"Abstract Wikipedia"

[Poetlister]

"Abstract Wikipedia" is the latest great new idea from the WMF.

Abstract Wikipedia proposes a new way to generate baseline encyclopedic content in a multilingual fashion, allowing more contributors and more readers to share more knowledge in more languages. It is an approach that aims to make cross-lingual cooperation easier on our projects, increase the sustainability of our movement through expanding access to participation, improve the user experience for readers of all languages, and innovate in free knowledge by connecting some of the strengths of our movement to create something new.

Meta

[Midsize Jake]

Maybe it's just me, but if I were a Wikipedian I'd be fighting tooth-and-nail against this sort of initiative. There's almost no way this isn't going to be abused, and probably in ways they haven't even imagined yet. And it's bad enough that they're using Wikidata as the basis for it in the first place.

I'd also object to the use of the term "abstract" in this context, if only because it shows disrespect for the likes of Wassily Kandinsky and Jackson Pollock.

[Poetlister]

The name is only provisional; see here. Remember that the WMF is basically a fundraising organisation. They have no doubt concluded that this project will encourage donations, and never mind what the Wikipedia community thinks.

[WanderingGuest]

There's almost no way this isn't going to be abused, and probably in ways they haven't even imagined yet.

That sentence describes everything ever invented.
I don't see any problem with this; just looks like a Wikidata mod that will it more human readable with the added benefit that it'll be significantly easier to machine translate. Pretty cool.

[Bezdomni]

Anagrams to I saw tribe pick data. Sounds about right.

e.g. Edward Colston (T-H-L)'s philanthrop root is (still) up 125% (4 -> 9) since Bristol residents dragged his statue through the streets and dumped it over the bridge.

also

Prick bite: I was data.

Bitter about being mauled by the Mob? Who?! Me?

[AndyTheGrump]

Also pick data write bias

[Zoloft]

Also pick data write bias

I saw pita bread tick.

[Giraffe Stapler]

Also pick data write bias

I saw pita bread tick.

Baited pricks await.

[Zoloft]

Also pick data write bias

I saw pita bread tick.

Baited pricks await.

Winner!

[Smiley]

https://meta.wikimedia.org/wiki/Abstract_Wikipedia

2021: Launch Wikifunctions on the Beta Cluster
2022: Launch Wikifunctions in production; start community discussions about the architecture
2022: Start development of Abstract Wikipedia proper
2023: Integration of Abstract Wikipedia into Wikipedia proper and other sister projects

[Smiley]

Here is a conversation and decision we need to have before launch of
Wikifunctions:

*How should the contents of Abstract Wikipedia and Wikifunctions be
licensed?*

Since the discussion is expected to be potentially complicated, let us keep
a single place of record for discussing this question:

https://meta.wikimedia.org/wiki/Abstrac ... discussion

We would like the discussion to go on for four weeks and that we have some
form of consensus by December 20th. This is not planned to be a vote
(although it might have votes in it and it might even be closed by a vote
in case no other form of consensus finding works out).

22 Nov 2021, 10:24

Thanks. The key question to my mind is whether abstract content and the
resulting foreign-language text output should be CC0 (like Wikidata) or CC
BY-SA (like Wikipedia).

The difference is that with CC0, re-users do not have to credit Wikimedia
or Wikipedia for the material they use. Voice assistants like Amazon Alexa,
Apple Siri and Google's Assistant along with search engines like Google and
Bing would no longer have to say that they got the material from a
Wikimedia project. They would also be free to copyright any derivative
works.

I think both of these results are undesirable, for reasons aptly described
by Heather Ford in her Wikipedia@20 chapter, "The Rise of the Underdog".

Here is one part of the chapter that speaks to this:

"... Wikipedia’s facts are now increasingly extracted without credit by
artificial intelligence processes that consume its knowledge and present it
as objective fact.


"As one of most popular websites in the world, it is tempting in 2020 to
see Wikipedia as a top dog in the world of facts, but the consumption of
Wikipedia’s knowledge without credit introduces Wikipedia’s greatest
existential threat to date. This is not just because of the ways in which
third-party actors appropriate Wikipedia content and remove the links that
might sustain the community in terms of contributions of donations and
volunteer time. More important is that unsourced Wikipedia content
threatens the principle of verifiability, one of the fundamental principles
on which Wikipedia was built.


"Verifiability sets up a series of rights and obligations by readers and
editors of Wikipedia to knowledge whose political and social status is
transparent. By removing direct links to the Wikipedia article where
statements originate from, search engines and digital assistants are
removing the clues that readers could use to (a) evaluate the veracity of
claims and (b) take active steps to change that information through
consensus if they feel that it is false. Without the source of factual
statements being attributed to Wikipedia, users will see those facts as
solid, incontrovertible truth, when in reality they may have been extracted
during a process of consensus building or at the moment in which the
article was vandalized.


"Until now, platform companies have been asked to contribute to the
Wikimedia Foundation’s annual fund-raising campaign to “give back” to what
they are taking out of the commons.[23]
But contributions of cash will not solve what amounts to Wikipedia’s greatest
existential threat to date. What is needed is a public campaign to
reinstate the principle of verifiability in the content that is extracted
from Wikipedia by platform companies. Users need to be able to understand
(a) exactly where facts originate, (b) how stable or unstable those
statements are, (c) how they might become involved in improving the quality
of that information, and (d) the rules under which decisions about
representation will be made.


"Wikipedia was once recognized as the underdog not only because it was
underresourced but also, more importantly, because it represented the just
fight against more powerful media who sought to limit the possibilities of
people around the world to build knowledge products together. Today, the
fight is a new one, and Wikipedia must adapt in order to survive.


"Sitting back and allowing platform companies to ingest Wikipedia’s
knowledge and represent it as the incontrovertible truth rather than the
messy and variable truths it actually depicts is an injustice. It is an
injustice not only for Wikipedians but also for people around the world who
use the resource — either directly on Wikimedia servers or indirectly via
other platforms like search."

There is a lot at stake in this discussion.

25 Nov 2021, 1:52

[Smiley]

[Moral Hazard]

Maybe it's just me, but if I were a Wikipedian I'd be fighting tooth-and-nail against this sort of initiative. There's almost no way this isn't going to be abused, and probably in ways they haven't even imagined yet. And it's bad enough that they're using Wikidata as the basis for it in the first place.

I'd also object to the use of the term "abstract" in this context, if only because it shows disrespect for the likes of Wassily Kandinsky and Jackson Pollock.

Abstract Wikipedia should be Lambda Wikipedia, to stand up against phallocentrism and heteronormativity and to honor the lambda calculus (T-H-L) (abstraction (T-H-L)).

[Zesty]

For what its worth, I think its a pretty good idea.. not that I understand it, or Wikidata.

[Jim]

For what its worth, I think its a pretty good idea.. not that I understand it, or Wikidata.

Please expand.

[Smiley]

It could be good if they don't just import a load of worthless crap from wikidata and allow ravenous multinationals to rebadge and resell everyone's work, which unfortunately, all be it unsurprisingly, appears to be the plan.

[Zesty]

For what its worth, I think its a pretty good idea.. not that I understand it, or Wikidata.

Please expand.

Well the whole thing with wikidata was computer readable Wikipedia data right? And this builds on that by automagically building articles using "WikiFunctions"? Soon we won't need editors at all!

[Smiley]

Soon we won't need editors at all!

That's the best thing about it.

[AndyTheGrump]

The real solution would be to automate the reading process too.

[Smiley]

The real solution would be to automate the reading process too.

[Ritchie333]

Just the mention of the term "Abstract" makes me wonder if this is a plot by a bunch of architecture astronauts.

[Giraffe Stapler]

From the Signpost: Wikimedia Foundation's Abstract Wikipedia project "at substantial risk of failure"

[DanMurphy]

WMF, never change.

A response authored by eight Foundation staff members from the Abstract Wikipedia team (published simultaneously with the Fellows' evaluation) rejects these recommendations... Vrandečić's team then goes on to reject the evaluation's core recommendations, presenting the expansive scope of Wikifunctions as a universal repository of general-purpose functions a done deal mandated by the Board (the Wikimedia Foundation's top decision-making authority), and accusing the Google Fellows of "fallacies" rooted in "misconception":

[Bezdomni]

From the Signpost: Wikimedia Foundation's Abstract Wikipedia project "at substantial risk of failure"

Maybe you meant to point to this highly puzzling article?

Wikifunctions is entity Q104587954 at Wikidata where its official website redirects to Wikilambda on Meta, which redirects to a page titled Abstract Wikipedia. This page points to phabricator, as it appears there is still not a website associated with Abstract Wikipedia, WikiLambda, or Wikifunctions (?)

But, never fear, there is a logo!



ps: the above is not it.

ps: it does appear to me that the Google workgroup agrees with me that constructors should be part of Wikidata, cf. the discussions long ago here and on WD about the predicate/property "banned in"...

I see that the Wikidata "card" for Ulysses still does not manage to encode the book's banning in the US and UK, only that Bloomsday and United States v. One Book Called Ulysses are predicative "significant events".

Oh well, such is life. I dropped trying to mess with that long ago as it was clear it would only lead to foodfights.

[Giraffe Stapler]

From the Signpost: Wikimedia Foundation's Abstract Wikipedia project "at substantial risk of failure"

Maybe you meant to point to this highly puzzling article?

Yes, I did. Thank you.

Wikifunctions is entity Q104587954 at Wikidata where its official website redirects to Wikilambda on Meta, which redirects to a page titled Abstract Wikipedia. This page points to phabricator, as it appears there is still not a website associated with Abstract Wikipedia, WikiLambda, or Wikifunctions (?)

Did you check the beta cluster? You didn't check the beta cluster, did you?

Wikifunctions

[Bezdomni]

I didn't check the beta cluster.

{ 
    "Z1KA" : "Z5",
      "Z5KA": {
         "ZWD40": "Z6",
         "ZA0K": "Z8"
        }
 }

Thanks.

[orangepi]

I don't have time to ignore these idiots. Glad to see you people doing it for me.

[Goodbye01]

Soooooo, Wikifunctions was released to the public and I saw it mentioned on social media. My initial thought was "wow finally a wiki for programmers that is going to be better than rosetta code" So I got involved a bit and saw some problems with it and the "community" that they supposedly want to build.

I started a section on the project's talk page. No response from WMF yet, but I have a bad feeling..

[No Ledge]

Their current Wikifunctions Catalogue is rather thin (yawn). I suppose in fairness, they're just getting started with reinventing the wheel.

[tinyboxs]

All of the non-linguistic stuff is just an extra. All they want is to create a crowdsourced interlanguage, an ancient concept from the days of symbolic NLP (c. 1950–90), surreal to see in age of LLMs. The key difference from normal interlanguages is that they expect sorry editors to eventually work directly in the deep and gloomy mines of the symbolic representation, rather than using it as a step in translating from another natural language. To create the interlanguage, they also need people to produce countless "renderers" to translate from each interlanguage construct to the equivalent in each available language, which is the actual purpose of Wikifunctions. They already have representations of the terms stored on wikidata. IMO their best chance would be to ask Wiktionary nerds, or maybe just to get GPT to generate them based on grammar guides, but this project is idiotic either way.

Here’s an example of what the finished interlanguage editors are expected to write in might look like, from Denny Vrandeci’s paper:

Code: Select all

Article(
  content: [
    Instantiation(
      instance: San Francisco (Q62),
      class: Object_with_modifier_and_of(
        object: center,
        modifier: And_modifier(
          conjuncts: [cultural, commercial, financial]
        ),
        of: Northern California (Q1066807)
      )
    ),
    Ranking(
      subject: San Francisco (Q62),
      rank: 4,
      object: city (Q515),
      by: population (Q1613416),
      local_constraint: California (Q99),
      after: [
        Los Angeles (Q65),
        San Diego (Q16552),
        San Jose (Q16553)
      ]
    )
  ]
)

bleh

[Giraffe Stapler]

IMO their best chance would be to ask Wiktionary nerds, or maybe just to get GPT to generate them based on grammar guides, but this project is idiotic either way.

A decade ago this was an interesting (but flawed) idea. 1n 2023 it feels a bit silly. It's on a level with asking volunteers to add tags to Commons images when you know that you could already do a decent job of it with an AI image classifier. I'm sure Google or Microsoft or Facebook or etc would be happy to get a little good press by helping out on this kind of job.

[Goodbye01]

All of the non-linguistic stuff is just an extra.

Thanks for summarizing my concerns in a brief but concise sentence.

[Banderas22]

[I registered an account just to say this so please don't shit on me too bad.]

Wouldn't Abstract Wikipedia resolve - to some degree - the issue of BLP battle zones where every suspect thing a person has done gets jammed into the lead?

I'm guessing Abstract Wikipedia might produce something like: "Jane Smith is a Nobel Prize-winning chemist who holds the Bob Roberts Chair in Chemistry at Johns Hopkins University. Her work focuses on the synthesis of some original persistent carbenes."

Whereas EN.WIKI is more like: "Jane Smith is an alcoholic Nobel Prize-winner. As a high school student she was arrested for DUI and later went into treatment for alcoholism. A 1991 paper on which she was a non-corresponding, contributing author was criticized over several punctuation errors."

[tinyboxs]

[I registered an account just to say this so please don't shit on me too bad.]

Wouldn't Abstract Wikipedia resolve - to some degree - the issue of BLP battle zones where every suspect thing a person has done gets jammed into the lead?

I'm guessing Abstract Wikipedia might produce something like: "Jane Smith is a Nobel Prize-winning chemist who holds the Bob Roberts Chair in Chemistry at Johns Hopkins University. Her work focuses on the synthesis of some original persistent carbenes."

Whereas EN.WIKI is more like: "Jane Smith is an alcoholic Nobel Prize-winner. As a high school student she was arrested for DUI and later went into treatment for alcoholism. A 1991 paper on which she was a non-corresponding, contributing author was criticized over several punctuation errors."

I have no idea why you would think that.

[Giraffe Stapler]

Wouldn't Abstract Wikipedia resolve - to some degree - the issue of BLP battle zones where every suspect thing a person has done gets jammed into the lead?

Why would you think this would change? (That's a real question.)

Perhaps if the idea is for Abstract Wikipedia to generate boilerplate articles with the correct facts filled in, but I don't think that's what anyone wants. Or maybe it is. To be honest, I never had any faith in it amounting to anything so I haven't looked very seriously into where it's headed these days.

[Banderas22]

Maybe I don't understand how AbWP works. Is it machine-generated content from WD data fields?

[Goodbye01]

[I registered an account just to say this so please don't shit on me too bad.]

Wouldn't Abstract Wikipedia resolve - to some degree - the issue of BLP battle zones where every suspect thing a person has done gets jammed into the lead?

How? The only way I would imagine this to work is for the BLP POV pushers to be gatekeeped by the harder-to-understand interface for editing Abstract Wikipedia, since it is obviously going to be harder to edit Abstract Wikipedia than enwiki.

[Ming]

Wouldn't Abstract Wikipedia resolve - to some degree - the issue of BLP battle zones where every suspect thing a person has done gets jammed into the lead?

Why would you think this would change? (That's a real question.)

Perhaps if the idea is for Abstract Wikipedia to generate boilerplate articles with the correct facts filled in, but I don't think that's what anyone wants. Or maybe it is. To be honest, I never had any faith in it amounting to anything so I haven't looked very seriously into where it's headed these days.

Perhaps it's because, like Agassiz, AbWP is weaker in the concrete.

[tinyboxs]

Maybe I don't understand how AbWP works. Is it machine-generated content from WD data fields?

No. Wikidata will mainly serve to translate terms. But there's no way that will be as simple as substitution, since the grammatical roles of individual terms get thorny across many languages, so Wikifunctions needs to hurry up and implement a system for directly querying Wikidata from generator functions. But Wikidata doesn't even have the information they need. Wiktionary does, but not in a machine-readable format. For example, where on Wikidata can you determine that the equivalent of the English verb "like" in Japanese (好き) has the subject and object roles flipped? Since this information is nowhere, will they need to hardcode it into a generator function, or even create a whole new Interlanguage construct just for the concept of liking?

[Banderas22]

Maybe I don't understand how AbWP works. Is it machine-generated content from WD data fields?

No. Wikidata will mainly serve to translate terms. But there's no way that will be as simple as substitution, since the grammatical roles of individual terms get thorny across many languages, so Wikifunctions needs to hurry up and implement a system for directly querying Wikidata from generator functions. But Wikidata doesn't even have the information they need. Wiktionary does, but not in a machine-readable format. For example, where on Wikidata can you determine that the equivalent of the English verb "like" in Japanese (好き) has the subject and object roles flipped? Since this information is nowhere, will they need to hardcode it into a generator function, or even create a whole new Interlanguage construct just for the concept of liking?

Got it. Makes sense.

[No Ledge]

I just slam dunked on JD Forrester

[tinyboxs]

I just slam dunked on JD Forrester

An embarrassing way to put it.