Establishing serial Wikidata records for Penn's Deep Backfile
Welcome, and thank you for working on this project! We are
making sure that serial issues in
Penn Libraries Deep Backfiles have
Wikidata
entries that clearly identify them and distinguish them from
other serials. This is a necessary step in our publishing copyright
information about Penn's historic serial holdings. It also will help
us establish linked
open data for serials Penn owns that can be combined
with information that others have freely provided and linked
related to those serials.
Many of the serials have Wikidata entries already, but
some do not. Some serials may also have Wikidata entries that are
not linked to our Deep Backfile table because they do not have an
ISSN given in our Franklin catalog recorded in their Wikidata record.
Once a serial has a Wikidata entry, we can link the copyright information
that we have created at Penn with that entry, and that copyright information
will be linked from the Penn Libraries' Deep Backfile table (as well as any
other Deep Backfile
table that includes that serial.)
Before you start, you should have a Wikidata account.
(If you already have an account
on Wikipedia or another Wikimedia service, that should also work for Wikidata.
If you need an account, you can
create one here. Do not create a new account if you already have one.)
Then you should should get familiar with editing Wikidata.
If you have not had much experience editing Wikidata, or want a refresher,
take the "Wikidata Basics"
Tours
provided by Wikidata for new editors, or some similarly useful training.
Once you're familiar with editing Wikidata, and are ready to start
working with serial records in Wikidata,
here's what to do:
- Sign into your Wikidata account. (See above for how to create one if needed.)
- Go to the Penn Libraries deep backfiles page.
- Find a suitable serial with no entry in the "Wikidata" column, and either "Pending" or "Unknown" in the adjacent "First renewal" column. "Pending" serials are higher priority than "Unknown" ones, since those are serials where someone has already given us copyright data that we can link once we have a Wikidata record.
- Go to Wikidata. Search there for that name of that serial, and see if the results include a record for the serial. If they do not, search for the ISSN of that serial, and see the results include a record for the serial. Note that if the serial title is not in English, there might not be an English label for the serial's record, so you may get search results that only show the record ID and not the serial title you expect.
- If you find a record for the serial, and you verify that the Wikidata record refers to the same serial as in Franklin, but isn't listed with the ISSN we have in Franklin, add a statement to the Wikidata record with our ISSN. (But see the Complications
section below for what to do if our ISSN is incorrect.) Also, if there wasn't an English label for the serial, go ahead and edit the label to include one. (See below for how to fill in the "Label" field.)
- If you do not find a Wikidata record for the serial, create one. Select the "Create a new Item" link on the left side of the Wikidata window. This will take you to a form that you should fill in as follows:
- Under "Language", there should already be "en" entered. Leave this alone. It simply indicates that the data you are entering will be on the label that will be seen by people editing Wikidata in English.
- Under "Label", enter the title of the serial (in English if there is an English title, otherwise the title in Roman lettering). Include initial articles as they appear in the title, if they do.
- Under "Description", enter a phrase to distinguish the serial from other things that might have that title or label, if necessary. (For example, "British literary quarterly"). If the title is unique and not likely to be confused with anything else, you can leave this blank.
- Under "Aliases, pipe-separated", enter any other titles or phrases the serial is commonly known by. If there is more than one, use a vertical bar ("|", also known as the "pipe" character) to separate them. In most cases, you don't need to put anything here, since there will be opportunities to enter title statements in the record itself. But if there is a very common name for the serial that's different from the label, it can be useful to include it here.
- When you have filled in everything you need to in these fields, press the "Create" button at the bottom of the form. This will take you to the new record for your serial, showing you the information you have entered for the label and the Wikidata ID that has been assigned to it. (Record IDs in Wikidata consist of the letter "Q" followed by a series of digits.)
- On the record, click on the "Add statement" link to add statements about the serial. You will then be prompted for a property name, and once you enter one, will be prompted to assign a value to it. Add applicable properties from the list below, if there are not already present and you know their values. You do not need to add references for your statements (though you can add them if you want to and there's something you can cite).
- instance of (required for our purposes): This will prompt for a controlled value that indicates the class of the item being described. Appropriate values include "serial", "periodical", "scientific journal", "academic journal", "monographic series", "magazine", and "newspaper". There may be others as well. Pick the most specific value you can reliably state for the serial. If in doubt, though, pick a general value. ("periodical" is more general than most of the other classes mentioned here, and "serial" is more general still.)
- ISSN (required for our purposes): This will prompt for a free-text string. Enter the ISSN we have on record for the serial, including the hyphen in the middle (e.g. "0278-0968"). Do not include quotation marks, spaces, or other characters. Once you enter a value, the string will be made into a link that you can follow to verify your ISSN at the ISSN Portal.
- You can optionally add a qualifier to the ISSN. (If there's more than one ISSN in the record, you'll usually get an icon asking you to add one.) The qualifier that usually gets used is distribution format. "printed matter" is the standard value for this qualifier for print ISSNs, and "online publication" is what usually gets used for electronic ISSNs on Internet publications. (However, I've noticed that qualifying the print ISSN as "printed matter" is enough to get Wikidata to stop complaining about multiple ISSNs even if the electronic ISSN is not specially qualified).
- language of work or name (not strictly required, but once an ISSN is present, there may be an icon asking that this be provided): This will prompt for a controlled value that indicates the language used in the serial (e.g. "English", "French","Japanese"). Multiple values for this property are fine if appropriate, but one is sufficient.
- title (not strictly required, but once an ISSN is present, there may be an icon asking that this be provided): This will prompt for a free-text string, and once you enter the string, you'll also be required to select a language from a controlled set of values (e.g. "English"). The language should be the language of the title, which is often, but not always, the same as the main language of the serial. (For an example of an exception, there are some titles like "Acta Botanica Croatica" that are in Latin, though the contents of their journals are not in Latin.) If the language uses a non-Roman script, it's best to use that script in the title string, rather than a Roman transliteration. While multiple title values are allowed, Wikidata may currently throw up an "issues" icon if more than one title value is present. (However, it may still make sense to have multiple values when a serial has different official titles in different languages.) Include initial articles as they appear in the title, if they do. There does not seem to be a strong consensus on whether to use title case or sentence case in Wikidata titles, so you can use either.
- country of origin (not strictly required, but once an ISSN is present, there may be an icon asking that this be provided): This will prompt for a controlled value that indicates the country where the serial is produced (e.g. "United States of America", "France", "United Kingdom"). Multiple values for this property are fine if appropriate, but one is sufficient. The controlled values offered will include more than just countries, so make sure you choose a value described as an actual country, and not some other type of value. For instance, you can use "United States of America" (described as "sovereign state in North America") but do not use "United States" (described as "Wikimedia disambiguation page"). If you are not sure what country to specify, click on the "ISSN" link for the serial to see what country is mentioned in the linked ISSN portal entry, and specify that country. (Note: There is a similarly-named property called just "country", which should not be used with serials. Make sure you use the "country of origin" property specifically.)
- inception (optional) can be used for the start date, if known. It prompts for a year or other date. You may see other properties occasionally used for serial start dates, but this is the property recommended on the Wikidata:WikiProject Periodicals documentation page. If this serial has predecessors in Wikidata, the inception value should be the start date for the serial under this title and/or ISSN, not the start date of one of its predecessors.
- dissolved, abolished or demolished date (optional) can likewise be used for the end date if there is one and you know it. It prompts for a year or other date. You may see other properties occasionally used for serial end dates, but this is the property recommended on the Wikidata:WikiProject Periodicals documentation page. If this serial has successors in Wikidata, the value for this property should be the end date for the serial under this title and/or ISSN, and not the end date of any of its successors.
- replaced by (optional) can be used to refer to the immediate successor to this serial, if there's a record for it in Wikidata. You will be prompted for that serial's record; you can either enter its Wikidata ID if known, or start typing in the label of the record in question. Either way, if the record exists with that ID or label, you'll be given the option to select it as this property's value. That record should then be given a replaces statement referring back to this record, if it doesn't already have one.
- replaces (optional) can be used to refer to the immediate predecessor to this serial, if there's a record for it in Wikidata. You will be prompted for that serial's record; you can either enter its Wikidata ID if known, or start typing in the label of the record in question. Either way, if the record exists with that ID or label, you'll be given the option to select it as this property's value. That record should then be given a replaced by statement referring back to this record, if it doesn't already have one.
- JSTOR journal ID (optional) can be used to refer to the serial's JSTOR page, if the journal is in JSTOR. This will prompt for a free-text string. If you know the journal is in JSTOR, this can be useful to add to be specific about the journal being described, and to link to full text content. The value should be the last part of the JSTOR URL for the journal, and if you put in a correct value, you will be able to click on it, once published, to go to the JSTOR page for that journal. For example, the value "aegy" will be published linked to the journal Aegyptus on JSTOR.
- Feel free to add other statements as recommended in the Wikidata:WikiProject Periodicals documentation page, but don't take too much time doing this.
- Choose another serial in the
Penn Libraries deep backfiles
without a Wikidata record, and repeat.
Once a Wikidata record is present that states an ISSN that's also in our
Deep Backfile table, a link to the Wikidata record should appear
in that table the next time it's generated. (We usually generate the
Penn Serials Deep Backfiles table
multiple times in a business day; you can also ask John to regenerate
it manually if desired. There may be a slight lag
between the time a Wikidata statement is recorded and the time that
it's available in table regeneration, but any such lag
will generally not be more than an hour, and is usually much shorter
than that.)
At some later date, John or another administrator of our serials project
will add an Online Books Page publication ID statement
to the Wikidata record, and that will in turn automatically add links to our
copyright information record (as well as to our record for free issues,
if any) in our Deep Backfile tables after they're next regenerated.
Complications
You may occasionally encounter some complicated situations. Here's how to deal with some of them:
- The ISSN Portal indicates that the Franklin ISSN is cancelled, or incorrect (e.g. is for a different journal than the one we catalog under that ISSN). In those cases, try to find a correct ISSN using the ISSN Portal and/or WorldCat. Then email Rachelle and John
(using the email addresses at the bottom of this page)
so the ISSN can be fixed in Franklin, and accommodated if necessary in our Online Books knowledge bases. (Include the correct ISSN in your email if you discovered it, but email us anyway if you weren't able to find a good ISSN but know that the one we have is bad.)
- The ISSN Portal indicates that the Franklin ISSN is a "legacy" or unconfirmed ISSN. In those cases, see if there's a more authoritative ISSN for the serial (using the ISSN portal or WorldCat), and let Rachelle and John know if you find one. Otherwise, search for the ISSN in WorldCat, and if that search turns up the same serial we catalog in Franklin, go ahead and assign it to its Wikidata record.
- Wikidata has "lumped" or "split" the serial differently than we do in Franklin. It's common for different libraries to make different decisions about whether serials should be entered under different records or a common record when certain events (such as changes in publisher or title) occur. Wikidata may tend to "lump" serials under a common record more often than Franklin or the ISSN agencies will. (John will sometimes do so as well,
noting the changed title as an alternative title in the common record,
especially for trivial title changes or short-lived changes that revert later.)
If there is a separate Franklin record and ISSN for an earlier title of a serial that's only in Wikidata under its current title, and the title change is significant and persistent, you may need to create a new Wikidata record for the earlier title. Put the earlier title's ISSN in that record (and then remove that ISSN from the current-title Wikidata record if necessary). The two records can then be linked with replaced by and replaces statements, as described above. It's a good idea to also add or edit inception and dissolved, abolished or demolished date statements with dates that make it clear when the older serial title ends and the newer one begins.
- Wikidata has two (or more) records for the exact same serial (and not for different titles or time periods of the serial). Occasionally duplicate records get created in Wikidata for identical items. This can often happen, for instance, when a large body of data is automatically ingested into Wikidata. If you encounter multiple records for the same serial that you're trying to work on, see which record is more established. (It may have a lower number in its Wikidata ID, for instance, or have more statements associated with it, including links to Wikipedia articles.) Use that more established record to add or edit relevant statements. Then we'll merge it with the other(s). To make that happen, you can email John with the Wikidata IDs for the records that need to be merged. Or if you're an experienced Wikidata editor and have made absolutely certain that the two records refer to the exact same thing, you can try merging them yourself.
Problems, questions, contacts
If you have further questions, or encounter problems as you
work on this project, you can contact
John Mark Ockerbloom at ockerblo (at) upenn.edu, and/or Rachelle Nelson
at nelsonrr (at) pobox.upenn.edu.
Thank you for helping with this project, and for providing
linked data about Penn's serials to the world!