Wikidata talk:Requests for comment/Improving Wikidata documentation for different types of user

Please leave your thoughts, suggestions for additions below, we will work to synthesize it and integrate it into the main page. To start a new subject click Add Topic at the top of the page. It would be extremely helpful if you could provide some information of your background when leaving comments so we can understand the needs of different kinds of users e.g "I'm a long time Wikipedia contributor who has heard of Wikidata but never edited it before".

Thanks

--John Cummings (talk) 13:18, 18 May 2018 (UTC)

What kinds of users exist?

Please add the kinds of users that exist below, I'm sure there are many that haven't been considered or have only been touched on broadly

• Wikipedia editor wanting to contribute data to Wikidata manually
• Wikipedia editor who wants to reuse Wikidata data on Wikimedia projects
• Wikipedia user who wants to import datasets into Wikidata
• Database expert
• Partner organisation with data
• Partner organisation with subject matter expertise
• Wikidata outreach volunteer
• Academics and researchers
• Teachers and students
• Data reusers
• Resource creator
• Wikimedian in Residence

Thanks

--John Cummings (talk) 17:53, 24 May 2018 (UTC)

Wikimedia Commons contributors

For Commons: explain c:COM:Structured Data to us. (like we're five) - Alexis Jazz (talk) 13:52, 18 May 2018 (UTC)

slides c:File:Structured Data on Commons - What changes are coming? - presentation at Wikimania 2017.pdf

video starting at 33:00 [1] -- Slowking4 (talk) 19:13, 18 May 2018 (UTC)

Wikipedia editor wanting to contribute data to Wikidata manually

In terms of this user-type, personally I (a subject-specific user focusing on sports) have struggled with lack of documentation for three things:

1. Inputting tools When you are focused on a specific knowledge area you often find yourself repeatedly manually inputting the same things. The process lends itself to semi-automation. I forked User:Magnus Manske/wikidata useful.js to do ease my own editing, but in no way should that level of expertise be required to merely have a way of filling out frequently used data. There should be documentation to assist people with making such things. Ideally, it would be hugely enabling for manual input users (i.e. general users) if there was a global tool with an easily customisable interface to enhance the input experience.
2. Ontologies There doesn't appear to be any documentation or tools for mapping out a data ontology for a given subject. I find the design of Wikipedia Infoboxes somewhat helpful as a starting point. If any data project is to be useful then similar data needs to be input in a similar way. I don't see any way to build or share documentation on that. Visual mapping would be really useful here. The lack of framework for how we do ontologies is a barrier to both manual and automated growth of Wikidata. The property pages should have clearer information on how they are used, what data goes in there, and which ontologies they are found in. Similarly, where a data item is common part of an ontology (e.g. academic conference (Q2020153) would be in a conventions ontology) then there should be a way to note that on the item.
3. Wikipedia<-->Wikidata Users need clear documentation on (a) how to easily get structured Wikipedia data loaded onto Wikidata, and (b) how Wikidata can then be drawn down onto Wikipedia. I am an IT professional who works with databases every day and writes technical documentation – if I'm struggling to get a grasp of these things then god knows how that looks like to other users (magic?). For (a) I see tools with an easy GUI being the best solution for that (e.g. HarvestTemplates). We need documentation on how to both use and build such things. For (b) we need better documentation on both Lua coding and its applications. That part is always going to be quite technical, but there is no reason why most of that task can't be simplified into generalised Module calls, providing documentation for an example ontology that one can then build upon.

The above points are pretty technical for a manual input user, but I think we're at the point where this role still requires a high level of expertise to undertake. Greater documentation and education will build the path to make this role one which is no more threatening to edit than Wikipedia (I'd like to aim for easier!) Sillyfolkboy (talk) 20:56, 18 May 2018 (UTC)

@Sillyfolkboy:, thanks so much for this, its very helpful indeed :) --John Cummings (talk) 06:24, 19 May 2018 (UTC)

The document covers a lot of ground, but in the practical world it is very much a case of missing Inputting tools, as Sillyfolkboy describes it. When you upload a fair amount of artworks to Commons, it is simply too irritating to have to repeat and repeat the same steps in Wikidate to register the artwork. The lack of tools for either a copy-paste from a similar artwork, or some sort of connection between the data in Commons and in Wikidata, means that you have to manually repeat your inputs. And that means that a lot of artworks (and other items) simply do not get registered in Wikidata. --Rsteen (talk) 16:30, 23 May 2018 (UTC)

eCH-0205 Linked Open Data (Guidelines)

Maybe the following guidelines can provide some orientation: eCH-0205 Linked Open Data. – Obviously, the document doesn't cover all the Wikidata-specific information needed, but it may provide a general overview to people who are new to linked data and Wikidata. I would be glad to receive some feedback as to what points the document covers well (from the point of view of which target users?) and where it could benefit from improvement. The document has been published under a CC-by license. Thus, it can be forked and further developed or we can try to develop it together in view of the publication of version 2 (in the latter case, the Switzerland-specific stuff could probably be moved somewhere else). --Beat Estermann (talk) 16:08, 19 May 2018 (UTC)

Commons, Wikisource

Where you have the various "Wikipedia editor..." journeys, I think Commons & Wikisource users probably have distinct journeys, which also merit some thought. Given the current efforts at structured data on Commons, that journey is particularly relevant at the moment. (Probably true of WikiTravel, too, I just don't know much about that.) - Jmabel (talk) 04:58, 21 May 2018 (UTC)

Yes absolutely @Jmabel:, I'm not sure what these users would want to do with Wikidata, if anyone reading this is a contributor to Wikisource, Commons etc please do tell us (I've asked on the village pumps). --John Cummings (talk) 11:55, 21 May 2018 (UTC)

Well, with Commons, at least, there seems to be a major integration of the two under way. There were four sessions about that at the Wikimedia Conference. & it seems to me that many of the same considerations should eventually apply for Wikisource, and the only reason this isn't moving yet is that Wikisource is a bit small and something of an orphan. - Jmabel (talk) 15:04, 21 May 2018 (UTC)

Useful information in event write ups

There is useful information for this RFC in this event writeup for a Wikidata event in Cologne. Does anyone know of any other event writeups that could be use to understand what different kinds of user want to do on Wikidata? --John Cummings (talk) 08:36, 22 May 2018 (UTC)

Bloom’s Taxonomy Verbs

Bloom’s Taxonomy Verbs may be useful in thinking about what learning objectives people have.

John Cummings (talk) 08:42, 22 May 2018 (UTC)

Is this a RFC?

1. This is a massive document which could be a programme outline based on its structure
2. It appears to rely on an analysis or model for something like a "stakeholder journey", yet the RFC is asking for comment on topics
3. RFCs are conventionally for proposals
4. The document leaves implicit the issue(s) it aims to solve, and there is no explanation of why the chosen breakdown is accurate and optimal, nor is the expected outcome clear (which might be a proposal for a project)

--Fæ (talk) 13:59, 18 May 2018 (UTC)

Thanks @Fæ:

1. Yes it could be an outline for a programme if someone wanted to pick it up as one (I guess it could for a grant request even)

2. Yes, you're right, that needs to be clearer, I want to ask for both types of user and feedback from users on what they need.

4. Can you explain a little more what you mean by The document leaves implicit the issue(s) it aims to solve? I based the breakdown of user types on my own experience, I'm hoping people will suggest additions and corrections. I need to make this clear also.

Thanks again

John Cummings (talk) 14:58, 18 May 2018 (UTC)

I agree - this does not look like the Wiki ideal of an RfC to me. We do not have a process for getting comment on half-finished works. Usually when I see a document like this it is well planned by some group which already has a proposal for a direction in development and practice. There are lots of suggestions here but there is no one in the world who can recommend the best way to provide all this training. Most of these use cases probably have never been imagined outside the Wikidata context.

I see no problem with presenting this as an RfC in the sense that there is a lot of work in this and it is relevant enough to enough groups that I think anyone who is likely to be in the path of the RfC notice is likely to be a stakeholder in the direction of development for this document.

That said - wow, this is complicated. I am seeing that it imagines about 100 variations of documentation. All of it seems urgent to create soon and all of these might have high demand from groups which have a higher-than-normal-for-wiki likelihood of generating power users. I do not know what to make of this. Blue Rasberry (talk) 16:37, 18 May 2018 (UTC)

Thanks @Bluerasberry:, you're right this isn't the normal RFC, but as you say there doesn't seem to be another process this could take place in, it feels like an important conversation to have. I think the impression that its partly finished because I've made it a bit pretty, I'm just trying to provide a framework for others to feed into (much of what I've added is things I've learnt through conversations). I'm starting this RFC to avoid many of the issues that English Wikipedia has very technical and granular instructions and policies with no clear overall path through them, something which I guess (now you point out this isn't a normal RFC) is an outcome of the processes available being limited to high granularity rather than allowing overviews like this.

I think that variations will probably be needed in some circumstances but a basic document which fulfils most needs of most people feels possible, they will just appear on multiple kinds of user's journeys. I'm sure this kind of one size fits most documentation is more true for some areas (e.g an introduction to Wikidata) than others (e.g how to import data into Wikidata).

John Cummings (talk) 11:01, 19 May 2018 (UTC)

Rotate tables

Hi, I suggest rotating the tables 90deg, so that the short-list is along the top, and the long-list down the vertical. That would solve the horizontal-scrolling issue on thinner windows/monitors/mobile. Cheers, Quiddity (talk) 14:56, 18 May 2018 (UTC)

@Quiddity:, I agree this would be helfpul, do you know an easy way to do this without having to rewrite everything by hand?. Thanks, --John Cummings (talk) 14:58, 18 May 2018 (UTC)

@John Cummings: Sadly not. I took a look at doing it in google sheets, and it can be rotated, but won't retain the formatting. Ditto for using VisualEditor cut&paste. So editing with wikitext manually is the only immediately available method. Quiddity (talk)

@Quiddity:, I'll try and get round to it soon, it will be more important as the page gets filled up I guess. --John Cummings (talk) 07:08, 19 May 2018 (UTC)

So is what you want basically like this?

Learning goals	Documentation needed		Tools needed
	Existing resources	New resources	Existing resources	New resources

(if so, let me know, that should be pretty simple; if not, let me know what's different, feel free to do so by editing the above.) - Jmabel (talk) 05:11, 21 May 2018 (UTC)

@Jmabel: oh! I spontaneously decided to just fix it, after opening the page again, but I didn't check the talkpage here. Hence I missed your question. I did the rotation manually, but am curious what technique you would've used (?), so that I can potentially learn tricks for the future. Thanks!

I'll also ping @John Cummings: so that he can comment with his preference based on your split-column proposal above. Although I suspect that 5-column layout might be more complex than needed, especially given the width-restraints from those large graphics to the left of each section... Hmm, maybe if we move those graphics up, so that they're above the tables... {done} but revert freely. :) Quiddity (talk) 23:14, 28 May 2018 (UTC)

It also would have been basically manual, though using some patterns for cut-and-paste that I've noticed facilitate this, but can't easily describe (sorry). I was mostly just trying to confirm I understood the intent before proceeding. I see you've done it differently than I was suggesting, but if John feels your solution meets his intent, that's what's important. - Jmabel (talk) 07:06, 29 May 2018 (UTC)

Thanks very much both, I've adapted them to include numbers to each table in the attempt to describe the stages of the process of learning the skills. --John Cummings (talk) 13:03, 30 May 2018 (UTC)

Learning goals		Documentation needed		Tools needed
Order	Description	Existing resources	New resources	Existing resources	New resources