The Signpost

James Heilman
Related articles
Knowledge Engine

WMF strategy consultant brings background in crisis reputation management; Team behind popular WMF software put "on pause"
6 February 2017

Knowledge Engine and the Wales–Heilman emails
24 April 2016

[UPDATED] WMF in limbo as decision on Tretikov nears
24 February 2016

Search and destroy: the Knowledge Engine and the undoing of Lila Tretikov
17 February 2016

New internal documents raise questions about the origins of the Knowledge Engine
10 February 2016

More articles

Jimmy Wales
Discovery presentation with references to "federated open data sources", including non-Wikimedia sources, and public curation of relevance
Tom Cruise. Jimmy Wales pointed out to James Heilman that isn't able to answer natural-language questions about him like "How old is Tom Cruise?" or "How many children does Tom Cruise have?". Because Google provides answers to such question, Wales added, "users don't come to us".
+ Add a comment

Discuss this story

This is the best-grounded look at the whole Heilman affair since it began, aided of course by the digging you folks at the Signpost have done and by the addition of the actual email chain between Wales and Heilman.

What a tale of technical overreach, fiduciary irresponsibility, behind-the-scenes machinations, treachery and duplicity!

Magnificent wordsmithing by Andreas Kolbe. StaniStani 00:10, 25 April 2016 (UTC)[reply]

My compliments on another excellent piece of work, Andreas. You should really try to get these articles more widely distributed. -- Seth Finkelstein (talk) 01:28, 25 April 2016 (UTC)[reply]

  • Wow, usually when someone says that the other party "took things out of context", I assume that they meant that the discussion before and after the quoted section would lead to a different interpretation. I didn't think that Wales literally cherry-picked sentences out of a long discussion to make both sides of the discussion look radically different than what they were. I really don't understand why Wales and the WMF have been so ridiculous about this whole thing- they had an obvious problem, came up with an ambitious solution, it turns out that they couldn't really do it, and... they now feel the need to lie and cast aspersions and throw people under buses for it. Guys, if you want to be a big-shot "tech company in the field of education/charity", then you need to take tech company 101: not every neat idea you have works out, and the takeaway is to learn from it, not fire everyone who disagreed. --PresN 01:35, 25 April 2016 (UTC)[reply]

Sorry but this is just a bunch of misconceptions. A query dialog engine is not a Google competitor, it is not even close. (Why do I waste time on reading this?;/) Jeblad (talk) 06:21, 25 April 2016 (UTC)[reply]

  • Excellent work, Andreas. It is clear that ambitions went far too high. Chiswick Chap (talk) 07:04, 25 April 2016 (UTC)[reply]
  • These are Wikimedia Movement resources and the WMF is simply a steward of the resources. It is disclosure in normal English of our strategy / goals that I am currently requesting rather than full scale consultation. Also typically those most involved in a conversation are also some of the most informed , I agree w/DocJames on this my view we are not painted on the wall (we edit for hours work , logic dictates we should have a voice). While the general idea by Jimbo Wales is great, its a matter of "whether the end justify the means"? (lack of transparency)..NO.--Ozzie10aaaa (talk) 11:44, 25 April 2016 (UTC)[reply]
  • Of course. WMF prepares 35 million dollars for a "knowledge engine" but can't spare a couple thousand for digitizing public domain materials in "the global south" or "developing communities" or whatever their term is now. Priorities, priorities... and the people who speak out get shafted. — Chris Woodrich (talk) 15:51, 25 April 2016 (UTC)[reply]
    • Chris, more to the point, one might weigh up some of the spending on physical meetups, trinkets, and carbon-intensive travel and accommodation, against clearly high-impact tasks such as digitizing. Just my 2 cents' worth. Tony (talk) 15:58, 25 April 2016 (UTC)[reply]
  • Since everyone has (rightfully) praised Andreas for this penetrating article, I'd like to ask a possibly stupid question: what would this proposed search engine offer a user that isn't already available? Besides the usual search function, there are hyperlinks between articles, similar articles are grouped into categories, & similar materials on different projects (viz. Commons or Wikisource besides Wikipedia, or even other-language Wikipedias) have links in the article. And when Wikidata matures sufficiently, that will provide a means to search for material between projects. And while improvements to the search function could be made, it will help a user to mine Wikipedia for all related information. So if I want to know what the Wikimedia projects have about Tom Cruise or Queen Elizabeth II, it's not that hard to find it all at present. Far easier than the library card catalog (or Reader's Guide to Periodical Literature) I had to rely on as a student decades ago. So what would a search engine offer that a user doesn't currently have -- or is likely to have in the not so distant future? -- llywrch (talk) 17:09, 25 April 2016 (UTC)[reply]
    • The idea that simply duplicating the ability to answer a simple question like "How old is Cruise?" on Wikipedia's home page will pull readers off of Google's home page to ours (seems we do want to compete with them for "home-page market share")... the idea seems silly. How many readers are so helpless that they can't search for Cruise himself, and easily find his age in the infobox. The only reason to leave Google's engine is for a specialized search that it can't handle. We recently had a discussion about Semantic Mediawiki, which tries to answer more sophisticated questions. From that you'll see that we have a long way to go to catch up to the Wolfram Alpha knowledge engine. It might be less expensive to just buy that. wbm1058 (talk) 23:43, 25 April 2016 (UTC)[reply]
      • +1 The future would be an IPA (e.g. Siri, Cortana), not Wolfram, and some are even free software. --Molarus (talk) 01:02, 26 April 2016 (UTC)[reply]
      • Platypus (backed by Wikidata) can answer it, FWIW. --Ori Livneh (talk) 03:57, 26 April 2016 (UTC)[reply]
      • Which just shows that spending resources developing a search engine is wasted effort. If the future is something similar to the proposed semantic web, Wikidata is a strong first step towards that -- & already supports a few proof of concept examples. Further, IIRC those examples were developed without Foundation backing. All that having the WMF create another search engine accomplishes is to add another line to someone's resume. (And by saying "someone" I'm not trying to say Lila Tretikov in a cute way; as more information comes out, the more obvious it is that there are other people who are likely to be the real person behind the Knowledge Engine. Treitikov might have been only a scapegoat.) -- llywrch (talk) 16:08, 26 April 2016 (UTC)[reply]
    • The {{Orphan}} template features a nifty "find link" tool that is very helpful for creating links to orphan articles. This is part of the work of crowdsourcing for relevance. Doesn't Google's algorithm give priority to pages with a lot of incoming links pointing to them? So whatever "knowledge engine" we build will be more powerful if it has a stronger web of interwiki links to build off of. Just a little thing like a bot that ran Edward Betts' tool against our entire database of orphans and pointed out the most linkable ones would be helpful. Maybe I'll ask for it in the next round of the community wishlist survey, but that seems like a waste of time when the bulk of resources are directed elsewhere. wbm1058 (talk) 00:53, 26 April 2016 (UTC)[reply]
  • Another excellent piece by Andreas Kolbe, the best writer in the field of Wikipedia-focused journalism. So we see again that Jimmy Wales has some honesty problems. How soon is he leaving the Board? Chris Troutman (talk) 17:47, 25 April 2016 (UTC)[reply]
  • Gotta pile on and agree that Andreas rocks. Nice to see Signpost hitting its stride again and putting April Fools' Fortnight behind us (but for another ArbCom melodrama). wbm1058 (talk) 01:25, 26 April 2016 (UTC)[reply]
  • Agree, excellent piece of work. I hope we can make progress getting some answers. SarahSV (talk) 02:13, 27 April 2016 (UTC)[reply]
  • Well, Jimbo allowed the Signpost bot delivery and some users' comments posted after it to be archived off his talk page with no reply, which is his usual way of "responding" to things he doesn't want to respond to, so I think you can interpret that as his "answer". Note that he has edited his talk page since then. -- (talk) 17:51, 27 April 2016 (UTC)[reply]
  • Excellent article. I hope the WF doesn't spend tens of millions on a better search engine. --Frmorrison (talk) 15:09, 6 May 2016 (UTC)[reply]

Some thoughts

Hi. Sorry if this is a daft question, but this piece is marked as an op-ed. What opinion is being expressed?

Does anyone disagree that our internal search needs improvement? I would think that Andreas and others would be supportive of efforts to have free, open, and independent search functionality. Below other mission-critical services such as providing SQL and XML data dumps, search is pretty important infrastructure, especially as the Wikimedia projects grow.

If we took an input string such as "How old is Tom Cruise?" and broke it up into pieces, I think we could, with some effort, program this and similar queries to return specific data points. We could look at the most relevant Wikidata item (d:Q37079) to extract the "date of birth" field's value ("3 July 1962") and then do a simple date calculation to show that Tom Cruise is currently 53 years old. Or, if we can get the search results to be better, we can pull out and highlight specific data points alongside the search results.

After we solve "How old is [famous person]?"-type queries, we can add support for alternate phrases such as "What age is [famous person]?" Once we solve that, we can move on to programmatically answering other "easy" queries. I don't think what's being described here requires artificial intelligence or IBM's Watson.

You want a concrete opinion? The search results at Special:Search/How old is Tom Cruise? are currently terrible. Tom Cruise bafflingly doesn't appear in the top 100 results. If Tom Cruise did appear in these results, we could look at the search input, see that it uses a known keyword ("age" or "old"), and then extract that information programmatically to serve our reader/researcher more quickly. Who opposes doing this?

Let's talk about how we can improve search and what that will require. Does an organization similar to the Wikimedia Foundation (or the Knight Foundation, for that matter) need to be involved? What value do these organizations provide? I think there's plenty of room for intelligent and thoughtful discussion about priorities and functionality and serving our readers. Can we start now? --MZMcBride (talk) 03:23, 26 April 2016 (UTC)[reply]

Hi! I think that many people are aware of imperfections in our current search functionality. But I don't think that it is a good idea to try to build a searchengine that uses natural language processing to get answers from a semantic wiki. That seems far too ambitious to me. And let's face it, they are not Google, they simply do not have the people and skill required. To reach such a goal you need to split it up in smaller, more manageable tasks, and I think it starts with improving or even rewriting the current search functionality.
Wikipedians don't really need a search engine that tells them how old Tom Cruise is, because we got a template for that (in this case {{birth date and age|1962|7|3}} which renders as: July 3, 1962 (age 53)). Internet users in general may need such a search engine, but creating it is difficult and making it popular is even more difficult, and I believe that big companies like Google and Apple (and even Microsoft) who have been doing research into (and experiments with) this kind of stuff for a long time now are far more likely to create something that actually works. The WMF is not a software company, and I don't think they can compete with the big guys in this field (Google, Siri), so I think they should focus on their niche.
Personally I wish they would be far less ambitious. I do want them to improve the search engine, maybe even to rewrite it from scratch if they believe that that is the best solution, but please keep offering roughly the same functionality as before, with some improvements and additions, instead of trying to create something superambitious that is gonna be a waste of time and money in the long run. There are many smaller improvements possible, for example the MediaWiki software does offer the ability to search for links only in a specific namespace, but this functionality is disabled on WikiMedia projects, due to efficiency issues.
Imagine if they would successfully create a search engine that gives correct answers to questions in plain English. Imagine if people (who are currently using Google for this type of task) would switch to using this new search engine, built on open standards with open data. Then Google will immediately embrace, extend and extinguish it. The Quixotic Potato (talk) 15:33, 26 April 2016 (UTC)[reply]
Hi The Quixotic Potato. The Wikimedia Foundation sure hires a lot of tech folks (cf. wmf:Staff and contractors) if it's not a software company. What kind of organization do you think the Wikimedia Foundation is? :-)
As noted at MediaWiki#Searches and queries, the search back-end was basically rewritten/replaced in 2014. And there have been substantial improvements to the site search functionality since then. But we need to do better; I think we're all agreed on that.
Regarding the threat that Wikimedia projects face from Google, I wrote about that here: mailarchive:wikimedia-l/2016-April/083722.html. --MZMcBride (talk) 00:51, 28 April 2016 (UTC)[reply]
MZMcBride, I agree, the op-ed designation seems odd; this strikes me as simply good reporting that, in some areas and transparently, draws conclusions that could be construed as opinions. In most publications, this is simply referred to as "news reporting." But the Wikipedia world can be highly sensitive around the issue of neutrality, and this particular topic is highly sensitive. My guess is that's why it was presented as an op-ed. That designation signals that others might be welcome to submit competing interpretations. In that sense, I like the choice; Jayen466 (Andreas) is a Signpost editor, so it's good to be extra cautious about any impression that his own views and the editorial position or policies of the publication are getting blurred.
On the substance of the piece: Yes, I think everyone can agree that there is room for substantial improvement in Wikipedia/Wikimedia search. I think that has been broadly agreed by many people over the recent months. But I don't see that as a central question in this piece. A very important, unanswered question remains: was the board justified in dismissing a recently-(s)elected Trustee? Or was Docjames actually the only Trustee trying to do the right thing, in the face of a board apparently deeply tied to going about things in a bad way (standing by its Executive Director despite massive staff opposition and attrition, and neglecting to clearly communicate its ambitions to important stakeholder groups like volunteers and staff)?
That question is an important one, and this piece advances the effort to unravel it. -Pete (talk) 15:38, 26 April 2016 (UTC)[reply]

I don´t think we need a question/answering-software, rather an assistant that could do some tasks for editors and readers. We could start with writing into the searchbox something like: WD, show me the article about Tom Cruise, or WD, read out aloud the introduction of that article, or WD, tell me who wrote most of this article, or WD, show me all media files commons has about Tom Cruise. --Molarus (talk) 23:48, 26 April 2016 (UTC)[reply]
The question is moot. We are here to write an encyclopedia with no end goal and no deadline. Meanwhile, the Board foolishly panicked because the pageviews are down thanks to Google's knowledge graph. While the Board wants to lie to us in pursuit of the next hot thing, I'm happy as a clam to write articles that no one will ever read. We, the editors, are fundamentally different than the Board and incidents like this make plain the depravity of the folks in San Francisco. Chris Troutman (talk) 17:32, 7 May 2016 (UTC)[reply]
There is a category on Commons for media files related to Tom Cruise. And the searchbox already works, if you type in Tom Cruise you will go to his article. Reading aloud is not something a computer can do; try using a screenreader for a day (I did, because I was curious how blind people experience the internet). Trust me, it sucks. Asking who wrote most of an article is incredibly difficult to calculate, and the answer is useless. The Quixotic Potato (talk) 15:32, 27 April 2016 (UTC)[reply]
You think it's useless? I think it would be useful to know whether an article was mostly written by, say, a PR firm or publicist being paid by the subject of an article. I don't dispute that accomplishing this might be difficult. -- (talk) 17:51, 27 April 2016 (UTC)[reply]
Lots of things can be useful. Only some of them are practical. What you propose is not the latter. (talk) 22:48, 27 April 2016 (UTC)[reply]
We could start with writing into the searchbox something like: WD, show me the article about Tom Cruise. Ah, yes, eight words to get to the Tom Cruise article rather than two. How incredibly useful. (talk) 22:48, 27 April 2016 (UTC)[reply]
Sure, it's possible to cherry-pick a silly example. Can you think of a reason that our search functionality should not support examples such as "show me all media files about Tom Cruise" or "What is Tom Cruise's age?"?
A friend showed me <>. This gives the correct answer (53) using Wikidata and other Wikimedia sources as its back-end, as I understand it. Pretty neat! It fails for queries such as <> and <>, but queries such as <> work. This tool is fun to play around with. Like Wolfram Alpha, it's easy to confuse or break it, but we could incorporate this type of functionality into our internal search engine immediately. No making perfect the enemy of the good, especially if we can keep false positives low with more conservative logic (err on the side of being quiet). --MZMcBride (talk) 00:42, 28 April 2016 (UTC)[reply]
The point is that I have written "We could start". After we have installed an open source software that is able to do that, we could append our own code for more useful tasks. This is a low cost solution with quick results. The board was right that we have to adapt new technology, but it is not Google that we should try to follow. Part of the problem that the board is deciding such things behind closed doors is that they lack the experience of Wikipedia editors and WMF software engineers. --Molarus (talk) 17:50, 29 April 2016 (UTC)[reply]

About Jimmy's behaviour and character

  • The big issue for me here is Jimmy's lying and defaming. It is clear now that James's Facebook comment exactly, accurately reflects Jimmy's statements to James about the Knowledge engine but in his gaslighting email Jimmy accuses James of misrepresenting his (Jimmy's) position. We have a serial liar strutting about posing as our spokesperson, squatting on a board seat, defaming a hard-working popularly-elected volunteer.
  • Another thing: A new board member discovers there's been a plan to develop an internet search engine that could cost tens of millions of dollars, and takes it to the other board members. And Jimmy's response is evasive and ambiguous. When James points up the contradiction between

    "we are not building a search engine"

    and the Knight grant documentation's,

    "Knowledge Engine by Wikipedia will be the Internet's first transparent search engine ... the Knowledge Engine by Wikipedia, a system for discovery of reliable and trustworthy public information on the Internet"

    Jimmy says,

    "I'm not really sure what is causing your confusion here. Perhaps it is just the term 'search engine' which in some contexts may mean 'a website that one goes to as a destination in order to find things on the web, such as Google, Bing, or Yahoo' and in other contexts can mean 'software for searching through a set of documents and resources'. But I'm not really sure what your concern is..."

    Here is where the gaslighting begins in my opinion: It's bloody obvious what James's concern is. And Jimmy is acting as if there's nothing remarkable or potentially concerning going on and that "we are not building a search engine" can be reconciled with the Knight grant documentation.
  • And another: It's clear now that the WMF was waiting for the right moment to let the community in on this scheme. Jimmy:

    "For me, it's more of a question of what kind of consultation should happen and when. A commitment to explore a concept through an external grant doesn't strike me as the right point necessarily to engage in a full-scale consultation."

    So, James's concern that this was being kept from the community was well-founded. Jimmy didn't trust the community with this information. James did. --Anthonyhcole (talk · contribs · email) 01:22, 27 April 2016 (UTC)[reply]
Hi Anthonyhcole and Pete. What do you make of the fact that the vote on the resolution removing James Heilman was 8–2 (really 8–1)? --MZMcBride (talk) 01:18, 28 April 2016 (UTC)[reply]
I don't know what's behind that. But the trustees on that board seem to be in the habit of voting along with the rest of the board, to get along, to present an image of decisiveness, or something (a practice that seems juvenile and deceptive to me). Jimmy claims that he went into the meeting that sacked James with the intention of voting with the majority. James's proposing the motion to accept the Knight grant despite his misgivings may be an instance of that. That might explain the numbers. Or perhaps James deserved it for some (still) as yet unexplained reason. My point above isn't that James is a useful board member (though having worked beside him on the WikiProject Med Foundation board for years I can attest to his integrity and honour there), it's that Jimmy isn't. --Anthonyhcole (talk · contribs · email) 03:14, 28 April 2016 (UTC)[reply]
Hi MZMcBride. All hail the master of all master sockmasters. Starved for attention, much? Do us a favor and stay away when the grown-ups are talking. You have nothing of value to add here. DracoE 12:25, 29 April 2016 (UTC)[reply]
Hi DracoEssentialis. I think Special:Contributions/DracoEssentialis speaks for itself. :-) --MZMcBride (talk) 19:25, 1 May 2016 (UTC)[reply]
MZMcBride, I don't think it's particularly significant. It's easy for any group to get swept up in the moment for any number of reasons -- legitimate or illegitimate. If I didn't know anything else, my starting assumption would be that they reached near-consensus through a deliberation that was careful and thorough commensurate with the gravity of what they were doing. But in fact, we now know a whole lot of other stuff, which tells a pretty clear story of a group that was layering one bad decision on top of another, rushing this one due to external factors, neglecting to seek independent advice or mediation, underestimating the gravity of removing a board member, etc. etc. So no, I don't think the vote number is especially useful information in the face of all the other stuff we know several months in. -Pete (talk) 05:07, 2 May 2016 (UTC)[reply]

Was the related email, from Jimbo to Doc James of 30 December 2015 ever shared publicly? HolidayInGibraltar (talk) 19:10, 27 April 2016 (UTC)[reply]

No. --Andreas JN466 20:28, 27 April 2016 (UTC)[reply]
Jimmy selectively quoted from the email exchange on his talk, in a way that made James look like the unreasonable party. James was more than within his rights to make the context public. --SB_Johnny | talk✌ 00:42, 28 April 2016 (UTC)[reply]

Kudos to Andreas for an incisive and revealing exposé. Is Wales any longer appropriate as a WMF board member and self-appointed WP figurehead? Given the long, damaging record of evasions, obfuscations, manipulations, lies, misdirections, misrepresentations, distortions, and self-serving personal attacks, the answer couldn’t be more obvious. Writegeist (talk) 19:11, 29 April 2016 (UTC)[reply]

Same. I had doubts for some time, but this makes it very clear. Peter Damian (talk) 06:07, 30 April 2016 (UTC)[reply]

Data sources

"In recent months, [Jimmy] has multiple times referred to the possibility that 'non-WMF resources might be included in a revamped discovery experience' or that 'some important scholarly/academic and open access resources could be crawled and indexed in some useful way relating to Wikipedia entries' while insisting that any suggestions 'that this is some kind of broad Google competitor remain completely and utterly false.'"

Please don't forget Fox News appearing in a sample search result. --NaBUru38 (talk) 16:11, 29 April 2016 (UTC)[reply]


The Signpost · written by many · served by Sinepost V0.9 · 🄯 CC-BY-SA 4.0