Wikidata

Wikidata:Requests for comment/Reforming the property creation process/eo

< Wikidata:Requests for comment
This page is a translated version of the page Wikidata:Requests for comment/Reforming the property creation process and the translation is 10% complete.
Other languages:
An editor has requested the community to provide input on "Reforming the property creation process" via the Requests for comment (RFC) process. This is the discussion page regarding the issue.

If you have an opinion regarding this issue, feel free to comment below. Thank you!


THIS RFC IS CLOSED. Please do NOT vote nor add comments.

The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.

Closing this RFC with the following results:

  • Creating properties counts as an administrative action, for the purposes of fulfilling the activity requirement. (# Admin activity)
  • Users should not create properties when they have proposed that property. (#Recusal (3))
  • Only users who have the ability to create properties should close property creation proposals, i.e. setting status to done, not done or withdrawn. (#Discussions)
  • Property creator right must be only granted following a request on Wikidata:Requests for permissions/Other rights which must remain open for at least 48 hours. An exception may be made for the case of regranting of the right to former property creators and administrators where the reason for removal was uncontroversial or due to inactivity. (#Granting the right; #Regranting)
  • Users creating properties are expected to reply promptly and civilly to concerns about their actions. Repeated failure to do so may result in removal of the right, subject to the usual procedures. (#Accountability)

--Pasleim (talk) 18:32, 13 April 2016 (UTC)[reply]

This is a RFC to reform the property proposal process. Some of these proposals explicitly clarify community practices that already exist. Others serve as suggested improvements for the process.

Admin activity

For the purposes of fulfilling the activity requirement, creating properties will count as an administrative action. This will encourage administrators to participate in the property creation process.

Recusal

Recusal (1)

Users should not create properties when they have taken part in the discussion for that property.

  •  Support obvious. --Rschen7754 06:31, 20 March 2016 (UTC)[reply]
  •  Emot It would risk stall the PP-process to much. -- Innocent bystander (talk) 06:33, 20 March 2016 (UTC)[reply]
  •  Support We have enough property creators; a third voice can't hurt and it will encourage outside opinions.--Jasper Deng (talk) 06:39, 20 March 2016 (UTC)[reply]
  •  Oppose It's good to read that some otherwise only marginally active admins might be more likely to participate in the property creation proposal, if it's counted "as admin activity", but I doubt that this will substantially increase participation in a process that is already slow. --- Jura 06:42, 20 March 2016 (UTC)[reply]
  •  Oppose I'm actually opposed to this; in clear-cut cases or cases with little interest, I'm not very concerned with involved users making the property. What I do want is for those users to be a) able to justify the creation, b) not hostile when people raise issues with the creations, and c) capable of collaboratively move forward when issues are raised. Ajraddatz (talk) 07:28, 20 March 2016 (UTC)[reply]
  •  Support To avoid possible COI. --Succu (talk) 08:43, 20 March 2016 (UTC)[reply]
    {{s}} I feel like some discussions would have been decided differently, so this seems neccessary atm. Though this restriction may propably be removed again once WD has grown up a bit. --Nenntmichruhigip (talk) 09:31, 20 March 2016 (UTC)[reply]
  •  Weak support although I would prefer this to be phrased less strongly as there are exceptions that will be uncontroversial. Unless the need is urgent they should always attempt to get the attention of others before creating the property themselves, but should that not be successful it should be possible to go ahead. Thryduulf (talk: local | en.wp | en.wikt) 10:30, 20 March 2016 (UTC)[reply]
  •  Oppose It's often necessary for a property creator to first ask for clarification, or a translation, or to notify a proponent of an omission or error in the proposal. Even where someone has !voted, they should be able to close the proposal if there is clear consensus for or against, or if they concede that the opposing view has it. [I write from the PoV of someone who has created well over 350 extant properties]. Note that this is not a "community practice that already exists". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:52, 20 March 2016 (UTC)[reply]
  •  Oppose This is the wrong way around; I understand there is a risk of COI, but there are other means to address that. But the downsides (stalling, people deciding with less knowledge) are worse than the benefits (no COI). Mind you, even with the rule COI occurs: a user that joins the discussion late noting a favorable discussion for his POV, and then moves forward without participating in the discussion. (Sounds like a conspiricy? Yeah, pretty much like users that do participate in the discussion not taking a neutral perspective in the final decision.) Egon Willighagen (talk) 11:14, 20 March 2016 (UTC)[reply]
  •  Support it could work if there are enough property creators. Pamputt (talk) 13:36, 20 March 2016 (UTC)[reply]
  •  Oppose The property creation process is often stalled because there is not enough participation, that would make people even less likely to participate.--Micru (talk) 13:48, 20 March 2016 (UTC)[reply]
  •  Weak oppose That is too strict. User should not create properties if he proposed its creation (or was its sole or one of few supporters), but taking part of wide discussion does not sound as obstacle to creating properties. --Jklamo (talk) 14:04, 20 March 2016 (UTC)[reply]
  • Oppose, per Ajraddatz. --Vogone (talk) 18:29, 20 March 2016 (UTC)[reply]
  •  Oppose Too strict. --Fralambert (talk) 19:42, 20 March 2016 (UTC)[reply]
  •  Oppose too strict - it could prevent PC who are interested in a property to participate in a discussion - that would be counter-productive. Maybe if they are the contributor who proposed the creation, but participating is much too vague. --Hsarrazin (talk) 20:30, 20 March 2016 (UTC)[reply]
  •  Oppose counter-productive. Prevents anyone who asked for details on a proposal to create the property, or would refrain property creator to engage in the debate by raising valid points. -Ash Crow (talk) 20:54, 20 March 2016 (UTC)[reply]
  •  Oppose per above comments - participation in discussion can be neutral. I would support recusal for the proposer and all supporters & opponents (neither should create nor mark as "Not Done"). Supporters and the proposer should be able to mark a proposal "Withdrawn" at any time however. ArthurPSmith (talk) 21:57, 20 March 2016 (UTC)[reply]
  •  Oppose Currently the creation process lacks both discussion participants and property creators. This would discourage property creators to participate in discussion. Alternatively, if someone has taken part in a discussion (for example with a simple "useful property"), they cannot create it, even if there is clear consent.--Srittau (talk) 02:56, 25 March 2016 (UTC)[reply]
  •  Support per Jasper Deng. Jianhui67 talk★contribs 08:33, 5 April 2016 (UTC)[reply]
  •  Oppose In a lot of cases there is no problem. For instance in case of consensus among several participants. Or if one creates a property one voted "oppose", or rejects a property one "supported" in discussion. Or if one asked good questions, raised good points, without conlusion. Besides, the number of people involved is often nog large. Lymantria (talk) 16:52, 6 April 2016 (UTC)[reply]

Recusal (2)

Users should not create properties when they have a link to the organization issuing or related to the property.

Recusal (3)

Users should not create properties when they have proposed that property.

Recusal (4)

Users should not create properties when they have supported or opposed a property.

  •  Oppose See Recusal #2. --Srittau (talk) 03:02, 25 March 2016 (UTC)[reply]
  •  Support Actually I think a creator acting contrary to their stance in discussion (i.e. creating a property despite earlier opposing it, or closing a proposal as not done despite previously supporting it) would be fine. The other direction is the problem. There's an important element of judgment needed in property creation and I think anybody that has expressed any bias in the process should not act according to that bias to ensure there is a legitimate consensus in play. ArthurPSmith (talk) 15:51, 25 March 2016 (UTC)[reply]
  •  Oppose If a proposal has no objections, or clear consensus, why would this be an issue? Note that this is not a "community practice that already exists". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:55, 25 March 2016 (UTC)[reply]
who is doing the judging of whether there is a "clear consensus"? It should not be an involved party, in my view. Even if it seems obvious. ArthurPSmith (talk) 16:31, 25 March 2016 (UTC)[reply]

Recusal (5)

Users should not create or decline to create properties when they have taken part in the discussion for that property, and there are outstanding objections to the proposed outcome.

Diskutoj

Only users who have the ability to create properties should close property creation proposals (successfully or not).

  •  Support Rschen7754 06:31, 20 March 2016 (UTC)[reply]
  •  Tveksam I think we should encourage more activity in these pages, not add more red tape. -- Innocent bystander (talk) 06:39, 20 March 2016 (UTC)[reply]
  •  Support Property creators are (at least in principle) vetted by the community as users who can evaluate community consensus well. So anyone who can do so with property proposals should be encouraged to have the property creator right.--Jasper Deng (talk) 06:41, 20 March 2016 (UTC)[reply]
  •  Comment in practice, this would mean that only property creators and administrators could close and archive stale stuff. I don't think I have seen that many administrators archiving proposals and some property creators already tend to forget to archive discussions they closed, I'm not sure if this is going to lead somewhere ..
    --- Jura 06:52, 20 March 2016 (UTC)[reply]
  •  Neutral On the one hand, I absolutely hate the system of non-admin closures happening on enwiki, which is part of their drawn out and useless process for vetting admins. We don't need that type of ridiculous and worthless bureaucracy here, which somehow further separates admins from regular users. Anyway, on the other, I think that any help clearing out stale old proposals would be worthwhile, so neutral. Ajraddatz (talk) 07:31, 20 March 2016 (UTC)[reply]
  •  Comment What's the current procedure? Everyone can close, and if it's closed as to be created admins have to follow? Or is it everyone can close unless an admin would have to create the property? --Nenntmichruhigip (talk) 09:31, 20 March 2016 (UTC)[reply]
  •  Weak support with the exception that anyone may close their own proposals as unsucessful or withdrawn, and anyone may formally mark a proposal as withdrawn if it has been withdrawn but the status not set to such. Thryduulf (talk: local | en.wp | en.wikt) 10:30, 20 March 2016 (UTC)[reply]
  •  Oppose Anyone should be able to close a proposal if there is clear consensus against, or of the property requested already exists. (If there is clear consensus for, they should mark it as "ready", so that someone with the PC ability may take the necessary action). Note that this is not a "community practice that already exists". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:52, 20 March 2016 (UTC)[reply]
  •  Support Pamputt (talk) 13:36, 20 March 2016 (UTC)[reply]
  •  Support--Micru (talk) 13:51, 20 March 2016 (UTC)[reply]
  •  Support --Jklamo (talk) 14:08, 20 March 2016 (UTC)[reply]
  •  Neutral I don't have a strong opinion on this. I find it kinda weird when non-admins/property creators reject proposals, but I don't have a problem with proposers withdrawing their own proposals or other people marking proposals as withdrawn or done if they weren't already. - Nikki (talk) 17:28, 20 March 2016 (UTC)[reply]
  • Oppose, it doesn't matter who closes such requests. Obvious false decisions can be reconsidered by a second person if needed. --Vogone (talk) 18:31, 20 March 2016 (UTC)[reply]
  •  Oppose I don't see the problem of a closure of property proposal by a user. At worst, we can simply reopen it. --Fralambert (talk) 20:12, 20 March 2016 (UTC)[reply]
  •  Comment I think there is some confusion evident in this recommendation. I rarely see property proposal discussions "closed" in the sense of using the Closed template or something of that sort. What generally happens is (A) an editing of the status field for the property to indicate either that the property has been created (entering the property number in that field), or some other status change such as "withdrawn" or "not done", and then at some time later (B) archiving the proposal to the relevant archive page. Changing the status to "ready" should indicate that all the property documentation fields are filled in (good examples, related item, formatter, etc) and consensus seems to have been reached in discussion and the property should be created. Anybody should be able to do B (archiving) - and it would be nice if a bot took care of it so humans didn't have to be involved at all. There's no careful evaluation needed for archiving, though I think it's best to leave a space of a few days before archiving to allow for any post-creation (or post-"not done") comments from participants. On (A) - the property creator should enter the newly created property's number in the status, so that pretty much has to be done by a property creator (it is Step 4 in the creation process). "withdrawn" should probably be entered only by the proposer though there are some cases where it seems obvious and anybody should be able to (for example the proposer adds a new replacement proposal on the same proposal page, but forgets to mark the first as withdrawn). I believe the main issue of concern is who should enter "Not Done". as a status. I would recommend that this be restricted to uninvolved property creators or admins, and ONLY on the basis of either * lack of consensus after significant discussion (roughly equal oppose vs support comments) OR * a complete lack of supporting comments after an extended period of time (whether or not there are any opposing comments). This has not perhaps been completely clear in the Guidelines but I would have interpreted a lack of supporting comments over an extended period of time as manifestly lacking in consensus - as it says now, "Property creators should not create new properties unless consensus exists.". ArthurPSmith (talk) 22:45, 20 March 2016 (UTC)[reply]
    • There is a long-standing convention that "Consensus is assumed when there's no evidence of disagreement." Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:54, 21 March 2016 (UTC)[reply]
      • This is an important point - could other admins and property creators please weigh in on how they have interpreted "unless consensus exists" in the context of property creation? I note for example that Micru closed this discussion as "Not Done" with the note "Lack of Support" after several months and no comments of any sort. I assume Andy would have created the property? But we have many others currently in a similar state with no comments or supporting or opposing statements - should they be created, or archived as Not Done for lack of support? ArthurPSmith (talk) 11:32, 21 March 2016 (UTC) -- addendum - Andy seems to be referring to en:Wikipedia:Silence_and_consensus but note that it states "Silence and consensus does not apply when either a strong consensus or a mandatory discussion is required." - how does this apply in the case of property creation? Since it's a limited right and there are very good reasons for reasonably restricting the number of properties in Wikidata, it has seemed to me the requirement of consensus is something more than simple silence here. How do others feel? ArthurPSmith (talk) 11:55, 21 March 2016 (UTC)[reply]
        • How many cases have there been, of properties being deleted, or at least nominated for deletion, or issues raised at project chat/ admin noticeboard, because they were created after a "no evidence of disagreement" type decision? Some examples would also be helpful, as well as an indication of the level of disruption caused. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:19, 21 March 2016 (UTC)[reply]
          • No answer. Presumably none, then. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:14, 25 March 2016 (UTC)[reply]
            • Andy, look at the exasperation over on project chat about the large number of barely used properties, look at the long list of potential external identifiers that we've been slowly slogging through, look at the properties that *have* been deleted. Were many created without much discussion? I believe so but I'm not going to go digging up details for you. Unfortunately I have two major projects that just landed on my lap yesterday so I'm probably not going to have much time for wikidata stuff at all for a while. Please don't assume my lack of response means agreement. What we really need is a better consensus on the purposes behind property creation in the first place - what are we actually trying to do here? Structure data, yes. But we are trying to avoid the mess that dbpedia and freebase got into with huge numbers of unconnected statements. Good properties connect things together wonderfully. How do we ensure that we limit the created properties to ones that will be long-term useful? ArthurPSmith (talk) 16:39, 25 March 2016 (UTC)[reply]
              • "Exasperation"? There are a couple of editors making trouble a fuss over nothing, regarding a few properties for which plenty of data exists in Wikipedia or in external sites; but that has nothing to do with either this proposal, nor my question. I don't assume your lack of response means lack of agreement, far from it. It just shows a lack of evidence. Just as there is no evidence of your hypothetical unconnected statements or lack of long-term use with regard to the issues covered by this proposal. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:26, 25 March 2016 (UTC)[reply]
                • Andy, what is wrong with waiting to create a property until it will actually be used? There are clear benefits (as the things I mentioned above indicate) to keeping the number of properties relatively small. External ID's may be an exception to that going forward as long as they aren't things that may change or require maintenance, but even there I don't see a whole lot of point in creating a property and then waiting years before somebody uses it for more than 1 or 2 items. Presumably there is a reason why property creation is limited to certain trusted users. Maybe we can be clearer about what the underlying purpose is for that. ArthurPSmith (talk) 20:43, 25 March 2016 (UTC)[reply]
  •  Support Jianhui67 talk★contribs 08:38, 5 April 2016 (UTC)[reply]
  •  Support Lymantria (talk) 17:01, 6 April 2016 (UTC)[reply]

Exception

After 2 months of inactivity in the discussion, the property proposal can be archived by anyone.

  •  Support Sometimes we have old discussions that haven't reached any conclusive outcome and they remained inactive for months. Those discussions should be archived by anyone, preferably a bot.--Micru (talk) 14:07, 20 March 2016 (UTC)[reply]
  • Neutral. However, if passed, this should not include discussions with consensus for creation. A valid property proposal should not be allowed to fail through lack of timely volunteer effort to enact it. If this exclusion is not included, read this as an objection. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:26, 20 March 2016 (UTC)[reply]
  •  Oppose I think this would be problematic. Lots of proposals see no activity for months right now and saying that anyone can archive them after two months of inactivity implies that the proposal automatically fails after two months of inactivity. We need to get more people involved with proposals first. - Nikki (talk) 17:28, 20 March 2016 (UTC)[reply]
  •  Oppose Will cause a waste of time when those properties have to be re-proposed.--Jasper Deng (talk) 19:44, 20 March 2016 (UTC)[reply]
  • I think 2 months is too soon. --Rschen7754 19:52, 20 March 2016 (UTC)[reply]
  •  Oppose I don't think we have problem of a too long list of inactive proposal. --Fralambert (talk) 20:16, 20 March 2016 (UTC)[reply]
  •  Oppose there are many properties that wait longer than this before being created (e.g. in early February I marked some proposals as ready that had had no objections since November, they were created today). It shouldn't take that long but that is a separate problem that this will preclude fixing. Thryduulf (talk: local | en.wp | en.wikt) 22:10, 20 March 2016 (UTC)[reply]
  •  Oppose - see my comment just above - "Not done" should be added as status with an explanatory note pinging the proposer and other discussants, then after at least a few days to allow for further comments the archiving can be done by anyone. ArthurPSmith (talk) 22:48, 20 March 2016 (UTC)[reply]
  •  Emot - Can somebody explain how the archives works here? -- Innocent bystander (talk) 08:11, 21 March 2016 (UTC)[reply]
    • Property proposal discussions that are finished (Status in the state of Withdrawn, Not Done, or given a property number) are cut-and-pasted by a willing editor to the current Archive page at some point after the discussion has concluded. It's a bit ad hoc. ArthurPSmith (talk) 11:38, 21 March 2016 (UTC)[reply]
    • Further to Arthur's reply (which is correct) note that I recently changed the guidelines to reflect that practice; until then they said that proposals should be archived immediately an item was created, which was both generally ignored and not sensible. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:44, 21 March 2016 (UTC)[reply]
      • Sadly, my changes to that page have been mass reverted, and - among a number of errors and stale suggestions reintroduced - they now say that archiving should be immediate. See its talk page for more. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:13, 22 March 2016 (UTC)[reply]
        • I tried to fix things up (Jura broke some other fixes with his revert) but Jura has re-reverted my changes also. Please see a draft page here: User:ArthurPSmith/Draft:Property_Creators with changes from Andy and myself to better reflect the current process, and from Matej to fix some translation issues. Note I have also removed "Step 7" on adding to the list of properties as it shouldn't be necessary and has definitely not been done consistently recently. Jura seems to think we need approval from every property creator and administrator before making these changes - I would have thought the normal wiki editing process was sufficient to reach consensus on this, but apparently more is required? What do people suggest for moving forward with this? I note there are important issues of clarification of process that this discussion here has surfaced that should perhaps also be reflected on this page. ArthurPSmith (talk) 14:03, 22 March 2016 (UTC)[reply]
  •  Oppose This is a very bad idea: ignore and the proposal dissapears. Lymantria (talk) 17:11, 6 April 2016 (UTC)[reply]
  •  Oppose All proposals should end in an active decision to create or not create, not a "fade into obscurity". --Srittau (talk) 03:05, 25 March 2016 (UTC)[reply]
  • I think we also need to pay some respect to the proposer. Closing or not closing is not the ultimate goal of everything, it's a part of the process. So the proposer should have a right to comment the closure and answer question like "Is my problem solved ?" This would help us to be contributor centric, and not only closer centric. author  TomT0m / talk page 08:45, 25 March 2016 (UTC)[reply]

What counts as closing a discussion?

From the above it seems that not everybody has the same view about what actions constitute closing a discussion. For clarity in interpreting any consensus produced above, please support if you think the listed action does or should close a discussion, please oppose if you think it does not or should not. Thryduulf (talk: local | en.wp | en.wikt) 02:06, 21 March 2016 (UTC)[reply]

setting status to "done"

setting status to "not done"

setting status to "ready"

setting status to "withdrawn"

adding comment "done"

  •  Support

adding comment "not done"

adding comment "ready"

adding comment "withdrawn"

Discussion of 'What counts as closing a discussion?'

Eeek; that's a lot of headers! I think that marking properties as "ready" shouldn't count as closing, but everything else should be (done, not done, withdrawn). From what I know, that's how it's currently done. Ajraddatz (talk) 03:38, 21 March 2016 (UTC)[reply]
That's clear regarding setting the status, but what about comments that don't set the status? Thryduulf (talk: local | en.wp | en.wikt) 14:43, 21 March 2016 (UTC)[reply]
Not much difference I think. If someone comments as opposed to change the status, then anyone else should be able to go and change the status to reflect the closure. Ajraddatz (talk) 16:38, 21 March 2016 (UTC)[reply]
@Thryduulf: - are you responding to my comment? I don't think this is helpful - what I would support would be distinct rules on each type of closing (for "Withdrawn", the proposer should generally do it themselves. "Done" or "Not Done" should be limited to property creators or admins. "Ready" could be done by anybody, as could archiving after the discussion is closed. ArthurPSmith (talk) 11:41, 21 March 2016 (UTC)[reply]
I was responding to the whole section above at the time I started this subsection not to any one comment. Thryduulf (talk: local | en.wp | en.wikt) 14:43, 21 March 2016 (UTC)[reply]

Granting the right

The property creator right must be only granted following a request on Wikidata:Requests for permissions/Other rights. The request must remain open for at least 48 hours to allow for the community to comment on the request. A request will be successful as long as no substantial objections are raised that are supported by consensus.

Regranting

(Only valid if the first proposal passes) An exception may be made for the case of regranting of the right to former property creators and administrators where the reason for removal was uncontroversial or due to inactivity.

Accountability

Users creating properties are expected to reply promptly and civilly to concerns about their actions. Repeated failure to do so may result in removal of the right, subject to the usual procedures.

Diskuto

Cart before the horse

I'm glad to see the community are now discussing these issues, I would point out that my PC-role has just been removed, for breaches of some of the new "rules" proposed here, which did not exist at the time of my actions. Had they done, so, I would of course have been willing to adhere to them. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:03, 20 March 2016 (UTC)[reply]

Property planned use

I propose to add a new field to the property proposal template stating the planned use for the property, that way the property creators can know if it will be used to import data straight away, if the user plans to add the data manually, if it forms part of a wikiproject, or if it will become a dormant property. The creation of properties with an active supporter/shepherd could be prioritized.--Micru (talk) 14:03, 20 March 2016 (UTC)[reply]

I don't believe this is necessary. If a delay is holding someone up they can ping a PC (as people occasionally did me, and I was happy to respond as soon as possible). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:16, 20 March 2016 (UTC)[reply]
@Micru: I have always interpreted the "robot and gadget jobs" field in the Property Documentation template to be exactly for that purpose - to indicate if the property proposer (or others who have edited the proposal) has a plan to fill in or otherwise use the property values. I don't think we need a new field but perhaps that one should be renamed or clarified? ArthurPSmith (talk) 11:44, 21 March 2016 (UTC)[reply]
I've interpreted that field as indicating whether there is a potential for the use of automated tools, not whether I personally have plans to use automated tools (I don't generally use gadgets and lack the skills to run a bot. Thryduulf (talk: local | en.wp | en.wikt) 14:48, 21 March 2016 (UTC)[reply]
We should avoid creating the false impression (one which a handful of editors already attempt to impose) that a property must be used in volume within a short time of creation. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:34, 23 March 2016 (UTC)[reply]

Some open discussion (mostly regarding recusal)

I want to open discussion down here. Some observations:

  • The process we call Property proposals won't work with "you can't act on these proposals if you've discussed them/have an external COI". We actually need more users with the latter issue, as they are the ones best able to decide a property's fit-for-purpose.
  • A general inactivity in regard to proposals is the predominant problem "causing" people to close their own proposals (as successful). Forcing a recusal does not fix that problem.
    • A time limit on a discussion may help, where if consensus is not obvious after some period of time, the discussion should simply be closed as unsuccessful.
  • I generally agree with @Ajraddatz, Thryduulf, Jklamo: at #Recusal.

Other comments welcome on the general issue of recusal. --Izno (talk) 17:06, 20 March 2016 (UTC)[reply]

Regarding your second and final points, I reiterate that I only ever closed my own proposals once I had asked it if was OK to do so at project chat. If anyone objected subsequently, all they had to do was to get consensus at that (or some other suitable) venue. If they objected in a specific case, for example because they thought I had acted against consensus in the relevant discussion, the option of starting a deletion discussion on that basis was open to them. This never happened. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:11, 20 March 2016 (UTC)[reply]
Though I understand the temptation, don't make this about you. You can probably make the same discussion points (general agreement) without using yourself as a point. --Izno (talk) 22:15, 20 March 2016 (UTC)[reply]
Given that I am referred to, albeit obliquely, several times in others' comments (some dishonest and ad hominem) on this page, most pre-dating mine, it is not me to whom you need to address that comment. I am also, as I noted above, able to speak with the experience of someone who has crated over 350 properties in the last 13 months. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:02, 21 March 2016 (UTC)[reply]
Andy: I did not refer to you. Do not be tendentious, please. If you have input on this particular discussion which doesn't involve your specific case, please provide it, but otherwise, move along. --Izno (talk) 12:56, 21 March 2016 (UTC)[reply]
I didn't say that you did; I specifically said (emphasis added) "on this page". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:01, 21 March 2016 (UTC)[reply]
  • Izno - Two points: (1) I don't think the creation guidelines are clear enough (see my comments above particularly with regard to the definition of consensus - is "silence" consensus in this case?). With some more clarification I and I expect other property creators would act more quickly on proposals that actually are ready, or help with those that seem to be having trouble getting there. I do regularly review the proposal pages but rarely find something in a state I consider sufficiently ready to act on. (2) the current collection of proposal pages is a mess - it's not clear where certain property proposals belong (imagine for instance an Authority control property for Places in a Natural Science area that consists of standardized terms) and some of them are so long (some proposals having extended discussions spanning years) that it becomes very hard to review and understand what's going on. I think some sort of reorganization of the proposal pages would be helpful, along with some sort of regular cleanup of old discussions and proposals. More clarity in process might need to come first though. ArthurPSmith (talk) 13:36, 21 March 2016 (UTC)[reply]

Reorganization?

  • Izno - [duplicated and trimmed by Izno for section relevance] (2) the current collection of proposal pages is a mess - it's not clear where certain property proposals belong (imagine for instance an Authority control property for Places in a Natural Science area that consists of standardized terms) and some of them are so long (some proposals having extended discussions spanning years) that it becomes very hard to review and understand what's going on. I think some sort of reorganization of the proposal pages would be helpful, along with some sort of regular cleanup of old discussions and proposals. More clarity in process might need to come first though. ArthurPSmith (talk) 13:36, 21 March 2016 (UTC)[reply]
    • I would support the reorganisation of the pages; I share Arthur's concerns in that particular regard, and his example is well made. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:01, 21 March 2016 (UTC)[reply]
      • Agree about reorganisation. I'm mostly interested in those external website IDs, which mainly are at Authority control page, but recently I discovered, that they sometimes are proposed at other pages. --Edgars2007 (talk) 17:54, 21 March 2016 (UTC)[reply]
        • Would organizing by datatype be a good approach? Obviously properties are multifaceted and could be organized many different ways, but properties with the same datatype do tend to have things in common (particularly for special datatypes like geographic coordinates or mathematical formulas). And this has the advantage of being an exclusive feature to organize on - you know exactly where the proposal belongs once you know what datatype the property should use. ArthurPSmith (talk) 18:07, 21 March 2016 (UTC)[reply]
          • Perhaps, but it would make it harder for people with a thematic interest to see all the related proposals; and we often change the proposed datatype during a discussion. A better solution might be a single page (or, say, one per month) with notifications to relevant projects or sub-pages, rather like daily template-deletion discussions on en.WP. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:19, 21 March 2016 (UTC)[reply]
            I think we can still consolidate some stuff. For example, en:WP:WikiProject Council has 5 groupings of WikiProjects, which I've attempted to intersperse below with our groupings:
            • Culture and the arts
              • Organization and politics
              • Creative work
              • Term
              • References
            • Geographical
              • Place
            • History and society
              • Events
              • Organization and politics (duplicated deliberately)
            • Science, technology, and engineering (and mathematics)
              • Space
              • Economics
              • Transportation
              • Natural science
            • Wikipedia (meta)
              • Sister projects
              • Metadata
            The ones remaining include:
            • Generic
            • Authority control
              • I think this one should be merged to the individual other spaces. Even though it's routine, I would prefer to sort "authority control" to their specific domains. Alternatively, we can play around with e.g. putting all identifiers on one page and then everything else to other pages. This would help move the "fasttrack" properties to one page and the slower discussions retained on their current domain-pages.
            • Person
              • 5 identifiers: Merge to individual other spaces; 1 political item; merge to culture and arts
            • Unsorted
              • Sort them, duh. These should be merged to Generic, regardless of any re-org.
            Of course, we can also decide as we're moving them around that certain proposals should be on certain pages, but I think that's the gist of it. --Izno (talk) 11:43, 22 March 2016 (UTC)[reply]
            Izno - maybe, but I'm wondering how you expect "Organization and politics" to work if it's duplicated? Every proposal should be in both places? Or only cultural and arts organizations in the first grouping, everything else in the second? Where would properties related to educational institutions go? ArthurPSmith (talk) 14:23, 22 March 2016 (UTC)[reply]
            @ArthurPSmith: I mean that in the sense of "I had to place this in two buckets on this specific page because I think 'organization and politics' needs to be split between the two". Education I would expect to see in art and culture. --Izno (talk) 14:31, 22 March 2016 (UTC)[reply]
            @Izno: So I want to note that I generally  Support the above proposed rearrangement, however I think the details aren't quite what we need. Right now we have 16 property proposal pages with the following distribution of new properties proposed: Generic (7) Authority control (22) Person (6) Organization and politics (9) Event (0) Creative work (4) Term (2) Space (0) Place (2) References (0) Unsorted (7) Sister projects (3) Economics (25) Transportation (9) Natural science (20) Property metadata (2). Distributing them according to the above suggestion would give just 6 pages with I believe the following distribution of proposals (allocating Authority control and Person and Org & Politics appropriately): "Culture and the arts" 22, "Geographical" 2 , "History and society" 20, "Science, technology, and engineering (and mathematics)" 54, "Wikipedia (meta)" 5, "Generic" 15. I don't think that's a helpful distribution. In particular "Geographical" doesn't seem useful here - maybe merge with History and society? And the Science one is way too big, we'll be hitting Lua errors quickly. So I would split that. Maybe Generic and meta can merge also. 6 pages with around 20 proposals on each would be perfect. ArthurPSmith (talk) 20:58, 7 April 2016 (UTC)[reply]
            What is the purpose of that sorting, exactly ? Does it really matter that it's balanced ? We probably will have a perpetual unbalanced proposal proportion because of shifting in focus of editors, needs, and because one proposal can call other in the same field. If we seek balance we will constantly have to reorganize the property proposals. What we need is the most is readability and meaningful and useful organization.
            One of the feature I would actually want is not balance who seems purposeless by itself, except if there is an overflow of proposals that challenges mediawiki or this leads to a mess, but alignment of the type of datas to the sorting of the proposal. One could think of "domain alignment" : if a proposal is linked to a class of datas, for example "date of birth" links to human (but could be generalized), then the place where we put another property related to "human" should be the same. Because we don't really want to search all other the place the best place where to put properties. So we should have a crystal clear sorting process with readable rules. The problem is the same before or after the proposal : how do we find a property after the creation ? What are the entry points ? It would be really great if we could find a rule that applies in both cases. Why ? Because, just stay focused with the property proposals, it would avoid request for a duplicate. In short : we should think hard on how to link proposals to projects and to classes of items, in a "not so messy" way. I created a point that was intended to be a global entry point a while ago, in the sense that when you search for something you would find it by reading it and find the classes, the projects, the proposals and so on starting from the page Help:Modelling. With a moderate success for know but maybe it's time to rethink again all together because it seems the problem is still not solved :) author  TomT0m / talk page 10:53, 8 April 2016 (UTC)[reply]
            TomT0m - well I wasn't expecting it to be exactly and always balanced, but you are right that the balance does naturally shift over time, and sometimes quickly when a lot of new properties are needed in some area. I'm not quite sure if I understand what you are proposing though. Many properties have the {{Constraint:Type}} constraint implying that the property should only apply to items that are instances or subclasses of some principle class. Are you suggesting we group them according to that class they are being applied to? So there would be human (Q5) properties and then there's work (Q386724) and geographic location (Q2221906) and organization (Q43229) maybe. Is there a concrete short list of these that would cover most property proposals well? Or maybe the best approach is just to have some sort of monthly list of all proposals with pointers from relevant other lists (and wikiproject pages) as others seem to be suggesting here ??? ArthurPSmith (talk) 18:38, 8 April 2016 (UTC)[reply]
            @ArthurPSmith: I think that there is too many approaches already for (atually not really) sorting properties and that's finally a mess. The "global list of properties" fails for so many reasons that I don't think I have to detail. We have to rely on external tools that are starting to pile up for this to have a chance to work. The "by domain of applications" have a moderate success, mainly because this was a starting point but I think nobody is really totally satisfied with this. On the other hand regrouping by the classes of items the property applies is an approach that is omnipresent in the computing world. I'd like that for example we find a way to integrate initiative of mine like {{Property for this type}} and the corresponding property. Then, as there is a set of properties that is appliable to a class of item, this is actually a natural way to sort properties as you note. Pulling we also have a natural way to sort the classes themselves : the class tree/hierarchy. I'd like that we use more those "natural structure" in our process. This would help to consistently develop the hierarchies, the classes themselves ... For example one stuff I think is not really good in our property proposal process is that it's not so easy, before proposing a property or reviewing it, is to see if there is a duplicate property that could be used. Mainly one our way to discover properties could be to create a proposal and wait for someone more involved to point us to a solution to our problem ... this is a massive loss of time because instead of searching for a solution, finding it, and add the statement, we have to wait the answers of someone to answer, and that can take days instead of minutes, at best, with the cost of entering a bureaucratic process. This is a massive failure of our organization. I really think that instead of tweaking the bureaucratic process, we should really best think of our global organization, in particular how to integrate better the different parts of the wikidata day to day work : creating projects, proposing properties, creating class hierarchies ... All this is actually too loosely coupled. We don't have working and consensual guidelines that could help us achieve that. author  TomT0m / talk page 10:47, 9 April 2016 (UTC)[reply]
Do bear in mind that some editors are prepared to accuse other editors of "hiding" proposals, if they do not agree that the page on which the proposal was made is the correct one, even when the proposal gathered comments from other, uninvolved, editors, and remained visible for a period of months. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:17, 22 March 2016 (UTC)[reply]
Perhaps the answer to that is to have properties proposed on e.g. monthly pages and pointers to them placed on various lists, e.g. the SSSI (England) ID I proposed could be linked to from pages relating to Authority control, environment, Places (and/or United Kingdom if we want that specific) and designations. People could then watchlist whichever topics they are interested in and get a notification whenever a property is proposed. I'm thinking something similar to en.wikipedia's Deletion sorting project. The initial sorting would have to be done by a human (and I'm willing to put some work into that if people want) but the tracking after that (status updates and archiving) would be doable by bot easily enough (but I am not competent to write or run one myself). Thryduulf (talk: local | en.wp | en.wikt) 23:53, 22 March 2016 (UTC)[reply]
This is what I suggested, above (timestamp: 18:19, 21 March 2016). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:39, 23 March 2016 (UTC)[reply]
Ah, I somehow missed that comment, sorry. It should be obvious given how similar our suggestions are that I support yours. Thryduulf (talk: local | en.wp | en.wikt) 14:05, 25 March 2016 (UTC)[reply]

Revising Property Creator Guidelines

@Rschen7754, Innocent bystander, Jasper Deng, Jura1, Ajraddatz: @Thryduulf, Pigsonthewing, Jklamo, Izno, Succu: @Nenntmichruhigip, Egon Willighagen, Pamputt, Micru: @Vogone, Fralambert, Hsarrazin, Ash Crow: I have made a draft updated version of the guidelines for property creators (original is WD:PTC). So far the only changes are some edits to the "Steps" to better reflect or clarify current practice (leaving a space between creation and archiving, some property documentation details, no need for the old step 7) and some translation-related changes lost in a revert. The above discussion suggests some more changes are needed to clarify handling of property proposal discussions and perhaps some other procedures. When/if we can reach consensus on my copy I'll move it from draft to replace the current version. ArthurPSmith (talk) 16:01, 22 March 2016 (UTC)[reply]

This is already the subject of an ongoing discussion at Wikidata talk:Property creators#‎Undiscussed changes:. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:17, 22 March 2016 (UTC)[reply]

Frustrating reform

I'm afraid this reform won't really change anything to my concrete problems with the process, this is just minor changes. What about a proposal like
   Under discussion
Data typeMISSING
Example 1MISSING
Example 2MISSING
Example 3MISSING
 ? A lot of discussions, a lot of approval, a lot of time ... but still no real reason for the creators not to create it ? author  TomT0m / talk page 21:42, 22 March 2016 (UTC)[reply]
An additional word : it seems that this reform is directed to the process with the view of the creators. It's important, but ... what about the proposers themselves ? Don't we risk that we end up with an optimized process for the creators but that the consumers are totally forgotten and are actually discouraged to propose something ? author  TomT0m / talk page 10:11, 23 March 2016 (UTC)[reply]
I've made all my comments on this page from the PoV of a PC, obviously, but I've tried to keep proposers in mind, and suggest things that will make life easier for them. Do you see any particular barriers in what's being suggested, or have additional suggestions, in that regard? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:37, 23 March 2016 (UTC)[reply]
I raised concerns for a long time, the review process is either too long for some properties or too short, some are left in the pipe forever while some created very fast are quite immediately proposed to deletion or after some time which suggest that the proposer for deletion sometime completely missed the creation process for some reason ...
Actually reviewers are probably the keypoint in the process. Because it's not that easy to find by yourself a property if you're a newcomer, or to have forgotten that a property existed, it's easy to miss something, so a good knowledge of Wikidata is crutial to take a good decision, avoid mistakes and gives good advices to property proposers. So I think what we should improve first is the review process. Pages are pretty hard to follow, it's easy to mix something, you must have all the pages in your watchlist, proposal are badly sorted with duplicates ... And it does not improve with times.
Also the usage rule of some property are not always very clear, and a bad process (for users) have a lot of bad implications : users will avoid a proposal and will more easily broaden the properties or misuse an existing property if it's too hard or take too much time or if the result is random. From what I can see, the rumour that the process in Wikidata is too bureaucratic already spread out.  – The preceding unsigned comment was added by TomT0m (talk • contribs) at 12:15, 23 March 2016‎ (UTC).[reply]
Thank you. In understand - and share - your frustration. However, I was hoping for specific suggestions for improvements. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:21, 23 March 2016 (UTC)[reply]
I just reread a few stalled proposal, and it seems that when one or two people opposed, it's enough to stall the process into "inconclusive" and that situation can stay forever. I'd like that the debate beetween the persons that oppose and the people that are for really "has" a follow up. For example in the example I gave above Snipre opposes, I answer him but he does not answer to my answer, and we are stuck. Should we create a rule to avoid those kind of stucking, such as if the concerns of an opposed votes are actually answered, the vote should be cancelled as not motivated anymore if no answer from the voter with more solid arguments ? author  TomT0m / talk page 12:46, 23 March 2016 (UTC)[reply]
@TomT0m: Wouah, belle mentalité. If someone doesn't answer, this means he doesn't want to answer. You never try to ping me a second time or to let a message on my talk page. Just for your information I receive each day several pings from different pages, sometimes several pings from the same page and often I can't find exactly the position of the ping in a long page. Next time just ping a second time and if you can't do that I can deduce that the topic is not important for you. Snipre (talk) 20:02, 23 March 2016 (UTC)[reply]
@Snipre: This was not personal, you just noticed it's in a general discussion and I was asked for something concrete. Your vote was just the illustration of some pattern that can block a proposal, I just pinged you because you might have some idea, not to fill a complain request against you. author  TomT0m / talk page 20:48, 23 March 2016 (UTC)[reply]

Suggestion for complex proposals: review meetings

I understand the frustration of TomT0m and others with the slow reviewing process, however there are proposals that are too complex for a single property creator to decide on. I would like to propose a monthly IRC meetings for property creators to discuss and decide on proposals based on the comments/information received. I think it might be easier to generate decisions based on group input when consensus is not so clear.--Micru (talk) 12:42, 23 March 2016 (UTC)[reply]

I don't use IRC, I already have way too many communication channels (email, slack, trello, jira, gerrit, wikidata notifications ...). Can't we find a wiki way to do reviews? ArthurPSmith (talk) 14:41, 23 March 2016 (UTC)[reply]
Not everyone uses IRC, and we cannot require people to use it. --Rschen7754 14:44, 23 March 2016 (UTC)[reply]
I don't see a reason we couldn't have a property creator noticeboard. This could also serve as a "here, guys, my property is basically ready, please move it to ready/create it" place. --Izno (talk) 16:57, 23 March 2016 (UTC)[reply]
Good idea with the noticeboard. It could also be used to generate further discussion on tough cases. Ajraddatz (talk) 19:37, 23 March 2016 (UTC)[reply]
I've previously posted requests for PC action to the admin noticeboard and to project chat, with little effect. Indeed, aside from snide and irrelevant comments, a five day-old request on the former, in another matter, currently remains unanswered. There are requests on Wikidata:Properties for deletion which have been awaiting closure for months; and I recently posted on project chat about the lack of progress on the pages set up especially for discussing conversion of properties to the "external-id" datatype. Sadly, I doubt that a further noticeboard would solve the underlying problem. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:33, 23 March 2016 (UTC)[reply]
Right, but property creators don't frequent AN, and stuff gets lost on PC. Property creators can do nothing about PFD nor about ID conversion (though those may be the same people interested in those discussions). A single noticeboard would allow property creators to discuss difficult closures, request feedback from the community on "last minute-before-I-close this particular proposal feedback", provide a place for a weekly "I set these properties to created/ready/not created" and etc. --Izno (talk) 16:32, 24 March 2016 (UTC)[reply]
All admins are PCs. The example I gave of an unanswered request for admin action on AN is not a PC matter; and has been ignored by the admins active who do read it. My general point, illustrated with such examples, is that simply creating a noticeboard does not cause volunteers, in any aspect of Wikidata, to become more responsive. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:36, 24 March 2016 (UTC)[reply]

Just because all admins are PCs does not mean all admins have the knowledge to function as PCs; they are simply trusted to do so if they should decide to do so.

But your broader point is better-stated now. I would expect a property proposal noticeboard would be watched by the creators (and I certainly plan to), if no-one else. --Izno (talk) 18:53, 24 March 2016 (UTC)[reply]

I certainly concede your former point; it was made abundantly clear in recent discussion of PC matters at AN. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:09, 25 March 2016 (UTC)[reply]
Just a note that I would support a property creator noticeboard (or perhaps just some easy way to "ping" all property creators). Mostly I would expect just brief pointers to items like this RFC or particular property proposals or pages like the external id's one that need review, but lengthier discussion among property creators to try to obtain consensus on relevant topics would be helpful. Posting to "Project Chat" or the Admin noticeboard allows things to get lost among many other topics (only the few who happen to be paying attention at that time may catch the discussion). ArthurPSmith (talk) 14:56, 24 March 2016 (UTC)[reply]
We already have a lot of discussion place, not all of those are successful. But I think what's most important here is that a proposal keeps contextualised wrt. the rest of the wikidata ontology. Maybe we need a pre-proposal step where the needs are analysed and proposals of possible problem solving, possibly with existant properties, would be appropriate to spare a proposal. I think our point is "how do we represent knowledge", and new property is possibly not the best option. This step would be a good opportunity to maintain existing properties and not forgot older initiative and help organize existing properties, maintain a list of answers wrt. Wikidata current ontology ... author  TomT0m / talk page 15:13, 24 March 2016 (UTC)[reply]
To be more precise : maybe all begins with a place like WikiProject Ontology. Someone asks a question like "how do I add this kind of information". Two cases
  1. we already have an answer : give him our answer - maybe a page abot the kind/class of item at sake.
  2. We don't. We open a discussion, whose goal is to find an answer. We ping relevant wikiprojects.
    1. We find a solution using existing properties => we document this solution and add the question to our answer set
    2. We don't => we think of new properties, and maybe related issues. When done => we create the properties and document the specific problem
If the question is related to a specific Wikiproject, we give them a pointer to the answer.  – The preceding unsigned comment was added by TomT0m (talk • contribs).
That would certainly be overkill for most external-ID discussions. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:45, 24 March 2016 (UTC)[reply]
The answer should then be pretty short, or we may imagine a specific case :
Three cases
  1. This is an external ID => propose, ping the relevant project
  2. we already have an answer : give him our answer - maybe a page abot the kind/class of item at sake.
  3. We don't. We open a discussion, whose goal is to find an answer. We ping relevant wikiprojects.
    1. We find a solution using existing properties => we document this solution and add the question to our answer set
    2. We don't => we think of new properties, and maybe related issues. When done => we create the properties and document the specific problem. If the properties still have to go through the current proposal process, we choose together a sorting on where to put the proposal.
I also updated the steps to include initial triage of properties and better sorting ... author  TomT0m / talk page 17:05, 24 March 2016 (UTC)[reply]

New proposals

I'm concerned to see three new proposals added in the last few hours. These are sub-proposals of something I and others had already commented on. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:03, 25 March 2016 (UTC)[reply]

The above discussion is preserved as an archive. Please do not modify it. Subsequent comments should be made in a new section.
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Requests_for_comment/Reforming_the_property_creation_process/eo&oldid=2033209371"