Phabricator. We're happy to lend a helping hand. (And in general, I think this is a clever tool, but yes it is slightly out of the scope of this project.) — Trevor Bolliger, WMF Product Manager 🗨 19:08, 4 October 2017 (UTC)

Thanks for the responses, they are much appreciated! Beyond My Ken (talk) 20:49, 4 October 2017 (UTC)

@Beyond My Ken: First working prototype: https://wikipediatools.appspot.com/editorintersection.jsp (commit). There are a couple of things I want to add, including getting the list of articles from a category and collapsible article lists. The limits are in place so that a result is returned reasonably quickly, I intend to have an offline version that will be able to analyze more articles and revisions. Let me know if you have any suggestions. MER-C 06:23, 6 October 2017 (UTC)

Thanks very much, I'll take a look. Beyond My Ken (talk) 00:40, 7 October 2017 (UTC)

Hi guys, good work on this important matter. Want to ask (by giving a scenario) would this tool also cover all edits by a editor even if they did a username change (of their original account -to possibly bypass detection) and then socked using multiple accounts?Resnjari (talk) 04:46, 17 October 2017 (UTC)

@Beyond My Ken:, @MER-C:, @GoldenRing:, @TBolliger (WMF): forgot to ask, can this tool be used in other non-English Wikipedia projects (which have problematic editors socking but few administrators etc to keep track)? If its use can be more widespread it would be of great assistance to those wiki projects. Best.Resnjari (talk) 05:36, 18 October 2017 (UTC)

Username change -- yes, if done through the normal channels.

Other wikis -- easy, I just need to add the text box (likely sometime this weekend). MER-C 12:51, 19 October 2017 (UTC)

@Resnjari: done. MER-C 03:55, 21 October 2017 (UTC)

@MER-C:, thank you for the hard work on this important wiki tool. Much appreciated. Best.Resnjari (talk) 06:30, 21 October 2017 (UTC)

The ability to group 2 editors as one, versus a second editor

Let's take the following scenario: User:Banana claims that User:Apple and User:GoldenDelicious are likely sockpuppets who are both harassing him. In this case, we need to take all edits by either of User:Apple and User:GoldenDelicious, which are to the same page as any edit of User:Banana. עוד מישהו Od Mishehu 04:01, 4 October 2017 (UTC)

Oh, that's a smart way to solve for the multiple users in the tool while retaining the two-column layout. We could use other design treatments (color, icons, identations) to show if the edit/action was Apple or GoldenDelicious. — Trevor Bolliger, WMF Product Manager 🗨 19:10, 4 October 2017 (UTC)

a color highlight box around the name(s) would be slick Nathanielfirst 10:57, 9 October 2017 (UTC)

I was thinking about this too, multiple groups who which serve as supersets for one and more number of editors will be helpful. Pretty sure, it'd lead to some technical overhead + visual mess, but we don't know until we try, I think. --QEDK (愛 • 海) 18:35, 31 October 2017 (UTC)

Email privacy and harassment investigations

One issue which was brought up is emails. While I certainly wouldn't want all admins to be able to see all my email interaction with oth er users, I should certainly be able to:

1. See all my own sent and received emails, including which user was on the other side;
2. Permit all admins to see interactions with individual user of my choice.

עוד מישהו Od Mishehu 04:20, 4 October 2017 (UTC)

I'll have to do some digging to see if logging direct emails has been discussed before, but it's definitely worth looking into. If we build this into preferences, we'd also need to consider logging when the preference changes so a user wouldn't toggle the preference off to send a malicious email then back on to over their tracks. — Trevor Bolliger, WMF Product Manager 🗨 19:13, 4 October 2017 (UTC)

Random comments

I prefer variation 3 to variation 2, though I can't quite nail down why - it may just be aesthetic. Having the time between actions is helpful - it avoids the mental arithmetic of figuring out whether actions are widely spaced. Adding more than two users will be useful, especially in sockpuppet investigation use cases, though it's hard to see how this would be presented visually in variation 3. I generally prefer formats where I don't have to open each diff in a new tab, though loading all of them in one go may have performance implications (it's deliberately hard to do through the API for this reason). Presenting the edit summary in the timeline would be helpful, I think, though the bulk of them may space the timeline out enough that it makes the tool less usable overall. Consider putting extra information in mouse-over popups, though you'd have to consider the accessibility implications. Sorry for an unorganised grab-bag of thoughts. GoldenRing (talk) 06:50, 4 October 2017 (UTC)

@GoldenRing: The grab-bag of comments is totally fine, and actually helpful! In our next design iteration we will address a timline for 5 users, which should help us winnow our options. As for the performance of loading the diffs — I assumed they would all be lazy loaded on demand, which should help. Lucikly edit summaries have maximum lengths, and if we can parse the section title (as seen in the design) then most wouldn't take too much space. — Trevor Bolliger, WMF Product Manager 🗨 19:18, 4 October 2017 (UTC)

Real-time timeline

One issue with both of these views is the fact that we don't see the timeline to scale - that is, that incidents 2 minutes apart aren't twice as close as incidents which are 4 vminutes apart. We neeed some means of a to-scale view, with the ability to zoom out to get an overviw of the times when there were ineractions, and zoom in at dense times. עוד מישהו Od Mishehu 04:30, 6 October 2017 (UTC)

@Od Mishehu: I agree that there needs to be some visual indication(s) to show areas of high interactivity but I'm not sure a true scale would be all that useful — is there really a significant difference between a 4 minute delay and a 2 minute delay? — Trevor Bolliger, WMF Product Manager 🗨 16:17, 6 October 2017 (UTC)

There is certainly a diference between a 5 minute difference and a 10 minute one. Maybe the 2 vs. 4 is closer than would actually be relevant, but if User:Banana does a series of edits in 3 minutes, and User:Apple or User:GoldenDelicious starts reverting him 5 minutes after he started, that would certainly be relevant. עוד מישהו Od Mishehu 18:07, 7 October 2017 (UTC)

You could try a log(interval) scale. Alsee (talk) 09:37, 9 October 2017 (UTC)

A log scale is a great idea, or we can always build this as a toggle — "highlight interactions that occur within N minutes". — Trevor Bolliger, WMF Product Manager 🗨 15:27, 9 October 2017 (UTC)

Regarding 'What Information is important to show?'

Should we build in-line hide/show diffs? Or does a link to the diff suffice? imo yes build them, it suffices but why not build them

Should multiple successive edits by the same user be collapsed into a single block? imo yes

When one user sends a direct email to the other via Special:EmailUser? (not currently public information) imo, emails are a part of life, screwing someone's work up / time spent / and damaging the encyclopedia should be the focus of the tool... maybe admins need email info as extra evidence for a decision, but it is private, so both parties would have to give consent

Cross-wiki activity? imo, should be included

Is it important to show aggregated statistics or data about the users and their interactions? imo, all info you can get other than private emailing, is useful to help judge, right? so yes...

Would it be helpful if the tool allowed users to annotate the history, or to provide notes? imo yes, with text, and the extra venue for harrassement just closes the case faster as it's under a microscope at that point

Edits Apple and Banana make elsewhere, outside of their direct interaction. imo, yes, collapsible or optional

Nathanielfirst 11:25, 9 October 2017 (UTC)

Speaking as a very frequent user of Writ Keeper's Common History script, inline diffs are incredibly useful for investigation (indeed, I often forget that there's another way of looking at diffs...). I would say that inline diffs would massively improve the functionality of this tool - which, incidentally, looks exceedingly useful. Yunshui ^雲_水 20:49, 9 October 2017 (UTC)

Thanks, Nathanielfirst and Yunshui. (And thanks for sharing the Common History script, I'm a big fan.) I agree that with an investigative tool like this that "more is more" — the difficulties will be how do we gradually prioritize and add functionality without overwhelming the UI. We'll likely build a simple version that shows a straightforward timeline, and users can fine-tune their queries based on the case. — Trevor Bolliger, WMF Product Manager 🗨 17:47, 10 October 2017 (UTC)

Being able to hide/show diffs inline would be really nice. Just don't overlook that some diffs can be comparable in size to a large article.

Grouping consecutive edits by the same user is an interesting idea. We would need to be able to click-to-show all edits in that grouping. There's little point in merging two lines into one, so it shouldn't kick in unless it's at minimum [show 3 edits by same user].

Showing cross-wiki activity is a novel idea. We normally stay very focused on "local jurisdiction", and most users don't cross-wiki a lot. I don't think this functionality would be significant very often. It's a neat option if you can easily include it, but it's probably low-priority.

Regarding annotations: Diffs and other evidence generally need to be pasted to a talk page anyway. I think annotations would raise complexity and content issues for little gain.

I doubt there would be much value in showing "Thanks".

Things like account_age / editcount / number_of_pages_both_users_edited is valuable context. Two long term users will have a ton of incidental overlap. Two low editcount users won't overlap on many pages by accident. Data-items that only appear once in the header area are very cheap to include. It easily scrolls off the top, and the real issue is the size of each edit-event.

Adding color or other indication for reverts is helpful. It's great to instantly zero-in on revert wars and stalk-reverts.

Regarding showing "Edits by other users in between Apple and Banana edits" - isn't that reinventing the page-history link? Alsee (talk) 17:33, 11 October 2017 (UTC)

@Alsee: Thanks for the comments — there's a lot of open ideas and we appreciate hearing about all of them. We'll likely be moving forward with in-line diffs. And good point about mega-diffs. We'll probably want a collapse button/link that's always visible on the screen (stickied to the top of the browser and follows you as you scroll.)

Cross-wiki activity will not be an initial feature — we'll add it later. It's not always pertinent for single wiki investigations, but Stewards and others may find this useful for better understanding cross-wiki harassment. If we build it, it'll probably be an optional toggle, same for Thanks.

For the statistics. You're right that several long-tenured users will likely have many incidental edits in common, which is why I think the date range parameters are so important. I think if date filters are used, the statistics box should show both lifelong data and data just for the selected date range. (E.g. 500 common pages since 2001, 8 common pages from 1 July 2017 to 2 July 2017.)

I also agree that showing edits (or even a visual indicator that a page was edited by other users) is overkill. Seeing the diffs and timestamps will accomplish most of this, and the link to page history will cover the rest. — Trevor Bolliger, WMF Product Manager 🗨 18:50, 12 October 2017 (UTC)

• I think it would be great if the investigating admin is provided with the ability to control the flow and type of information they want displayed on the timeline. Particularly, filters might be useful. E.g. end date (to narrow down the amount of edits shown on the results if diff info has to be shown in full), type of pages (Wikipedia:, Articles for edit wars, Talk:, User talk:, etc.), whether abuse filter was triggered, the option to either display all diffs, or display those that are in non-article space pages only, or hide all for that particular query. The other thing I might suggest is to use small symbols to label the type of page that was edited. Harassment can be in all forms and shapes, and so it is arguably difficult to have a fixed formula of what information is important and what isn't. Providing filters allows the investigating admin to decide what is important, and then narrow down and pinpoint where and what type of offending behavior is occuring between users. Of course, there should be a warning that filters do not contain all information (and what information has been hidden if it is obvious), and that anyone using the tool should always start wide first and that should be the default search setting. - Mailer Diablo 23:51, 12 October 2017 (UTC)
  • Harassment can be in all forms and shapes, and so it is arguably difficult to have a fixed formula of what information is important and what isn't. Spot-on, I agree 100%. Great point that the tool should make it clear when items are filtered out, so decisions are not made on incomplete stories (especially if you can share a query with other users.) — Trevor Bolliger, WMF Product Manager (t) 00:34, 13 October 2017 (UTC)

Information density

One of the most obvious and most cited reasons people hate Flow is because of the horribly low information density. Flow has about 40% the density of Talk pages. I know, you've got piles of books and research and design specialists saying whitespace is good. But this isn't light entertainment reading, this is a workplace and the interaction analyzer is a tool. For whatever reason editors can handle higher density than standard "webdesign advice", we're used to higher density, we expect higher density.

When I looked at the two mockups, before I even read any of the text or looked at anything you put on the page, the only thought going through my head was "Noooooo..... I wanna keep the old analyzer!" The density appears to be something like 30% or 40% of the current analyzer. There's almost nothing on the page between all the whitespace.

It took me a while to be able to get past all of that whitespace and actually examine the limited content in the mockups. It seems to show a lot less than what the current analyzer shows. Most importantly, the critical links appear to be missing. Are they merely invisible because they're supposed to be accessible via hover or click effect? Alsee (talk) 21:13, 10 October 2017 (UTC)

@Alsee: I've been a MediaWiki product manager for seven years now, I understand the importance of information density. :) And I wholeheartedly believe that information density and a low cognitive load are not mutually exclusive. Long story short — there's going to be less whitespace in the final version, but some tools and UI elements will likely be collapsible and closed by default, or accomplished with icons/colors/etc. We'll decide the final sizing/spacing after we get the layout sorted.

These first draft wireframes hopefully convey the concept of what we're trying to accomplish — showing a linear chronologic story of the interactions between two users so a conduct dispute can more efficently and accurately be evaluated. Do you think this Timeline concept would help you better understand a conduct dispute? Is there anything it fails to do, or is a sequential understanding of events not as important as we believe it to be?

And let me be clear: we are not changing the existing analyzer tool. We're going to build a new tool, which will hopefully acomplish some things the existing tool does not. (That said, we're open to fixing some bugs in the analyzer if there are any you'd like to report.) — Trevor Bolliger, WMF Product Manager 🗨 22:24, 10 October 2017 (UTC)

Trevor Bolliger, thanx. Sorry if I went overboard, chuckle. Active/collapsible elements definitely allow opportunities to improve information density and cognitive load at the same time. It would be great if you penciled this in as a late stage design step. Explicitly calculate the ratio of how many edits are visible between the two tools. (i.e. links may be hidden behind an active element, but how many non-collapsed edits are fully visible.) I tried to zoom the mockup image to the same font size as the current analyizer. On my widescreen monitor the current analyzer showed about 28 rows of edit-events, with a full set of links, info, and long edit summaries. The mockup fit maybe 8.5 rows of edit events... although the mockup was hindered by being a narrow-width image. 8.5/28 = 30%. The new tool is obviously going to have differences from the current tool, but that's a good computation to check. Alsee (talk) 18:15, 11 October 2017 (UTC)

I can't guarantee that it'll be a 1:1 ratio between the tools — but I can guarantee that we're not adding in any design elements just because they're shiny or pretty. Our ultimate measure of success will be if this tool helps users accomplish important tasks. — Trevor Bolliger, WMF Product Manager 🗨 19:02, 12 October 2017 (UTC)

Centralized discussion

If you want more visibility and input, I'd be happy to add this project to {{Centralized discussion}}. Or you could make the edit. That would give a lot more visiblity than the announcement on Village_Pump Miscellaneous. Alsee (talk) 18:31, 11 October 2017 (UTC)

@Alsee: Yes! I would greatly appreciate any awareness we can drive to this page. Thank you :) — Trevor Bolliger, WMF Product Manager 🗨 17:56, 12 October 2017 (UTC)

 Done.

It's lucky that I checked back here. I never received your ping. I heard there was a recent bug with pings, and I guess this was one of the pings that got lost. Alsee (talk) 07:40, 19 October 2017 (UTC)

Trevor Bolliger, central discussion usually links to a talk page section. For consistency I'll create a target section below "Feedback", and point Central Notice to there. Alsee (talk) 05:13, 20 October 2017 (UTC)

@Alsee: Thank you! — Trevor Bolliger, WMF Product Manager (t) 14:40, 20 October 2017 (UTC)

Paid editing tracker

Might there be some way of tracking a very specific metric: new pages created and edited by 2+ new users? I believe this would be very helpful for finding paid editing. Thanks, GAB^gab 21:13, 12 October 2017 (UTC)

@GeneralizationsAreBad: It's technically straight forward, the only question is resources. It appears Rentier's NPP Browser fetches most of the relevant information, and should be able to get the remainder with very few additional API requests. MER-C 13:10, 19 October 2017 (UTC)

I have already started working on this. Not sure when I will find the time to finish it, but it's not too far away. Rentier (talk) 13:23, 19 October 2017 (UTC)

Feedback summary, Oct 16

Hello everybody. I’ve re-read all the comments here and on the talk page on Meta Wiki and wanted to provide a summary of all feedback for the Interaction Timeline so far:

• Mention and Thanks notifications could be shown, as they are publicly visible in the wiki revision history or logs. However, sometimes they may not be helpful so we will want to consider a filter to toggle them on/off.
• We’re very likely going to need filters for any non-essential data we represent in the timeline. These would most likely be a series of checkboxes to toggle certain elements in the timeline on/off.
• The two-column layout could be used to monitor more than two users by bundling the others into one of the columns. (E.g. compare ‘Apples’ against ‘Bananas’+’Carrots’)
• One vertical line (Variation 3) vs two vertical lines (Variation 2) is preferable. No explicit support for Variation 2 yet.
• Showing the explicit time calculation between users is helpful — it avoids mental arithmetic.
• In-line diffs would be preferable to new tabs, but beware megadiffs and slow performance.
• Seeing the edit summary is important, as long as they don’t take up too much space.
• Some information could be hidden under mouse-over, but consider accessibility needs.
• The timeline could be to scale to better illustrate when/how people are overlapping in their interactions — maybe a log(interval) scale.
• We should explore bundling multiple successive edits by one user into a single ‘block’, but will need to make sure there is a way to see the details for each specific edit.
• Showing cross-wiki activity could be helpful, but it would need a filter to toggle on/off.
• Showing aggregated statistics above the timeline is helpful
• It could be helpful to…
  • …annotate the timeline, but there are also some concerns about necessity and complexity.
  • …export results from the tool to wikitext for easy on-wiki reporting.
  • …show edits made by the other users outside of their direct interaction, but would definitely need a filter toggle.
  • …color code or use icons to indicate reverts.
  • …filter certain namespaces out of the timeline.
  • …represent namespaces with different icons.
  • …highlight the different usernames with different colors.
• The UI should warn when a filter/toggle is enabled that is hiding certain information
• This is a utility tool — don’t add too much whitespace. Be cognizant of information density.
• If we built a solution that allow users to specify if they want their emails logged (and visible only to admins) then this data could be shown in the Timeline.

Did I miss anything or misrepresent any feedback? And keep the thoughts coming! — Trevor Bolliger, WMF Product Manager (t) 22:12, 16 October 2017 (UTC)

I believe you missed the date cutoffs mentioned above.

Please also add links to diffs so one can copy the URLs to paste elsewhere. MER-C 13:14, 19 October 2017 (UTC)

Second draft of wireframes

Hello all;

Based on some of the feedback we've received we made some minor adjustments to the wireframes. The biggest change we made was showing how the in-line diff could function. There are still plenty of changes we'd like to make, such as adding in filter options and color coding/iconography for notable elements.

Second draft wireframes

— Trevor Bolliger, WMF Product Manager (t) 22:15, 16 October 2017 (UTC)

Wow, that's a very large image. It seems fine, though. It took me a moment to realize that was a diff open in the center. At least, I'm pretty sure that's what's going on. NinjaRobotPirate (talk) 00:13, 17 October 2017 (UTC)

Does anyone use Intertwined Contributions?

During a conversation with a Wikipedian on Friday about the Interaction Timeline, the user mentioned a tool called Intertwined Contributions. It can be found here: http://tools.wmflabs.org/ptools/intertwined.php

Does anyone use this tool? Does anyone have any thoughts about it? Does it have any advantages or disadvantages to the Timeline tool we are proposing? — Trevor Bolliger, WMF Product Manager (t) 22:16, 16 October 2017 (UTC)

One issue I see here is the fact that the wiki is chosen in a single dropdown (unlike other tools, such as this page search tool, in which the selection is made in 2 textboxes); this means that the dropdown is, by necessity, huge. I think that the field of the wiki should be split into 2 separate dropdowns. עוד מישהו Od Mishehu 04:48, 18 October 2017 (UTC)

Feedback

See the matching page here: Interaction Timeline for full details. In short, Interaction Timeline is a new tool being built by the WMF Anti-Harassment Tools team, similar to the toollabs Editor Interaction Analyzer.[1] Mock-up images and other discussions can be found higher on this page. The team wants feedback from us on what kinds of functionality and design would best serve our needs. The section Discussion Questions lists various questions the team is hoping we will help answer. Alsee (talk) 05:14, 20 October 2017 (UTC)

• It would be helpful to have the tool combine similar diffs by default (and allow granular separation as an option). For instance, if an editor makes five successive content changes to a page, I'm likely better off seeing "editor edited (page)" alongside their five edit summaries and the number of collapsed revisions (five). Click on the number to have the section split out into five different diffs, as the mockup shows now. Another idea: When talk page diffs are successive exchanges, instead of showing them as diffs on opposing sides of the timeline, collapse them into a single, centered node on the timeline, and have the diff expansion highlight syntax to show the adjacent talk page additions contributions of each editor. Similarly, if the issue is a content conflict in mainspace, use something like the mw:Help:Two Column Edit Conflict View and center to the timeline.

It'd also be nice to have some visual representation of volume, e.g., having the timeline nodes smush closer together when the interaction is rapid/intense and more spread out when paced over a longer period of time or punctuated by other editors also participating. I also don't think it's worth separating by dates—there's that one example with 1h25m marked in red, but since WP is running on UTC, the breakdown by calendar date is less relevant, or why not use the same red text to mark the hours between "overnight" interactions (if not already resolved by some other visual representation for intensity). Alternatively, the date can display as a small, centered heading—similar to that of iMessage—without breaking the cadence of the timeline.

On the smaller side, the descriptions become kind of redundant. If the name of the edited page came first (and perhaps were bolded or otherwise designed for distinction), the mainspace/talk/section edited or created would be self-explanatory (sections formatted as "Talk:Example#Discussion" or reformatted to "Talk:Example § Discussion" as is done in hatnotes). I'd also try to reduce vertical whitespace by moving the "diff" button to the outside. I can see how this tool will be useful for zoning out irrelevant diffs when adjudicating a conflict, but in most cases, I'd imagine the discussion is really going to be about three or fewer key diffs (n.b., not my area of admin expertise, but perhaps that makes me a target audience). The most useful part is going to be if/when editors can input diff URLs or otherwise specify the key diffs in an interaction so someone else can see both the core issue and, via the tool, the issue's context without being overwhelmed by minutiae. czar 07:34, 20 October 2017 (UTC)

• Great feedback, thank you @Czar:. I totally agree with the combining of diffs (and uncombining as an option, as successive diffs could be used to write and remove harassing messages) and with exploring alternatives to the two-column diff layout. Talk page edits are usually different than article page edits (additive vs. adding and modifying existing content) so we are also considering representing them differently if needed. Brilliant thought to merge them for both users if they're all successive. I hadn't thought of that but it would definitely cut down on the back-and-forth required.

We're looking into some sort of visual representation of volume for periods of high-interactivity. I'm not convinced that spacing/sizing is the right approach, but I'm 100% open to being wrong. I do agree that calendars are not as important, especially given the global audience of Wikipedia — if a user in Australia and a user in New York are in a dispute, the time between their responses will likely be different than two users in the same time zone. Good suggestion to look at iMessage — that text is there when you need it, nearly invisible when you don't. I hope to have another round of design explorations for both of these points in a few weeks.

Great point about space and redundancy. And yes, we'll do what we can to reduce as much vertical whitespace as possible. We'll find that right balance of cognitive load and information density.

This may be a bit too aspirational, but I think this tool could be the only URL needed for reporting a user conduct dispute to AN/I. Annotations, highlighting, pinning an icon to a specific diff/period of interactions could be used to call out the important evidence from each side. But this is definitely a later stage add-on feature. We're still working on making sure the core concept is useful. — Trevor Bolliger, WMF Product Manager (t) 22:46, 23 October 2017 (UTC)

Good, I'm glad. Those extended features sound very useful and worth adding to the project page's roadmap for others to see. Wanted to add a recommendation that as many parameters as possible be accessible from the URL structure (i.e., not a single hash) so as to prepare against link rot. And before implementing three-way interactions, single-user timelines would be even more appropriate, as I assume (1) there are more abuse cases of individuals with indiscriminate malice (at least via a survey of ANI on enwp) than there are two-way harassment cases, and (2) the tool would be particularly helpful in annotating a single user's edit history across multiple unconnected events. Related to ANI, editors rarely focus on one argument at a time as neatly as in the mockup, so it would be useful to be able to search/filter across edits and edit summaries for specific phrases, e.g., repeated insertion or discussion of a contested phrase over time, or repeated reverts that use the same explanation. Perhaps for the extended roadmap or for discussion elsewhere in the community health initiative, it might even be most helpful to let editors dynamically embed a single diff (with details) or annotated timeline directly inline within a talk page discussion, such as via a collapsible template, which would make the discussions more accessible. Appreciate your work on this czar 23:51, 23 October 2017 (UTC)

• Why not borrow from the existing method in which edits are listed on Special:Recentchanges, but broken up by day and with the added annotations (e.g. 1hr between edits)? We're quite used to these, and would be about the information density we're looking for. Having the timeline on the left will make it easier to incorporate more than two users in a clean manner (distinguish them by color or whatever means is appropriate). I agree edits to the same page without anything intervening should be grouped, one could have a mini-history listing in a single colored box, titled with the page name and username and corresponding to a single point on the timeline. The timeline should be sorted in reverse chronological order. MER-C 08:09, 20 October 2017 (UTC)
  • @MER-C: Thank you for your feedback! In the end, I expect we're going to build a variation of this Timeline that looks more akin to the class 'log' style of Recent Changes and History, especially if we want this to support investigations involving more than two users. For now, our design interviews and research to date has shown that the side-by-side approach helps tell the sequential story, so we're still pursuing that direction. The reason our timeline is chronological is because we feel it's important to understand the 'story' from beginning to the end. (e.g. Apples makes an edit, Bananas reverts it, Bananas yells at Apples, Apples yells back, etc. etc. etc.) I am also pretty sure we'll build in the functionality to sort this list reverse chronological — it's trivial to build and if it makes the Timeline usable for even one more user, then there is no reason why not to. — Trevor Bolliger, WMF Product Manager (t) 22:46, 23 October 2017 (UTC)

3rd Draft Wireframes

Hi all, We made our 3rd round of wireframes. I made sure to incorporate community feedback like reducing the amount of whitespace. I'd love to what you think of the differences between wireframes? We also shortened the text across all of the wireframes.

We are looking for feedback :) One wireframe has summaries that say "summary: an addition" for an example, and for another removes "summary" and just says "an addition." Is this easy to read, does this come across as a summary or should we include the word 'summary'? What do you think of the different kinds of diffs we've included? Does one look better or is more readable than the other?

let me know :)

• 
  Interaction Timeline 3rd Draft WireFrames

--CSinders (WMF) (talk) 21:39, 31 October 2017 (UTC)

Allow IP ranges in user inputs

I agree with limiting the number of user inputs to 2. However, each input should allow an IP range (CIDR subnet notation?) to help with anonymous contributors using dynamic IPv4 or IPv6 addresses. The output should include all activity from each IP range in the same column but also include the specific IP address used to make each edit. 174.226.129.181 (talk) 02:52, 2 November 2017 (UTC)

Yes, I agree this should be supported. I've added it to the list of requirements for the first full version of this tool, as documented in Phabricator task T179607. — Trevor Bolliger, WMF Product Manager (t) 19:09, 2 November 2017 (UTC)

Indirect harassment

Thank you for this tool, it is sure to be useful! One query I have (apologies if this is already covered elsewhere, I have not read all of the discussions) is about indirect harassment? Example: User Carrot dislikes User Tomato, because Tomato disagreed with Carrot in a discussion somewhere. Carrot then proceeds to bad-mouth Tomato in multiple fora, such as, "Well, you can't believe anything Tomato says, their judgment is suspect," or "Tomato is messing up those articles over there, you'll need to watch out for them," etc. These comments are typically placed on pages that Tomato does not frequent: Talkpages on other articles in topic areas where Tomato is not involved, User talk pages of users that Tomato does not interact with, Wikipedia talk pages of policies that Tomato has never edited, or even on subpages within Carrot's own userspace (like if they decide to create an "I hate Tomato" page). Would these kinds of things show up with this tool? Thanks, --Elonka 03:30, 2 November 2017 (UTC)

@Elonka: The concept of "only show pages where both users have edited" is our starting point. In the end I'd like this tool to be powerful enough to cover this exact scenario you bring up. This could be done in a few ways, including showing times when Carrots creates a link to User:Tomato (which would miss times when Carrot just says their username without a link) or by allowing users to input specific pagenames where only one of the two users has edited. (Or we could build both!) Would you have any other ideas on how to include this information in the Timeline?

We'll keep this idea in mind and I'll be sure to capture it in our next round of feedback summary, which I use to prioritize the features we add. Unfortunately due to their complexity these features likely won't be part of the first version we release. — Trevor Bolliger, WMF Product Manager (t) 19:16, 2 November 2017 (UTC)

Thanks for considering this. I think the tool would be most useful if it just did a text search on comments made by Carrot, rather than looking for a username link. Where I have observed this kind of behavior, User Carrot is not trying to bring Tomato's attention directly to the comments, more it's usually been a case of Carrot complaining about Tomato in fora that Carrot sees as their personal space, speaking to users that Carrot perceives as allies. Sometimes it can become of more concern, as the allies pick up the call, then they start badmouthing Tomato, and it spreads virally from there, to the point that Tomato can offer a comment somewhere, and then suddenly get negative reactions from users they've never interacted with! The easiest way to find this (when I've done these searches manually) is to check Carrot's contributions list and look for talkpages where Carrot has had substantial (more than 3 or 4) entries of participation. They are most likely to be article talkpages, or user talkpages. Very rarely it would go as far as edit summaries to articles, like where Carrot might say, "I have to clean up Tomato's mess again." Flagging those summaries would be very powerful. --Elonka 00:06, 19 November 2017 (UTC)

@TBolliger (WMF): I'd like to add that this is a distressing form of harassment. It's difficult to deal with, because if you comment (as the person being discussed), you expose yourself to more of it, and if you don't comment, you have to sit back and watch yourself be undermined. Would it not be relatively easy for the WMF to count how often Carrot mentions Tomato in posts and edit summaries? Carrot wouldn't create a link to User:Tomato in this type of harassment, because the point is to talk about Tomato not to them. (A more difficult form to detect is when other words are used instead of "Tomato", but where everyone on the page knows who is being discussed. This is often done by linking to one of Tomato's posts, e.g. "you know who" with a link.) SarahSV ^(talk) 04:21, 19 November 2017 (UTC)

@Elonka: @SlimVirgin: Thank you for providing more examples and ideas. I hope we can find a solution to make these types of comments easier to find and use in conduct dispute investigations. Abbreviations, allusions, and nicknames make automatic detection incredibly difficult, and in the end will likely still need human verification. We received a proposal for a feature to make auditing a single user's edits easier, which could make the existing manual processes more time efficient. I've also created phab:T181101 to make sure this topic does not get lost in the shuffle. As I think about it more and more, there may need to be a different tool entirely. This is an Interaction Timeline and we may need to consider building an Uninteraction Timeline. If we do, we will want to make it easier to combine the results. — Trevor Bolliger, WMF Product Manager (t) 22:19, 21 November 2017 (UTC)

General feedback

Hello Trevor, and other participants and commentators. Thank you for your work. I have been involved in a number of investigations, now years ago. At that time there was no tool support available, and I can personally attest to the difficulty of determining the timeline of a dispute. I can also attest to the difficulty of sharing a hand-crafted timeline with others in a way that inspires confidence. I have a few comments.

• Though often there are one or two primary disputants involved, many if not most disputes are multi-party, and even for bilateral disputes it is important to show selected actions of uninvolved third parties on a timeline. These may be minor actions like making a cleanup edit or responding to a question, or may be a revert or warning.
• It may be helpful to automatically include edits by others that appear on the disputants' talk pages, since in practice most of these will have some bearing on the matter at hand.
• It is often the case the logs are an important part of the timeline, especially when the actions of an administrator are being questioned. Since these disputes are often the most damaging to the social fabric of the project, it is of particular importance that they be handled in a transparent fashion making the usefulness of a timeline important even though the total number of these cases is smaller.
• It is important to be able to include deleted edits when the tool is operated by an administrator, since some of the most damning evidence can end up being deleted for various reasons.
• In response to your questions, it is rare for disputes to occur across wikis in a way that requires a considered response. We do see, for example, crosswiki linkspam and crosswiki vandalism, most of which is now caught by automated tools. It is rare for contributors to enwp to make substantial contributions elsewhere, and this is especially true for contributors that are working in areas that are conflict-prone.
• I believe it is important to be able to annotate the results in a reasonable way after they are generated, perhaps using other more general-purpose editing tools, as this is necessary to share conclusions with others.

Policy wise, the greatest difficulty posed by stalking/bullying allegations is that the borderline cases require insight into the motives of those involved. There is absolutely nothing wrong with following the contributions of someone who has a history of questionable edits for the purpose of scrutinizing those edits with particular care and cleaning them up, or indeed reverting them, on their merits. Contributors should feel confident that they may do so. I would hope that the ready availability of tools like this does not result in that confidence being undermined.

The Uninvited Co., Inc. 19:29, 3 November 2017 (UTC)

@UninvitedCompany: Thank you for sharing your feedback, it's input like this that helps us make the best possible tools. We're hoping that this Timeline feature can reduce the amount of time to investigate, discuss, and respond to user conduct disputes.

• We're almost certainly going to need a solution for showing more than just two users. We've discussed having two versions: a two-column vertical timeline layout and a more traditional horizontal log layout. We're also exploring a three-column layout, but I have reservations about usability and cognitive load, relative to the log format.
• I hadn't thought about filtering in additional edits (such as all edits on both users' talk pages) but I can see how that could be helpful. We were already planning on building a function to filter out edits from specific pages or namespaces so we can also add in the converse to filter in edits from other users on specified pages.
• Publicly available log entries will be added very soon after our initial version.
• We'll have to plumb up permissioning so only the appropriate users can see deleted revisions and edits to deleted pages, but I agree they are important pieces of information for some disputes. In your opinion (and I'll check with the WMF's legal and security teams about this as well) if both users have edited on a page that was deleted, can we display a message to non-permissioned users along the lines of "User1 and User2 have edited a page that was deleted. Admins can view deleted edits." ?
• Cross-wiki functionality might be helpful for stewards. This functionality is down near the very bottom of our prioritized list of features.
• Annotations will certainly be interesting to work into this. If we're building in user authentication for viewing deleted revisions, then it should be straightforward enough to attribute annotations to usernames. My current thought is to allow users to attach icons (pushpins, stars, etc.) or highlight certain edits/log entries. Comments and discussion should be kept on-wiki or privately in email when appropriate.

I'd hate for this tool to have a chilling effect on good-faith wikignomes. If we're successful, this tool should make it even easier to realize which claims of harassment are baseless. But if this tool brings more harm than good, we'll bury it deep in the woods and work to repair what we've broken.

Have a nice weekend! — Trevor Bolliger, WMF Product Manager (t) 23:01, 3 November 2017 (UTC)

Thank you, Trevor. In response to your question, in my opinion, exposing information about deleted pages would create a possible attack surface we would prefer to avoid. For example, an attacker could use such a tool iteratively to determine authorship data for a controversial deleted page if one author was known. An attacker could also accumulate a reasonably complete list of deleted pages edited by a particular user by using the tool iteratively querying disputes with the target user and each administrator in turn plus certain bots, since nearly all deleted pages are edited by an administrator or a bot at some point prior to deletion. Seems farfetched, but in the past there were far more ambitious campaigns waged to ferret out nonpublic information. The Uninvited Co., Inc. 17:46, 6 November 2017 (UTC)

Yes, it makes sense that someone could cross-reference enough users to determine who created or edited on a deleted page. It's already common knowledge that certain usergroups can see information non-privileged users cannot, so I don't think this feature is necessary. — Trevor Bolliger, WMF Product Manager (t) 21:39, 6 November 2017 (UTC)

Test the alpha version of the Interaction Timeline

Our team is proud to announce that an alpha version of the Interaction Timeline is now available for testing! We wanted to get the tool in your hands so you can let us know if we're headed in the right direction.

The Interaction Timeline can be found at https://tools.wmflabs.org/interaction-timeline/. If you provide two usernames and a wiki, the Interaction Timeline will display a chronologic list of edits made by the two specified users on pages where both users have edited. For example, if User:A edits on Washington, Texas, and Kansas and User:B edits California, Texas, and Kansas the tool would only show their edits on Texas and Kansas.

Known limitations

• Due to current API limitations the tool only retrieves a maximum of 1,000 edits. This is our top priority to resolve. (T179702) The tool works like this:
  1. In the background, the tool retrieves the first 500 edits by both users from the provided start date (or first 500 edits of all time if no start date is provided.)
  2. In the background, the tool then generates a list of commonly-edited pages between the two users, from all time.
  3. On the timeline, the tool only displays the edits to the commonly-edited pages within the 1,000 edits that it already retrieved.
• Performance is not optimized, and there is no loading indicator to show when the tool is still working to generate the results.
• There is no error message when there are no results.
• The tool does not support IP address, only usernames.
• Typing in the date range only allows you to type one character at a time.

Examples to test

1. User:Test-apples and User:Test-bananas on test.wikipedia.org
   • Link to timeline: https://tools.wmflabs.org/interaction-timeline/?wiki=testwiki&user=Test-apples&user=Test-bananas
   • Scenario: Wikihounding. Bananas makes a bad edit, and Apples then follows them around to other pages they've edited.
2. User:Test-carrots and User:Test-durian on test.wikipedia.org
   • Link to query: https://tools.wmflabs.org/interaction-timeline/?wiki=testwiki&user=Test-carrots&user=Test-durian
   • Scenario: "He said, she said" dispute, likely boomerang. Durian criticises Carrots until Carrots snaps back. Durian then cleans up their trail and reports Carrots to ANI.
3. User:Derby pie and User:Sweets lover on test.wikipedia.org
4. User:Tinker toys and Baby rattle on test.wikipedia.org

Discussion topics

• For comparing disputes between two users, how does this two-column concept compare to other existing tools? (e.g. Edit Interaction Analyzer, Intertwined Contributions, Intersect Contributions. Special:Contributions, etc.)
• What information or functionality is missing that would make this tool more useful than your existing tools or workflows?

Thank you, and we sincerely look forward to your thoughts on our work, and your feedback on where to take it next. Trevor Bolliger, WMF Product Manager (t) 21:33, 13 November 2017 (UTC)

Alpha comments

See also: meta:Talk:Community health initiative/Interaction Timeline § Alpha feedback

Is this the right place for comments on the alpha release? Feel free to move my comments if not. First off, I like the interface, but I'm expecting to see more features as development proceeds. Just observing a few things as I go through:

• The guide mentions that the tool retrieves the "first 500" edits if no timeframe is specified. Surely this means the 500 most recent edits? To my mind that would be much more useful than the 500 oldest edits.
• The tool does not currently seem to sense interactions in all namespaces. I compared two accounts that only have interactions in User: space and the tool returned no results; however I compared two other accounts which have interactions only in Talk: and could see the interaction. The "old" editor interaction utility includes all namespaces by default, but also allows filtering by namespace or multiple namespaces.
• Future feature request: since this got me thinking about namespaces, it would be incredibly useful if the tool could compare interactions across many wikis. Harassment frequently spills across multiple projects.
• Only supports accounts, not IPs: I see this is listed at #10 under "what's next" but it should be #1, the thing you fix before you fix anything else. If all a malicious user has to do to avoid detection by this tool is log out, then the tool is worse than useless. If for some reason it's not possible for this tool to include IP edits then just stop working on this and assign your resources elsewhere.
  • Related: it would be a big help if the tool would support IP CIDR ranges. But the existing one doesn't, so anything beyond "works with IPs" is gravy.
• Nitpicking: the username entry field auto-complete is case-sensitive; if I type an account name in the wrong case it doesn't come up in the drop-down.
• Nitpicking some more: if I type "enwiki" in the wiki field, en.wikipedia.org is the last result and I have to scroll to get to it. If I type "english" I get nothing. If I type "wikipedia" every wiki is listed, or so it seems, and enwiki is lost in the middle somewhere. I think it should be easier to find the largest project.

More to come when I have more time to test things out. Overall it looks very promising, notwithstanding my comments about IP addresses, and the team should be proud of what's come so far. Ivanvector (^Talk/_Edits) 21:04, 15 November 2017 (UTC)

Thank you, Ivanvector for your comments and kind words! (Yes, this is the best place to leave your feedback.)

• Because the Timeline shows the edits chronologically from oldest to newest, we grab the oldest 500, not the most recent. This will be a moot point once we get full date range support, currently our top priority. (phab:T179702)
• The tool should show all edits in all Namespaces. There's nothing intentional in the code that removes certain namespaces. I've filed phab:T180647 to investigate and fix. (My first hunch is that the tool might have failed and never showed you any results. Refreshing the page sometimes fixes this. phab:T180072 will show a loading indicator and phab:T180083 will show error messages, which should make it easier to know if the tool is loading, failed, or has no results.)
• Great recommendation, thank you. We're currently focusing on making the Timeline incredibly powerful for investigating two users on one wiki, but if we're successful we'd love to expand it allow for more complex investigations (e.g. 3+ users, multiple wikis, etc.)
• This is exactly the type of prioritization feedback we're looking for. I prioritized IPs lower because I don't see many IPs reported to ANI, but this tool isn't for me — it's for you and other active Wikipedians. I still don't think IP support is more important than loading indicator, error messages, and full date range support (as the tool often feels unusable without them) but I agree that IP support would be more useful than items 4-9 on our list. We'll look into CIDR, but no promises. Our current implementation depends on what the existing MediaWiki APIs can give us in terms of contributions.
• These bugs are very annoying, and I've filed them at phab:T180087 and phab:T180084.

Thanks again, and I'll post an update in a few weeks when we've implemented more improvements. — Trevor Bolliger, WMF Product Manager (t) 23:59, 15 November 2017 (UTC)

Some initial observations:

• If a user doesn't exist, say so. This is semantically different from there being no results.
• Please add the usual user links to the header: Example (talk · contribs · deleted contribs · logs · filter log · block user · block log)
• A request has been filed against the API for range contributions. I've been waiting for this one as well (I was the one who originally requested it on AN).
• The tool uses only ~60% of the horizontal space on my 1920x1080p screen (Firefox 56 on Linux), with the edits themselves taking up only ~50%. Is the extra space going to be used for annotations (or should they be placed under the relevant boxes)?
• Dates should be presented in a country-neutral manner i.e. 2017-12-31.
• I revdeled one of the edit summaries on the edit war example (let's pretend it contains egregious personal attacks). The resulting edit summary displayed on the tool is blank.
• Minor edits should not just be indicated, they should be filterable.

Looks good so far. MER-C 05:02, 16 November 2017 (UTC)

Thanks for the feedback, MER-C. The username selection only accepts valid usernames. We are building in error messages in phab:T180083 which will provide more context why there are no results (tool died vs. no results to show). I've created phab:T180722 to add in the userlinks. Your thoughts on the ticket would be greatly appreciated!

Yes, the whitespace. :) Long story short: this is not the final design. We used Bootstrap to get this alpha built quickly and favored function over form. I underestimated the importance of addressing this issue. In the coming weeks we will be integrating standard Wikimedia OOjs UI elements (phab:T180090) and will make sure the information density is improved and optimized both for standard laptop monitors as well as large desktop monitors. This is also why we're not using YYYY-MM-DD date format yet. (phab:T180725)

As for deleted revisions and pages — we eventually want to display deleted content in the timeline, but will first need to build in account authentication and permissioning. We'll likely get to that in early 2018. (In the #General feedback section above, UninvitedCompany and I previously discussed revdeleted content.)

Great idea about filtering minor edits. We're adding in byte change and minor visual indicators soon, and filters will likely be January. Over the next two weeks we will build a loading indicator, error messages, full date range support, IP user support, and improved information density (via Wikimedia UI components). Our full V1/beta list of requirements can be found at phab:T179607. Is there anything else critical we're missing? — Trevor Bolliger, WMF Product Manager (t) 19:03, 16 November 2017 (UTC)

Apologies if I wasn't clear: something intelligible should be shown if the edit summary was revdeled, blank can easily be confused with no edit summary and the revdel, as pointed out above, is meaningful information whether or not the user of the tool is an admin. The admin stuff is worthwhile doing, but as you said, it's for later.

Instinctively, I would add the user links just above the interaction summary (presumably under the input fields and above the results). Speaking of which, I'd add a little space or an indicator to visually divide the inputs and the results as there is no submit button to do so (it doesn't have to be this obvious). I can easily see the dropdown beneath each user working for a group of users versus another group of users, though. I don't particularly mind which you adopt. MER-C 09:37, 17 November 2017 (UTC)

@MER-C: Thank you for clarifying. I've created phab:T180829 to correct the revdel'ed edit summary issue. I also agree with your point about visually separating the inputs from the Timeline results. We're going to do a visual refresh (with a focus on increasing information density) so we will likely address it at the same time. — Trevor Bolliger, WMF Product Manager (t) 21:42, 17 November 2017 (UTC)

A few comments:

• Similar to Ivanvector above, actually finding enwiki in the list of wikis is frustrating. I get not building bias in, but designing UIs to make the most common usecase easiest is also a thing.
• I'm finding that comparing real editors doesn't produce very useful results; I end up with long strings of edits by one editor that I'm not really interested in. Consider [2] for a random example. Although B!sZ and I cross paths fairly often, I don't think this helps find any of those cases. In fact, it implies that B!sZ hasn't edited WP:AN or WP:ANI since the 6th of October, which I'm pretty sure isn't the case - is this the 500-edits-limit coming into play?
• Per some others above, it seems a bit daft to wrap the page titles and edit summaries while only using about 1/3 of the horizontal space. This is always going to be something that's much higher than it is wide, so I think it makes sense to use all of the horizontal space.
• Dates should be displayed in the user's locale.

GoldenRing (talk) 13:05, 17 November 2017 (UTC)

• @GoldenRing: Thank you for your comments! We just decided on a solution for the wikiname input, which we will fix in the next few weeks. We'll also made the timeline show more than 500 edits, which is a crippling limitation for interactions such as the one you provided.
• I regrettably underestimated the importance of focusing on information density. When we get to phab:T180090 in the next two weeks we'll be sure to optimize for standard laptop screens and large desktop monitors, meaning there won't be any excess whitespace on the sides. Would it be acceptable if dates used the international neutral YYYY-MM-DD format? — Trevor Bolliger, WMF Product Manager (t) 21:42, 17 November 2017 (UTC)

A few tests with me, NinjaRobotPirate, as one of the subjects:

• In the demo, it shows a pretty table where both sides have data.
• In this test of me vs Drmies, only my side has data. Drmies and I have both made a substantial amount of edits to a wide variety of pages, so I figured maybe this was causing trouble.
• In this test of me vs Erik, only Erik's side has data. I restricted it to 2017-01-01, thinking this might help.
• In this test of me vs Erik, both sides have data. It was restricted to 2017-11-01. Still, there should probably be more overlap between us; we're both members of the same WikiProject.
• In this test of me vs Erik, both sides have data, but now it's kind of lopsided toward me. For example, Drive (2011 film) shows up on my side but not his. This raises the question of why it's even there. This time I restricted it to 2017-11-10 (the past week). I think maybe a large volume of edits and/or long history (10+ years of editing) might be troublesome for the alpha version? NinjaRobotPirate (talk) 23:02, 17 November 2017 (UTC)

• @NinjaRobotPirate: I think you found a bug! 🕷 While looking into the example of you and Erik from January 1, 2017, a few of your edits should definitely show. You made your 500th edit on January 12, 2017 so it should show these edits on the timeline, such as this or this revision. Sure enough, if I add an end date to your example, both of your edits show up. Damn. We'll have to fix this, we need this tool to be reliable. (But thanks for providing such clear examples!) — Trevor Bolliger, WMF Product Manager (t) 00:08, 18 November 2017 (UTC)

Feedback summary, Nov 22

Hello all! Here is an abridged summary of all the feedback received since I posted the last update:

• Common feedback, and our top prioritizes to address:
  • The Information density is bad, there is too much whitespace, and the tool requires too much scrolling. This will be one of our top priorities to address, and we plan to show progress in the next two weeks. (phab:T180090)
  • The limit of 500 edits per user makes the tool virtually unusable for many scenarios. (phab:T179702)
  • The username field should be case agnostic. (phab:T180084)
  • The wiki field is obnoxious. It could autofill based on where ever the users interact the most, or the logic should be improved to suggest enwiki first. (phab:T180087)

• Feedback about prioritization:
  • Support for IP users is needed. (phab:T180653)
  • In-line diffs are needed. (phab:T180417)
  • Display error messages if there are no interactions to display. Right now the tool just looks broken. (phab:T180083 and we'll build a loading indicator in phab:T180072 in the next few weeks)

• Other new feature requests:
  • Make the usernames sticky as you scroll. (phab:T180245)
  • Group edits by timespan to make it easier to identify incidents of rapid reply
  • Combine successive diffs by the same user, but allow the option to see each edit individually.
  • Build a log style of this tool
  • Support IP CIDR ranges
  • Minor edits should be indicated (phab:T180415) and should be filterable.
  • We will eventually need support for more than 2 users
  • We may want a filter to show edits from other people to both users’ talk pages
  • We will want to display log items on the Timeline
  • Admins should be able to see deleted revisions
  • It would be helpful to allow for annotations of the Timeline that could be shared with other users.
  • It would be helpful to compare interactions across multiple wikis simultaneously
  • Add userlinks somewhere in the header (phab:T180722)
  • Display dates in country-neutral manner (phab:T180725)
  • If an edit summary is revdeleted, it should show as revdeleted, not blank. (phab:T180829)

• Defects to investigate/fix:
  • A bug was found that made some interactions show as entirely blank. This may be a result of the 500 limitation, but we’ll still investigate it and fix if needed.
  • We received a bug report about the tool not showing all namespaces. We’ll look into it. (phab:T180647)

Many of the suggestions that do not have tickets yet are also documented on the parent Phabricator task phab:T179607. Let me know if I misstated or missed anything. New feedback welcome, as always! I'll post an update around December 6, when we hope to have a next version ready to test.

Thank you! — Trevor Bolliger, WMF Product Manager (t) 23:46, 22 November 2017 (UTC)