Wikidata talk:Requests for comment/References and sources

Definition

Please give definition and example of references and sources. Personally I don't see the difference. Snipre (talk) 08:02, 17 April 2013 (UTC)

If I understand correctly source is a set of data concerning a document and reference is the use of one source (partially or completly use) in order to source one claim ? Is it correct ? Snipre (talk) 08:21, 17 April 2013 (UTC)

Yes, the source is a whole set of data, the reference is the pointer that allows you to select the relevant data to the information that you want to uphold.--Micru (talk) 14:07, 17 April 2013 (UTC)

Too specific

This discussion is too specific to books problem: references have to include articles (scientific journal, newspaper) and media (TV, radio, web). By integrating those parameters we have to exclude all documents which are not used as reference for claim if we want to avoid addition of unecessary data which will increase the size of wikidata without any profit: this means data from Wikisource and Commons not used in sourcing wikipedia items can't not be uploaded in wikidata as first approach. Snipre (talk) 08:21, 17 April 2013 (UTC)

This discussion it is not only about books. A source can also be a video (with a reference pointing to a specific part of the video, or quoting some sentences), or a journal, or a web.

About the "increase the size of wikidata without any profit", I think the benefits would be:

1. To support the claims more effectively
2. To keep source information being machine readable too
3. To avoid repetitive work

--Micru (talk) 14:07, 17 April 2013 (UTC)

If you extend the concept to all types of sources you have to limit the possibility to enter data in order to avoid that wikidata becomes a database for everything. That's something which is important for the development team because they don't want to see ressources used just for data compilation: the first idea of wikidata is to support wikipedia. For example there are 94 editions of the book CRC Handbook of Chemistry and Physics, the idea in the first step is to have as sources data only for editions used as references and not others. In other terms sources not used as references will be deleted. If this is clear, there is no problem. Snipre (talk) 16:08, 17 April 2013 (UTC)

Do you have a link showing the dev team view about the limits of Wikidata they should know ? I have not shown a lot about that, execept a quick discussion with one of the dev on IRC, who seemed to be on edge with the development right now :) TomT0m (talk) 16:29, 17 April 2013 (UTC)

Ask user:Aubrey for the link. I know they had a discussion with the development team on last Saturday. Or my talk page for the discussion I had with him about that. Snipre (talk) 16:44, 17 April 2013 (UTC)

And I ask the development team some weeks ago: Wikidata:Contact_the_development_team#Need_comment_from_development_team Snipre (talk) 17:32, 17 April 2013 (UTC)

There are several things to consider. Adding every edition of a book -- I don't think Wikidata is the rightest place for that, if that data is not used in the Wikipedias. Also, there are plenty of external DBs for that. I am a bit less concerned about the increase because of technical reasons, but I wonder if a community can really keep up with editing that. Why add items that you don't need? References are as flexible as they are in order to allow to express references, and at the same time use items for sources. --Denny (talk) 14:22, 18 April 2013 (UTC)

I also oppose importing unused data, it would be just... too much.--Micru (talk) 19:39, 18 April 2013 (UTC)

Clarifying the terminology confusion

Overview of a Wikidata statement

I just realized that in Wikipedia the words "notes" and "references" are used instead of what I called "references" and "sources" in the RFC. However, what wikipedia articles say is:

• references: Reference is a relation between objects in which one object designates, or acts as a means by which to connect to or link to, another object., but also In scholarship, a reference may be a citation of a text that has been used in the creation of a piece of work such as an essay
• source: Source text, in research (especially in the humanities), a source of information referred to by citation
• citation: Broadly, a citation is a reference to a published or unpublished source (not always the original source). More precisely, a citation is an abbreviated alphanumeric expression (e.g. [Newell84]) embedded in the body of an intellectual work that denotes an entry in the bibliographic references section[...]

Then Wikipedia:Citing sources states that A citation, or reference, is a line of text that uniquely identifies a source. And depending on the article, there are the sections: "notes", "references", "bibliography", "footnotes", "recommended reading", etc. In a way it is true that we shouldn't call a "source" to something that is not the source itself, but a "reference to the source" or "what helps to find the source of information", and maybe it is equally wrong to call a "reference" to "a selection/annotation of information inside the referenced source". For all this reasons I suggest that we spend some time finding the right wording and the right definitions for the terms that we want to use and that will prevent further confusion. My suggestions are:

• What is labeled now in the graph as "reference (colapsed)": claim source
• The wikidata item that might be linked from the claim source box and that represents a body of information upholding the claim: source item
• When instead of a source item an html link is provided: source link
• The additional information that selects the relevant information for the claim inside the source item or the source link: annotation

Please note that this are just concepts not implemented yet and what I called source item might be a any item (which I don't recommend).--Micru (talk) 00:38, 19 April 2013 (UTC)

Office hour about sources

Heya :)

Denny offered to do an office hour to help with answering remaining questions you might have about sources. I've set this up. Details are here. I hope you can make it. --Lydia Pintscher (WMDE) (talk) 21:26, 19 April 2013 (UTC)

Just a reminder: This is in a few hours. --Lydia Pintscher (WMDE) (talk) 12:47, 25 April 2013 (UTC)

Office hour log.--Micru (talk) 18:00, 25 April 2013 (UTC)

integration of non english speakers

Hello :) ,
in german Wikipedia are growing critics, that bots are filling in content without sources. It's my aim to give Wikipedians from German Wikipeida the ability to understand, what happend here in the ROC. Maybe some people can help to explain the content and get in discussion in german on our project page. We are on the way to centralise discussion for integrating Wikidata content in German Wikipedia, because there are several locations of discussion at this time. Greetings and thank you for your work, Conny (talk) 10:08, 25 April 2013 (UTC).

Some decisions

Not a lot of decision during the office hour. Ok, so as we have to go forward I propose the two possible options. There are only two solutions because all other solutions will be a nightmare for people creating lua models or templates with inclusion syntax (if you don't see why there are only 2 solutions just try to code the inclusion syntax for reference data able to manage all possibilities: authors defined in the reference section of the statement, authors defined in the item describing the book, date..., edition, place of publication, editor,....

• All data in the source section of the statement. No possible reuse
• Most of the data in one item which can be reused but implies several items for each edition.

I just give an example: I have one book which contains data about 5000 chemicals. And for each chemical I have at least 2 data and sometimes until 4-5 data. If I want to source the data I prefer to use a item for the reference including all data of the reference (ISBN, editor edition,...) or I will need to use a bot to add the same set of data in each source section of the statements (it is not a problem in my case because I am preparing a bot to perform my data import from an excel file where I stored my data). Snipre (talk) 18:41, 25 April 2013 (UTC)

To illustrate, I give both possibilities in the same item ethanol where the statement CAS registry number is sourced with option 1 and where the statement chemical formula is sourced by option 2. And for information both data are coming from the same book and from the same page. If you don't see a big difference click on the linked item defined under property stated in. And now think what is a good solution for sourcing 10 or 15000 statement about chemicals Snipre (talk) 18:56, 25 April 2013 (UTC)

For me the simplest solution is to allow reference item when a manifestation of a book is used more than x time as reference. Then we can keep the number of items describing a particular example of a book quite low. But the best solution to avoid confusion between a work and all of its manifestations is to create a new domain for reference item where references can be designed by RXXX and not QXXX. Snipre (talk) 19:06, 25 April 2013 (UTC)

If each reference gives the ISBN and page in the book then it can only be reused to support statements on that page of that edition. Statements referring to other pages or editions need a different reference unless "Reference" is being used here for the whole book in which case it might as well be an item (QXXXX). There doesn't seem to be a good reason for a new namespace. Filceolaire (talk) 14:24, 8 May 2013 (UTC)

If you have a look at the example I did you will see that page property is not included in the source item. again the idea of a source item is its reusability so if you integrate the page property in the item, its reusability disappears. 178.237.94.235 19:57, 8 May 2013 (UTC)

Adding cited webpages as sources

I would like to see persistently cited webpages (say, any citation that has remained in an article for over a month) get an entry. This should include metadata about the webpage (which could auto-fill the citation template), as well as links to online caches of that page. If we one day start gathering information about source reliability, that sort of information would also be connected to the same global entry for the source. Sj (talk) 16:28, 18 May 2013 (UTC)

See Help:Sources for the parameters needed to cite a webpage. But I don't think we can create an item for each webpage or website. First because this kind of data is quite unstable over a long time even for institutional websites where updates often modify the website structure (data are still available but the url is changed) or website/webpage can be deleted without any warning. So for reliable data we need to focus on hard documents or very well organized system like scientific journals. For databases, the best solution is to avoid the use of url and to provide the ID of the element in the database (like data used in authority templates). Again the database structure can change but the main parameters like ID stay constant. Reliable data means stable and secured data and this is often not the most accessible ones. Snipre (talk) 17:01, 18 May 2013 (UTC)

Option: A separate entry for each expression of a work, and for distinct editions

• Minor editions – multiple editions with the same ISBN (for instance, multiple printings, with tiny corrections) – should have a single entry. Information on the various printings and their date range could be included as a field.
• Each major edition – with separate ISBN, or from separate publishers – should have its own entry. (These are truly different works; they often have different introductions and indexes. See for instance On The Origin of Species.)
• Each expression – the collection of all reprintings and translations by different publishers – should have its own entry. (This is the level of abstraction that usually gets its own Wikipedia article)
• Each work – the collection of all expressions of that work in different formats and contexts – should have its own entry.

This will improve the precision with which people can refer to different abstraction-levels related to a single concept or work.

Comments

• Support as cleanly structured option. Sj (talk) 15:13, 18 May 2013 (UTC)

  I do not think this will produce "too many" entries - the vast majority of works are only published once. We do need a clean way to visualize this FRBR-style tree for a single work, but listing all editions in a single entry doesn't solve that problem. Sj (talk) 15:10, 18 May 2013 (UTC)

Options being removed from Project page.

I've noticed that some of the options in the Editions discussion have been removed or shifted (it's happened twice). The reasons given are that redundant conversation is being removed. I feel angered by that because I need to see the full text of everybody's opinions in one place. I also need it to be true that user opinons are only deleted when consensus is made to do so. I request that no more subtractive edits are made to the edition discussion please. Maximilianklein (talk) 17:11, 23 May 2013 (UTC)

We are speaking above options since one month so I think I know what is consensus. And don't mention consensus about removing when you don't try to find a consensus when proposing a new option. I am not against new propositions but I don't want to spend another month in general discussions especially when nothing new is proposed in term of practical solution.

I hear that you are becoming frustrated that this discussion is lasting a long time, is that correct? I'm not sure how to help your frustration. When you say "I don't want to spend another month in general discussions" I'm curious to know how much long you think this discussion should take, and what you will want to do if we take longer than that? Maximilianklein (talk) 00:25, 25 May 2013 (UTC)

Next time try to put first new solutions in the talk page and once you get consensus put it in the main page. Don't share general discussions between talk page and main page in order to keep the main page clear and showing the result of a real consensus defined 'after discussions.

Ok, I'm happy to put new solutions on the talk page before they go on the main page. I still feel a bit hurt that you didn't put my new solution on the talk page, but instead deleted the text. But I'm happy to put that behind now and move forwards, proposing new solutions on the talk page first if we agree on that as a rule? Maximilianklein (talk)

Before we start an edit war just explain what is new in the proposition of Sj compared to what is already proposed ? 1) We are speaking about edition so work expression is not concerned 2) work edition defined as item is already proposed in the second option 3) what is major or minor edition ? no information is given so this needs refinement. Snipre (talk) 18:12, 23 May 2013 (UTC)

I don't want to get into an edit war either. It seems like you want to keep the main page clean and readable by removing discussion you think doesn't belong there. And I want to have all the discussion in one place so we don't have to read both the discussion and the main page to get a full understanding of what's going on. Normally this isn't a problem but on some Project pages it is the style that I've seen to have discussion on the main page and not on the talk page. Maybe you can give me a good rule for what should go on the talk page and what should go on the main page, so I can understand the difference and avoid an edit war? Maximilianklein (talk) 00:25, 25 May 2013 (UTC)

I think on the subject of 1) we are talking about when to represent work and when to represent the edition, and when to represent both, I think the topics are related so it's OK to talk about works in an edition discussion. 2) The second option I see is each edition as a seperate item, not each work as a seperate item. I'm confused about what you mean by work edition? 3) I see that SJ in his proposition defined major editions to have new ISBNs and minor editions to have same ISBNs. Although I don't personally agree with recording multiple editions with seperate isbns. Maximilianklein (talk) 00:25, 25 May 2013 (UTC)

Properties set for book available

I created publication date (P577) so the minimal set of properties to describe a book and to source a statement with a book is available. Only the way to store information is missing. Without a clear decision or if new developments are necessary for storing edition data, I will propose to store source data in the source section under the statement like it will be proposed for all other types of sources (Help:Sources). This will ensure a homogeneous solution until a optimal solution is available for book edition.

Is Wikidata a wiki ??

Wikidata is a wiki. This means that anyone can edit. Anyone can add or change assertions. Many assertions can be supported by references or sources and it is relevant that there will be sources added to assertions eventually.

At this state of the game it will be a disaster when people who are not active on Wikidata force sources on every assertion. I have added many assertions on Wikidata, at the last count over 2800, none of them are sourced at this time. The process was that I used lists on a subject in Wikipedia, I read the articles and included statements from the texts and from info boxes. I am considered a trusted user in that I am careful about the assertions that I add and have been recognised as such.

When Wikidata is to grow a community, it is important that such activities are allowed even encouraged. At this stage, Wikidata is adding information at a phenomenal pace and, the information is coming from many sources. It is added by bot and by hand and many sources provide the information that is added. There are many, many more millions of assertions that will and should be added in this way.

How sources should be added is open for discussion. When a specific book states something as a fact, it means not that much when other publications disagree. Adding sources to assertions means not that much either when the sources can not be handled in the manner or in a similar manner of a nano publication. (this is open technology and I am sure these people can help us with adding their functionality to Wikidata).

In conclusion we are adding millions of assertions by hand and by bot. We will start adding sources to these assertions. But they will not make them facts, they will make them assertions with some claim. Thanks, GerardM (talk) 06:55, 3 June 2013 (UTC)

Statements entered manually won't be deleted, however it is important to raise awareness about the need of sources and hopefully some day we can start some action to complete the sources all statements, including your 2800 assertions.--Micru (talk) 17:45, 4 June 2013 (UTC)

Guidelines about labels

I do not think we have guidelines about labels, especially about translations. We need the original title somewhere, as it is needed for citations, but for that the "title" property is a better fit than the label, where there can always be synchronization issues across languages. So what should we have in the label ? The original title ? A translation like in Q13414893 ? Or even an alternative format like in Q13409386. --Zolo (talk) 07:04, 3 June 2013 (UTC)

I think the label should always be translated, the original title that holds the source should probably be in the original language. I think Q13409386 is a case of Harvard or APA citation. Probably better to put that into the aliases. At a later point in time the search function should be intelligent enough to recognize that "last name" + "year" means that it has to look for those article that have that kind of citation (generated out of name of author and publication date). Then inputting that into the aliases would be redundant anyway. --Tobias1984 (talk) 07:21, 3 June 2013 (UTC)

The (now extended) property P357 (P357) can have the original title when used in the work item and a translated title when used on an item representing the translation edition. The main item can also have labels in different languages.--Micru (talk) 20:45, 3 June 2013 (UTC)

Sourcing with a book

For good you can't source with the work item of a book: you always need to use the edition of a book. If only the original edition exists the question is now: do we you have to create a second item after the work item ? For me yes in order to keep the extracting code in client as simple as possible. Then it is necessary to describe clearly where each property has to be defined: this again in order to develop a simple code which can look for the correct value in the correct place of the wikidata structure.

1) Check if the book does exist as item in Wikidata

If it doesn't then create an item for the book and populate with these properties if there are relevant: author (P50), P357 (P357), P392 (P392), original language of film or TV show (P364). Other properties can be added if necessary: based on (P144), genre (P136), movement (P135), illustrator (P110), follows (P155) and followed by (P156) (in case of a serie).

2) If it does check if the book edition does exist as item in Wikidata

If it doesn't create an item for the book edition, put the edition number, the language and the publisher in the description of the item (to differentiate the edition item from the work item) and populate with these properties: publisher (P123), place of publication (P291), publication date (P577), edition number (P393), language of work or name (P407), ISBN-13 (P212), edition of (to create). If necessary add editor (P98), part of the series (P179), volume (P478). Then other properties can be added illustrator (P110) or translator (to create).

3) Source your statement with stated in (P248), pointing to the book edition item and add the property page(s) (P304) and its value behind the property stated in (P248).

Snipre (talk) 18:17, 3 June 2013 (UTC)

Why not to link both items, "Stated in" and "edition"? And we still need a property to link edition items...--Micru (talk) 20:30, 3 June 2013 (UTC)

This is done with the property edition of to link the edition item with the work item. Stated in is used, in which case do you want to use it ? Snipre (talk) 20:41, 3 June 2013 (UTC)

And what about the other relationship (Work item with its edition)?

For your question, what I meant is: instead of using only the edition item as source, use also the work item.--Micru (talk) 20:51, 3 June 2013 (UTC)

Template:Harvard citation (Q8860997)

Français : Quel est l'usage final des sources dans les éléments ? Je m'explique : une grande partie des propriétés que l'on ajoute aux éléments sont prévus pour être utilisées par les infobox des différentes Wikipédia. Les sources doivent-elles y être utilisées ? Dans ce cas, il y a un petit problème : plusieurs Wikipédia utilisent le Template:Harvard citation (Q8860997), qui est une abréviation de la référence avec juste le nom de l'auteur et l'année de publication. La propriété author (P50) ne permet pas d'avoir uniquement le nom de l'auteur, sans le prénom. Il faut aussi que la propriété publication date (P577) soit obligatoire. Ne faudrait-il pas ajouter une propriété "Référence Harvard" ?

English: What is the aim of the sources in the item ? I clear my mind : most of properties that we add to items are supposed to be used in infoboxes inside different Wikipédia. Should sources be used also ? In this case, there is a little trouble : several Wikipédia use Template:Harvard citation (Q8860997), that is en abbreviation of reference with just the author's surname and the year of publication. Property author (P50) does not allow to have just authour's surname, without his firstname. Property publication date (P577) must also be mandatory. Shouldn't we add to reference item a new property, "Harvard citation" ?

Français : Je pense que on doit attendre le resultat de la proposition nom de famille. Si on a cette information et la publication date (P577), alors Template:Harvard citation (Q8860997) peut être generé automaticament. Si no, on peut initier la discussion pour un proprieté "Référence Harvard"

English: I think we should wait for the outcome of the surname proposal. With this information plus publication date (P577), Template:Harvard citation (Q8860997) can be generated automatically. If not, it is possible to start the discussion to create a "Harvard citation" property.

--Micru (talk) 13:20, 4 June 2013 (UTC)

We actually can't generate Harvard Citations because we don't have a field for primary and secondary author. We need a qualifier for the author property to do that. --Tobias1984 (talk) 07:14, 6 June 2013 (UTC)

Authority control

I was thinking about Micru's comment about the paper with the 3171 authors (I think this one). How would a person or a bot even begin to create such a source-item. The journal only lists the initial of the first name and the last name. How in the world would it be possible to create items for all those authors. 6 of them have the same last name (Zhang) and two of those have the same initial (Z. Zhang). If I find any of the names on Wikidata, how can I be sure who is meant. I can't find any VIAF or other identifiers on that page. This is another example of things that have be written up somewhere in the sourcing guidlines. Otherwise we will attribute authorship to the wrong people and create hundreds-of-thousands of duplicate items. --Tobias1984 (talk) 21:13, 4 June 2013 (UTC)

See also discussion around taxon authority property proposal in biology. TomT0m (talk) 21:26, 4 June 2013 (UTC)

This discussion is outdated and should be archived. --Succu (talk) 21:31, 4 June 2013 (UTC)

How so ? Do you say that because it is old or because you came up with a solution ? Because the exposed problems justifying the need for this property seem very similar to me, so if it is closed it should be looked at again. TomT0m (talk) 10:54, 5 June 2013 (UTC)

One question about taxons, in Piciformes (Q25934), why is it using taxon author (P405) instead of instance of (P31) taxon (Q16521) plus author (P50)? --Micru (talk) 18:31, 5 June 2013 (UTC)

@Tobias1984: have you tried looking for them on ORCID (ORCID iD (P496))? A possible solution for those cases could be to have a property called "secondary author(s)" with datatype string. Main authors (usually sourced and identifiable) would get their own item and would use author (P50), the rest would be pasted as text in this secondary author property.--Micru (talk) 02:15, 5 June 2013 (UTC)

A secondary author field or (unverifiable author field) would be a good idea. As soon as a person has an item and some kind of authority control number people are going to think that he is the legitimate author. But if people assign more publications to this person just because he has the same name will wrongfully give credit to the person with the same name. I can already see author-items where one can look at a 300 year publishing record in 7 scientific fields :) --Tobias1984 (talk) 06:35, 5 June 2013 (UTC)

I have started the property proposal for unverifiable author with datatype string.--Micru (talk) 17:55, 5 June 2013 (UTC)

Reference, sources and Template:Cite book (Q92570)

Français : Bonjour, plusieurs Wikipédia utilisent le Template:Cite book (Q92570) pour insérer une référence dans un article. On peut imaginer qu'il y aura dans le futur une Phase avec un "Groupe de travail des références" similaire à la phase 2 avec le Groupe de travail des Infoboxes. Je pense que les propriétés des références doivent être utilisables comme paramètre des Template:Cite book (Q92570). Qu'en pensez-vous ?

English: Hello, several Wikipedia use Template:Cite book (Q92570) to add a reference in an article. We can imagine that there will be in the future a phase with a "Reference Task Force" like the Phase 2 with Infoboxes task force. I think that the properties for reference must be usable as parameters of Template:Cite book (Q92570). What do you think about this idea ?

Français : On a dejà, plus or moins, les propriétés nécessaires pour Template:Cite book (Q92570). S'il ya quelques qui manque, il est possible de les suggérer ici. Je pense que le meilleur moment pour démarrer un tel groupe de travail serait après la propriété "page Web" est disponible.

English: More or less all properties necessary for Template:Cite book (Q92570) are already available. If there are some missing, it is possible to suggest them here. I think the best moment to start such a task force would be after the web page property is available.

--Micru (talk) 23:16, 6 June 2013 (UTC)

The simplest is to creat a new template in lua which extracts all information from an item. And instead af creating one template for each type of document we can expect to have only one template Source which can handle all sources. Snipre (talk) 00:22, 7 June 2013 (UTC)

Guidelines

Can somebody add examples for guidelines?--Ahonc (talk) 13:49, 8 June 2013 (UTC)

A statement is unsourced, now what?

We just talked about this proposal on irc. It's mostly about how to source. I'm sure we'll figure something out. I'm more concerned about what happens if statements are not sourced. Gerard already voiced his concern about this. So I want to import 10.000 claims based on Wikipedia, but I don't have an external source. Are you going to prevent me from doing that, or even worse, remove the claims? That would kill this project. We need a lot of claims to be successful in phase2 (infoboxes). Having overly strict guidelines means that phase2 will never be usable. I would propose a more pragmatic approach like Wikipedia: Our aim is to get everything sourced, but we accept that we have (lot's of) claims for which we still need to add sources. Articles start as a stub and grow over time to featured article status, we should have a similar grow path for our items. Multichill (talk) 19:51, 9 June 2013 (UTC)

Before you import 10.000 claims from wikipedia I believe we should review here and see if there is a way we could add sources to these, even if it does mean your bot needs work. Filceolaire (talk) 21:22, 9 June 2013 (UTC)

We can't prevent you to import that amount of data, but be warned that this will be considered as a provocation by some contributors. The best solution is to clear that question (massive import without sources) through a RfC. Personally I find that data without data source are useless mainly because some pedias won't use them because of their strong policy about sourcing. For that subject we can't decide ourselves without taking care of the opinions of the different pedias: wikidata is working with all pedias and not alone or for one or two pedias. Snipre (talk) 11:19, 10 June 2013 (UTC)

@Multichill: My personal opinion is that if such data imports are made, then we should have a template that is displayed in the corresponding Wikipedia entry that says: Wikidata needs help finding sources for some of the claims. The people working on the entry know the literature better than a bot and can quickly improve the item. After all the claims are sourced (could be displayed with green color for the boxes) the template in the Wikipedia article disappears again. Checking sources and checking data templates are standard practice in all Wikipedias, so why not adopt this in a fancy automatic way. Look at en:Hydrofluoric acid. The green tick-marks in the chembox show that the data entry has been checked. But does that mean that all the data without a tick-mark should be deleted? In my opinion no. It is just easier to involve the whole Wikipedia community in this sourcing effort. Data entries could have a little symbols next to them, asking the reader to validate and source the claim on Wikidata. After this was done the symbol disappears from the infobox. --Tobias1984 (talk) 12:06, 10 June 2013 (UTC)

New property for 'date retrieved'

The discussion at Wikidata:Requests_for_comment/Time_DataType_Properties#Date_retrieved has identified a problem with using point in time (P585). New property 'date retrieved' proposed to use for info sourced from web pages and databases. Filceolaire (talk) 12:06, 8 June 2013 (UTC)

Work consisting of multiple volumes

There are no guidelines to cite them. --Succu (talk) 21:57, 6 June 2013 (UTC)

Like for page: create one itemfor the multiple volume edition and add the volume in the source section of the statement. We don't need to specify in the edition item the number of volumes and this will ensure a maximum reusability of the edition item. Snipre (talk) 23:08, 6 June 2013 (UTC)

What about different subtitles and different publishing years? --Succu (talk) 16:04, 9 June 2013 (UTC)

Put every feature of each volume of the same edition in the source section of the statement. Keep the edition item as reusable as possible. Put if you have an example this can be interesting to analyse it in details. Snipre (talk) 18:29, 9 June 2013 (UTC)

Example 1: The Cactaceae (Q1388502): Four volumes. Only the publishing year is different. Vol I: 1919, Vol II: 1920, Vol III: 1922 and Vol IV: 1923. Used in Arequipa (Q4068983). --Succu (talk) 20:07, 9 June 2013 (UTC)

The page describing the edition can have a statement for each volume (new property 'volume number (string)', with a qualifier giving the publication date.

If all the volumes have the same details then I suppose they could be covered in one statement - new property 'number of volumes (string)'.

Any source referring to that edition will need to include the 'volume number' as well as the page in the source additional data.

I think that will work. Filceolaire (talk) 17:47, 10 June 2013 (UTC)

Example 2: Sukkulenten-Lexikon (Q13426044): Four volumes. Different editors, subtitles and publishing years. I created four items: Sukkulenten-Lexikon. Einkeimblättrige Pflanzen (Monocotyledonen) (Q13426046), Sukkulenten-Lexikon. Zweikeimblättrige Pflanzen (Dicotyledonen) (Q13427059), Sukkulenten-Lexikon. Asclepiadaceae (Seidenpflanzengewächse) (Q13427111) and Sukkulenten-Lexikon. Crassulaceae (Dickblattgewächse) (Q13427103). I hope this is correct. --Succu (talk) 18:20, 11 June 2013 (UTC)

A Work ID

A chart showing how a work identifier could fit with Wikipedia and Wikidata

Hi, I've been in discussions with a part of OCLC who are thinking about making an identifier for "Works" that is higher level than ISBN or OCLC number. It won't be in production for at least another year, but I wanted to share this concept drawing we made about how it would fit into Wikidata. Maximilianklein (talk) 17:14, 12 June 2013 (UTC)

Examples:

An item describing a work in English Wikipedia: The Great War and Modern Memory (Q4117225), en:The_Great_War_and_Modern_Memory,

Some detail about what catalog librarians have recorded about the various "editions" of that work: in worldcat

Thanks for sharing, Max. Is the tag "concepts" supposed to represent the "aboutness" of a work? We might need a property for that too.--Micru (talk) 17:57, 12 June 2013 (UTC)

As I understand it, yes. Maximilianklein (talk) 19:58, 12 June 2013 (UTC)

@Micru: I think depicts (P180) is fine for that. --Zolo (talk) 20:29, 12 June 2013 (UTC)

@Zolo, depicts (P180) is for visual objects, but I'm thinking that category's main topic (P301) could be "stretched" to include creative works.--Micru (talk) 03:04, 13 June 2013 (UTC)

Source from WEB page

I could not find any example of such source in the wikidata yet, so I am assuming it is still not available. There is also mentioned in the comments that the URL are not stable. They might disappear some day. So, when it happens and it would be possible to put URL as source, would it be possible to define alternative URL to web.archive.org to keep the material still sourced, or use such links directly when the original web sites do not exists? I want to put some sources to properties about cities in Poland, which are available in the official web pages of the seat of local government. Such things are trivial in wikipedia. Paweł Ziemian (talk) 21:18, 15 June 2013 (UTC)

URl property is still not available so webpage can not be used for the moment. Then about the use of webpage as source this will be accepted but there is always the risk that data will disappear one day. If you have a solution to avoid this situation feel free to propose a solution. From what I know about web.archive.org you need to be sure that the website you will use as source is available on web.archive.org in order to create a link there. Is it the case for your case ? Snipre (talk) 22:17, 15 June 2013 (UTC)

My experiences with the web.archive.org are there are not the pages I needed, however, once browsed in live view through the archive, it makes the page archived and available in a few months, when the index to the archived pages is regenerated. I tried it once and it worked. Currently I am waiting for a few next pages I "archived" recently. This method is useful even for wikipedia articles. The method of web citation in wikipedia with the original link + optional link to archived version of the page seems very practical and should be implemented some how in wikidata to make them more robust before they vanish into thin air. The URL property shall have some complex form with a few subproperties i.e. "main-address", "access-date", "online", "archived-address", "archived-date". The presentation form shall generate link to "main-address" if the page is available, or "archived-address" otherwise. Alternatively there shall be a set of such properties when web source is provided, and "archived" variants are optional in the set. But the problem "what to do when the link does not work, and there is no archived version available" still remains. Paweł Ziemian (talk) 20:10, 16 June 2013 (UTC)

Is there some example for Source from WEB page ? Actually, I have done (in Battle of Bagrevand (Q2137777)) :

• stated in (P248) = Charles Cawley, "Medieval Lands", 2006-2020 (Q13419312) (the general website)
• page(s) (P304) = http://www.fmg.ac/Projects/MedLands/ARMENIA.htm (the url of the page where information comes from)

Is it correct ? Odejea (talk) 15:37, 1 August 2013 (UTC)

No as explained the URL datatype is not available and page(s) (P304) is not designed to accept that kind of data (see Wikidata:Database_reports/Constraint_violations/P304#Format. Look at Help:Sources for the correct labelleing of sources. Snipre (talk) 13:34, 2 August 2013 (UTC)

For questions about how sourcing 1) see Help:Sources 2) ask questions on the talk page there