Hoi,
Charles, thank you for the analysis. You missed some points. It is not only about the Black Lunch Table, there are other organisations involved in the development of Wikimedia content that make a similar use of the "catalog" property. Organisations like the Smithsonian, the Library of the Botanical Garden of New York.. The problem with the BLT is indeed a conflated issue and one issue you do not mention has to do with notability.

When catalog is used to build an unsorted list of items that are of relevance to that project / organisation, it follows that when an item is created and identified (by the organisation) as part of its catalog, the item is not only notable but also necessary for the functioning of the catalog. It follows that like other properties that have a similar status ("award received" is one), the sheer fact that an item identified as such is not eligible for deletion. The problem becomes trust. Do we trust an organisation, now the Black Lunch Table, to actively work on this data.

With the "award received" it is **accepted** notability. Here it is enough that a Wikipedia article or an external source identifies a Joe Blogg as the award winners and this serves as a reason to add an item and add the winning of the award as a signal that the item is not to be deleted. The purpose of these items except for completeness, is that they serve as an indicator that this item is similar to the other award winners and this is one reason why automated tools may suggest them for new articles in (a) Wikipedia.

With catalog and with the Black Lunch Table, there is no external source that helps people who "sit in judgement". It does not help them to easily pass or fail. With the Black Lunch Table we are talking about "visual artists of the African diaspora" the project is explicitly started to bring visibility to notable artists. So finding sources is not easy. It can be done and I am thankful to Pasleim for the good work he has done and is doing.

Finally, it is about purpose; what is the purpose of Wikidata. The problem we face when we talk about Wikidata that many arguments are used that have nothing to do with Wikidata but everything with Wikipedia. Original research is done with Wikidata. Why else the inclusion of all the bio-medical data? Why else the scaffolding of data through award winners, generations of royalty, the scaffolding by means of the "catalog" property.

One additional thing to consider; in a "scholarly paper" about Wikidata it was mentioned that the implementation of constraints had an dampening effect on the diversity of Wikidata. With the full frontal attack on the work of the Black Lunch Table organisation, we face a direct attack on the diversity of Wikidata.
Thanks,
       GerardM



On 6 January 2018 at 02:34, Charles Horn <charles.horn@gmail.com> wrote:
Hi Jane, Erika, and everybody else,

As a relatively recent contributor to Wikidata, I have been struggling to understand the objections to the Black Lunch Table's use of the catalog property and the points of view behind this discussion. I have read all the emails, and all the linked to discussion and talk pages, and finally had to go back to tracing all the Q and P entries and their parent classes mentioned in those discussion, to come up with my own understanding of the situation from basic principles. It has taken hours, and given how much effort has gone into the various discussions in the different locations it is frustrating that (as far as I have seen) no one objecting to the usage has been sufficiently clear what the problem is, or the exact nature of the 'abuse'. 

I don't think it is as obvious as some are claiming it to be. Nor do I think all people claiming the abuse are even aligned on what the valid issues are, since no one has been specific, and in the discussions, various issues are being conflated.

Given that people here have different backgrounds, languages, are focused on different tasks, and may not be as intimately familliar with some P and Q items as others, I believe it is worth spelling out exactly what the nature of the problem is, especially since I think every one is aware of the good faith nature of the Black Lunch Table, and presumably aligned in the goal of making Wikidata better.

I'm replying to Jane's email because this seems like the closest statement to what I can determine to be a simple valid complaint, I'm going to try to set it down here clearly so others can agree or disagree whether it is the entirety of the objection, or if there is more to it:

Q28781198:Black Lunch Table is P31:instance of  Q21025364:WikiProject

but is being used to populate P972:Catalog, which is supposed to represent an instance of Q2352616:catalog
The objection is that anything entered into P972:Catalog MUST be P31:instance of Q2352616:catalog

Is that it?

Perhaps this is obvious after the fact, but I don't think it was obvious from reading any of the initial objection notifications, or requests for deletions, or subsequent follow ups. When the usage is made in good faith, the objection does need to be spelled out so that everyone involved can check their assumptions and understanding.  

If that is the entirety of the objection, I have to agree that it is technically sound, and hopefully that provides a basis for a solution that keeps everybody happy.

Some other comments which I hope do not cloud my attempt at clarifying things:

It seems clear that the Black Lunch Table Wiki Project has a list of artists they are interested in, so I consider that the Black Lunch Table catalog is real thing. From my reading of Q2352616:catalog there is nothing that strictly specifies how it must be represented or published, nor can I see anything clear about how public it needs to be to be valid. These properties would appear to be inherited from its parent class Q386724:work  I accept people will have different opinions about what quality of publishing or how public something needs to be before it is 'good', but it does not seem like a clear cut thing, and the arguments could go on and on without much productive outcome. 

Insisting that BLT publish their list elsewhere could be trivially done to technically meet some publication threshold, and still not be good enough for some, and would not be a productive use of resources or change anything in a meaningful way. It might be nice if it were made more accessible elsewhere, but I don't think it is a technical requirement for BLT to accomplish what they are trying to do. Feel free to disagree, but the point I'm trying to make is that I can't see how an objective threshold can be set here that couldn't be unproductively argued about either way. BLT has a real catalog, opinions as to its quality, or quality of availability can vary. I'm sure polite requests to improve things can be made.  

One place where another user attempted to clarify a technical objection to BLT use of catalog is: "
  • More technically: the way how the catalog property is used by BLT is technically incorrect, which was mentioned several times and very early as well. Catalogs are typically qualifiers to catalog code (P528), or used in exhibition items, but not standalone on items about humans. Editors who spend a lot of time to fix wrong property use are not happy if we allow permanent exceptions.
—MisterSynergy (talk) 08:57, 4 January 2018 (UTC)"

I think the above misses the point, and stumbles on unclear definitions in the description of P972 https://www.wikidata.org/wiki/Property_talk:P972#Documentation  which claims the 'Domain' is 'exhibition' which seems an unnecessary elevation of one catalog type example to some special status. Below is says 'often exhibition catalog'. Often is fine, it does not mean always though. And 'typically qualifiers to catalog code' may be true, but again it does not prevent it being used without a catalog code. The Documentation on P972 could be clearer, and I consider that 'Domain': 'exhibition' to be plain incorrect on that page, which does not help clear interpretations for anyone.


I see no requirement that the usage of P972:Catalog needs a catalog code (P528) (I don't think anyone has claimed this as an primary objection against BLT catalog usage, but then it hasn't been clear). Catalogs without codes seem valid in principle, and my reading of the Wikidata properties is that this is perfectly fine in Wikidata too.

There seems to be some confusion about the subjects on the BLT catalog being thought of as attendees of BLT events (I've read this on both sides) -- individuals on the BLT catalog are there because they are (potentially) notable artists, and the purpose of BLT is to add all the appropriate links to show this notability. From what I have read, if an individual is not really notable and the links don't exist, that entry will be removed _by_ BLT, as that is what the project is about. I would take it on good faith that anyone on the BLT list is  notable.

Perhaps this raises questions about whether stubs or placeholder entries are acceptable on Wikidata? Clarity on this would be beneficial to many other editors and users of Wikidata. Do entries have to be added to Wikidata fully formed and perfect, or is there some understanding that items have differing levels of quality and can always be improved? I don't know. I think very clear statements on this would be useful, as my reading of https://www.wikidata.org/wiki/Wikidata:Notability  
2) It refers to an instance of a clearly identifiable conceptual or material entity. The entity must be notable, in the sense that it can be described using serious and publicly available references.
is that the threshold for meeting this is very low, and 'identifiable' means that someone could do it if they tried, not necessarily that all the links and supporting evidence are immediately available. I could write more on this, but again it feels like I am conflating issues which isn't helpful.

To summarise  my take on things, I started out neutral and just wanting to understand what this storm in my inbox was about. Now if I had to chose a "side", I would side with BLT.

Erika, I hope you don't take my attempt at clarification above as my personal objection, I'd like to see things resolved in a way that allows you to proceed with your project!

My attempt at a solution would be to create a BLT catalog which is an instance of Q2352616:catalog and somehow belongs to or is associated with Q28781198:Black Lunch Table, and use that in all the existing P972:Catalog locations.

Honestly I don't know if that will silence all objections, and it seems a pretty trivial semantic re-shuffling to add another layer, but it feels like a technically correct solution to what looks like a valid technical objection.

Last thing, apparently consensus advice was given to use P972:Catalog in this way, all I can assume is that either the advice was somehow mis-communicated or recieved, and the intent of the advice was always to use a BLT _catalog_ for this purpose, rather than a WikiProject, or that no one really thought about that aspect of it and the consensus advice was given and received all in good faith and acted upon by BLT as intended. Either way, it doesn't seem like that much of a big deal, mistakes are made and things are overlooked. 

If all that is wrong is Q28781198:Black Lunch Table needs a Q2352616:catalog  instance to use for their purposes, can we get some confirmation consensus and move forward?

I have already invested enough time into trying to understand this that I'm willing to help out and make it happen!

Charles Horn.

Wikidata user: Salpynx 





On 6 January 2018 at 10:55, Jane Darnell <jane023@gmail.com> wrote:
I object to your use of the catalog property to link to something that is not a catalog. I don't see why my objection leads you to expect me to offer an alternative way to track your project. I am not responsible for your project and don't understand what it is. If you can't understand that then you should not probably not be editing Wikidata.

On Fri, Jan 5, 2018 at 2:33 PM, Brill Lyle <wp.brilllyle@gmail.com> wrote:
Hi Jane,

Actually, "the rest of your email is irrelevant" illustrates the problem. I am a bit baffled at this statement.

The rest of the email is the whole point, and dismissing it illustrates the actual problem here. If Wikidatans don't want to hear about or learn about the context of the problem that needs to be solved, then what's the point of anything here? I was trying to provide background and overview of this problem in my email. If you aren't interested in learning about or hearing about what is trying to be done, then a non-indepth understanding of the issue won't work to provide a solution. 

This is not a casual, rigid WIKI:Rulez situation. The overarching effort is a goal to integrate Wikidata into the Wikipedia outreach and page improvement / page creation process for multiple projects. We are trying to solve a problem with Wikidata. We are trying to use Wikidata in an outreach project in a new way, a way that previously had consensus and had implementation that was effective and super functional. If the consensus won't meet community standards, please help us solve the problem by helping us to figure out another solution.

The bottom line is that we need to be able to tag items with a unique identifier to connect the Wikidata items to various outreach initiatives. In some way. If that basic functionality is deemed to be not allowed on Wikidata, is deemed to threaten and weaken Wikidata metadata (the latest complaint, along with accusations of the project work being original research, which is a newly creatively inaccurate characterization), if it is deemed not welcome, then let us know. It will negatively affect project outreach and integrated holistic engagement of Wikipedia editing with Wikidata but if that's the bottom line and community consensus, let us know before further work is done. None of us want to waste our time here if the free digital labor is not welcome. 

- Erika


Erika Herzog
Wikipedia User:BrillLyle

On Fri, Jan 5, 2018 at 3:18 AM, Jane Darnell <jane023@gmail.com> wrote:
Yes to exactly this part of your email: "Gerard and I thought we had consensus on this, but apparently not. We need to find some solution that will address all concerns."
The rest of your email is irrelevant to using the property for "catalog" on person items on Wikidata when there is no catalog. Please just publish the catalog somewhere and then link to it from your "Black Lunch Table" item. If you don't have a catalog and the project itself is building the catalog, then this property is definitely the wrong way to go. I have tried to read through the material you made available, but I still don't see why this project needs any special property at all when you can create listeria lists from unordered lists of item numbers. If you have a list anywhere on a Wikipedia project, you can also run queries using Petscan.

On Fri, Jan 5, 2018 at 6:12 AM, Brill Lyle <wp.brilllyle@gmail.com> wrote:
First off, thanks so much for the support and assistance in understanding the work being done here. Thanks to those editors who restored the wholesale deletion of the catalog property.

Secondly: While the Black Lunch Table is unique in both its scope and outreach, other projects are using the category in actual "real-world things." They are not internal WikiProjects.

An example of this is the GLAM project for Colección Patricia Phelps de Cisneros (CPPC). CPPC is a large and active private Latin American art collection that is improving coverage of Latin American artists on the projects, with the intention of adding at minimum articles in English, Spanish, and Portuguese. Which is why Wikidata is so helpful, for it's language neutral interchangeability of the scaffolding of metadata and the establishment of notability via VIAF and other identifiers.

The CPPC GLAM project has multiple task lists and SPARQL queries in Listeria tables. CPPC also plans on doing an image donation to the Commons in the next 6-12 months as the project develops and as full metadata is collected and implemented in the most robust Wikidata-centric way. But first the publications and artist metadata needed to be populated.


This is not a WikiProject, but is a GLAM (Galleries, Libraries, Archives, Museums) initiative.

So should GLAM outreach and Wikipedia:Meetup projects like BLT have something specially created to cover this type of outreach?

I believe that other potential outreach institutional partners would want to be implementing usage of Wikidata in this way as well. With this Wikidata-in-a-Box approach, the idea is to expand and improve upon common outreach requirements (like task lists), setting up a replicable structure and process that reduces administrative burden and doesn't require re-inventing the wheel over and over again. Because the fact is that there is definitely an exponential need for this work -- and this need is only going to increase and expand in scope, hopefully. As long as things like what happened here don't happen again and discourage this work and destroy outreach efforts. 

So it would help to have consensus of some type to support this outreach going forward.

Gerard and I thought we had consensus on this, but apparently not. We need to find some solution that will address all concerns.

Thanks again,

- Erika

Erika Herzog
Wikipedia User:BrillLyle

On Thu, Jan 4, 2018 at 11:21 AM, James Heald <jpm.heald@gmail.com> wrote:
Better to use P4570, or a new bespoke property, since the things these people are being tagged to be part of, or participants in, like "Black Lunch Table", are not external real-world things, but internal wiki-world projects.

It is useful to maintain a distinction between the two -- it helps to avoid the confusion that has been the root of the issue with P972.

 -- James.



On 04/01/2018 16:10, Thad Guidry wrote:
"relatedness" or "tagging" is typically handled generically in Wikidata
through the use of "part of" and "has part" properties.
They work terrifically well to apply some generic classification needs such
as those of the Black Lunch Table efforts.

So, an alternative to the current modeling could be...

Are they only persons ?  if so, mark them as "participant of" ->
"Q28781198" Black Lunch Table
Are the topics needing some "tagging" for classification sometimes more
than persons ?  if so, mark them as "part of" -> "Q28781198" Black Lunch
Table

-Thad



_______________________________________________
Wikidata mailing list
Wikidata@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata



_______________________________________________
Wikidata mailing list
Wikidata@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata



_______________________________________________
Wikidata mailing list
Wikidata@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata



_______________________________________________
Wikidata mailing list
Wikidata@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata



_______________________________________________
Wikidata mailing list
Wikidata@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata