Completing Metadata

Introduction

This page will discuss some of the general processes about how an editor goes about completing records within the edit system and other suggestions for best practices related to managing metadata records that have been uploaded into the Digital Collections.

For specific information about how individual fields are used or how values are formatted to complete records, please refer to the other guides listed in the left-hand menu.

Considerations Related to Metadata Creation

Overview

The primary goal when we create metadata in the Digital Collections is to help users find items that they will find interesting or useful.

What Information Should Be Included?

Information should match the item and represent it as accurately as possible (e.g., not describing a clipping as a whole newspaper, or using a title that only applies to part of the content)
Any available information about how the item was made and what is in the content to help a user find a specific item or know if it has the information they need

Why Is Information Split Into So Many Fields?

Some information primarily describes what the item is and how it was made – e.g., creator, contributor, publisher, creation date, resource type
Some fields primarily describe what kind of content the item has and whether it will be useful to the user – e.g., content description, subjects, coverage
We mainly try to keep this information separate and not duplicate it across these fields
Occasionally information in could overlap and may be duplicated to reflect that in specific cases, e.g.:
- A self-portrait is made by the person who is in the image (creator and named person would be the same)
- A report could say “this is what happened today” (creation date and coverage date would be the same)
- Most photographs will have a creation date and coverage date that are the same (since the time of creation captures what is happening at that time)

Why Are There So Many Rules About Formatting?

We want information to be added consistently because there may be overlap between items in different collections or we may want to use the values for interface features
- Values entered in some fields are used for filtering and browsing (e.g., explore by…location, subject, date, etc.), but this relies on standardized formatting
- Many types of values are “clickable” in records if a user wants to find “more items” that have the same subject, the same creator, from the same series etc. – but this only works if terms are entered the same way – and these often apply across items described by many different people

Where Should I Even Start?

This page gives a general overview of what you should think about when doing metadata
The Quick-Start Guide (also linked in the left menu) gives a summary of the most important things to know about every field

How Do I Know When the Record is Done?

Generally, when you have entered all of the information you have available (and all placeholders are replaced or removed), you have finished the record
The page on Reviewing Metadata Records provides information on how to check a record to see if anything is missing/incorrect (the summary view may be useful for this)

Questions or Issues

If you are working on metadata and discover that you have questions or come across a different kind of item than you have described before, you have a few options to find more information:

For formatting or usage questions
For example: A name formatted differently from most common names Whether certain information should be included in a particular field An unusual date Information for an infrequently-used field (like source) A field that you aren’t sure applies (like citation/publication title) A qualifier that you haven’t used before (like a creator role, such as “attributed name” that you want to verify before using)	Review the guidelines to see instructions and examples, by: Using the navigation at left Looking at the top of the field in the edit system and clicking on a link: “Help” to see the “Quick-Start” summary of the field highlights and most frequently-used examples “Guide” to see the full guidelines including detailed suggestions and examples, additional resources, etc.
If you find an unusual item
For example, an item: different than others in a particular collection that does not have standard template values without standardized descriptive guidelines (like correspondence) that doesn’t fit in a specific resource type category	Check to see if there are project-specific guidelines or examples in the documentation for the fields Try searching in the Digital Collections for similar items to see how they have been described before
To get more context about an item
	For published items, try looking in the library catalog or WorldCat/OCLC to see if it has a record Consult any accompanying information (inventory sheets, finding aids, collection descriptions) Occasionally, it may be helpful to search for more information online (just be mindful of how long you spend on this, especially for an individual item)

General Considerations

A couple of notes about entering metadata and records in general:

If you open a record and do not end up making changes, DO NOT publish the record (this will save a duplicate copy); if you are new to the system, the System Quick Start Guide can give you an overview of navigating and saving your work
There is no way to add mark-up or styling to any of the text in metadata fields (e.g., italics or bold)
Any existing encoding characters or mark-up should be removed if you are pasting text from elsewhere (e.g., MARC sub-fields from catalog records, some PDF punctuation is translated as encoding strings, etc.)
Except for non-display notes, all text in all fields will be visible when the record is made public – any comments or placeholder values in the template or added by editors will display if not removed (more about templates below)

Editing Records

Completing Records from Templates

Most records are uploaded with at least some pre-filled values or placeholders and may also have notes or reminders. The goal of these values is to save time so that editors don’t have to enter the same value frequently (e.g., if all of the items in a collection have the same creator) and placeholders are intended to:

[1] highlight information that is required or expected for the record (e.g., titles or descriptions)

[2] remind editors to check whether information applies to a specific item, especially for less-frequently-used fields (like source, which may apply for certain items like clippings)

[3] to suggest formatting or values for greater consistency (e.g., subject values from controlled vocabularies that may apply)

Some things to keep in mind:

All placeholders must be replaced by appropriate values or removed if they do not apply
If you do not know what information should be entered, review the guidelines for that field	Examples of placeholders: [title] or {{{add title}}} {{{name}}} {{{add named contributors}}} YYYY-MM or YYYY-SS Volume #, Number #
In some cases, the type of placeholder may signal what kind of information ought to be entered	For example: {{{month YYYY}}} > > December 2008 {x, #} p. > > vii, 300 p. (i.e., Roman numerals and pagination are likely) Fiscal Year YYYY > > Fiscal Year 1965 (for a series title, the format must match other titles in the series, in this case it requires a single year)
Some values may apply frequently, but not to every item in a collection; for example, we might include a creator name that applies to 90% of the records so that editors do not have to add it every time, but there are still a handful of records where the creator name should be changed or removed. This also applies to suggested subject values that should be removed or changed if they do not apply to the specific item.
For certain collections, there may be additional notes included as placeholders to remind editors about special issues These will display publicly if they are not removed	Examples of notes: {{{look for similar records in the collection that are already completed}}} {{{add relevant subjects}}}
Some placeholders have notes to proved suggestions about how to enter information or where to find values The notes should be deleted and values need to be verified, entered, or removed	Examples: “Issue #.” – whole number issue Bowman, Inci A. {{{or name}}} “Issued {{{Quarterly}}}.” — adjust this based on cover/title page
When removing an existing value or placeholder that does not apply, use the “x” button to completely remove the whole entry If a placeholder includes qualifier values (e.g., a creator placeholder that has a type/role pre-chosen), removing only the text will leave an empty entry with a label that still displays publicly and these can only be found by looking at every individual record.

Helper Tools and Warning Messages

The record summary view includes an overview with fields flagged that may need review. Here are some examples of things that the system might identify as possible problems.

Values in the summary view are highlighted in red when the system finds a problem with validation
	This is an example of a validation error because there is a contributor role with a missing role or type – all creator/contributor entries must include both a role and a type
	Here, a date value in a date field (either date or coverage date) does not meet EDTF standards, so the date is invalid; this example is a placeholder (intended to be replaced with a “real” value) but any date entry that is not formatted correctly will be flagged
Values in the summary view are highlighted in yellow if the system suspects that it may be incomplete or incorrect, but needs review
	In this case, the record has a number of placeholder values that the system has tried to flag based on values like “{{{” in the title and “#” in the description. Placeholders should be fixed or removed.
	Some items are uploaded with very few values or placeholders. The system also flags missing required fields, like content descriptions and subjects (explained in the warning text).
Summaries may have a mix of red and yellow highlighting depending on the conditions
	The contributor placeholder is flagged because of the curly brackets but the type is red because there is a value in the name portion but not the type (invalid) – the whole entry would need to be removed or updated with a correct name, role, and type
Some values get flagged but are not incorrect – when in doubt, consult the relevant field guide for usage/formatting
	The title field is flag due to characters that may be placeholders but in this case, the “uniform” title is used to represent the general pattern for a serial title and may have gaps for information that is different in individual issues/years
	These subject values were flagged because of the word order, but happen to be valid Library of Congress Subject Headings in this case

While editing a record, some fields have a level of validation; if a field value seems wrong, the field will be highlighted in the sidecar along with a message explaining what needs to be checked:

This warning appears if a title, creator, contributor, or publisher value is more than 50% capitalized.

In rare instances, these values may be correct

This warning appears if a value in the date or coverage date field does not conform to EDTF specifications (check guidelines for more info)

This warning appears if a subject value does not match the vocabulary qualifier – this only works for vocabularies that have search modals

Issues with Items or Records

Here are some general guidelines to make things easier for administrators and other editors who may be looking at items in the edit system:

If you believe (or have determined) that an item should remain hidden for any reason, add a note to the main title with a brief explanation. This way, if another editor is viewing hidden records to see why a collection is not fully visible, it will be apparent from the Dashboard why something is not public. For example:
In the patent collection we know some items have scanning issues or are outside the scope of the collection (i.e., not patented by/assigned to Texas inventors) so editors append text to the end of the existing title (usually an identifier) marking what is wrong, e.g.:	SKIP (reissue) – usp017/re011446 Adding-Machine — NOT TEXAS SKIP last page blacked out SKIP: 01086715 – Not a Texas Patent. Baling-Press – MISSING FIRST PAGE 1407533 - Incomplete
Sometimes there are items in an archival collection that have been scanned for preservation but are found to contain personal information or to have other privacy/usage concerns and need to remain hidden until staff members can resolve any issues, e.g.:	[title] * * Home addresses - keep hidden * * [title] – contains SSN, leave hidden [title] – KEEP HIDDEN for privacy concerns
If an item is a duplicate of another item (i.e., the content is identical and we would likely not want to keep both), it is helpful to include the ARK of the other version so that anyone reviewing these records can verify that they are duplicates or decide whether both versions should be kept, e.g.:	[title] – duplicate of ark:/67531/metapth34017
If an item should be deleted (usually because it is a duplicate), it should be clearly marked in the title
Note that only collection managers or administrators should be making this determination; records marked to be deleted will be suppressed and no longer viewable in the edit system (i.e., this cannot be undone)	[title] – DELETE, duplicate of ark:/67531/meta# DELETE, duplicate
Generally, records that have been made public should not be re-hidden, but there are exceptions (e.g., identification of duplicates, take-down requests, etc.).
In cases where a choice has been made to re-hide a record, a note should be added to provide any additional context and information not included in the title.	Non-Display Notes: This version contains consent, keep hidden but do not remove. To remain hidden due to personal contact information – and possible copyright issues

Here are some general categories of problems and how they are handled:

There are two copies of the same item and we only want to have one. In this situation, the duplicated item (usually the one of lesser quality or added later) is marked as a duplicate and left hidden. If verified by an administrator, the duplicate will be permanently deleted.
The record and item/s are mismatched, e.g., pages from two items were grouped together or two pages of a single text were separated by mistake. In these cases, the titles are marked and an administrator will re-organize the files to replace the items, for example:
- If a two-page document was accidentally uploaded as two items, all of the images for both pages will be uploaded in place of one of the existing items (that metadata record will not change); the other item will be deleted.
- If two newsletter issues were accidentally uploaded as a single item, the images for all of the pages of the first issue will be uploaded in place of the existing item and the images for all of the pages of the second issue will be uploaded as a new item with a new metadata record.
There is an issue with image quality, for example, a photo is rotated and displays sideways or upside-down. In this case, an administrator will correct the original file and it will be uploaded in place of the existing item (the metadata record will not change).
There are permissions issues with a specific item that contains personal information, copyrighted material, etc. Usually, this can be marked in the title and the record remains hidden until someone in charge of the collection can get more information.