Before creating a new collection, it is best to spend some time planning. Here are some questions to consider. Think about them and plan before creating your collection:
What sort of content do you have? What sort of metadata? Who will use the collection and how? How will the collection grow? How will images and media be contributed? Are they being scanned? Imported from a digital camera? Created digitally in Photoshop? Will they come in batches or one or two items at a time? Is there multimedia? How large will the images or multimedia be? How will the metadata be created? Will it come from text files, comma-delimited, tab-delimited, XML, or perhaps Excel-formatted files? Will metadata be entered by hand via Inscribe or will it come from a read-only database? Determine if the data is technical in nature, process-oriented, minimal, or complex. How is it created? What are the processes that it undergoes (is there a validation or data integrity process)? Do you use controlled vocabularies or external hierarchies? What types of fields do you use (numeric, text, date)?
The following sections will help you to think about these issues and plan your collection.

Designing a Metadata Structure

There are many metadata standards for cataloging visual and non-visual media, including the VRA Core, MARC, Dublin Core, CDWA, Object ID, CIMI, and EAD. Each of these standards has benefits and drawbacks; some are more suited for describing books, others for slides, others for complex objects or multimedia, etc.
Think about how the data is organized and what you want to do with it. Your specific needs will inform your choice of cataloging standards.
Below is a list of resources where you can learn more about data standards:

• • Introduction to Metadata - Metadata Standards Crosswalk http://www.getty.edu/research/conducting_research/standards/intrometadata/metadata_element_sets.html
  • Categories for the Description of Works of Art (CDWA) http://www.getty.edu/research/conducting_research/standards/cdwa/
  • The CIMI Profile http://www.loc.gov/z3950/agency/profiles/cimi2.html http://www.unt.edu/wmoen/projects/Z39.50/cimi/Profile/appendixc.htm
  • Dublin Core Metadata Element Set Version 1.1 http://dublincore.org/documents/1999/07/02/dces/
  • MARC 21 http://www.loc.gov/marc/
  • Object ID Project http://www.object-id.com/
  • VRA Core Categories, Version 3.0 http://www.vraweb.org/vracore3.htm

Components of a Catalog Template

A metadata schema or Catalog Template represents the way that data is organized within a collection. A Catalog Template consists of fields or groups of fields that represent one complete record. A Catalog Template represents the most complex data record that you plan to catalog.
More specifically, the components of a Catalog Template are:
Fields: Fields are the simplest units that can be cataloged, such as names, dates, or simply, text.
Records or Field Groups: Records are groups of fields and represent how data is organized either for display or logically grouped for cataloging.
Object Records: An object record represents a complete data record in Insight.

Fields

Fields are the lowest-level building blocks of a data structure. Each field represents one unique section of data. Fields can be Numeric, Short Text, Long Text, or IDs. Multiple Fields can be combined to create Field Groups and Records.

Field Validation

Depending on a field's use or content, it may be important to restrict input of certain types of information. Field Validation Rules enable the Insight administrator to require users to input valid data into a given field within Inscribe.
Insight Studio enables an administrator to add different types of validation rules, including checks for required fields, numeric and numeric range validation, verification against an external hierarchy such as the Getty's Art and Architecture Thesaurus, and date validation.

Special Validation – Controlling Access by User Rights

For more complex data input scenarios, administrators can control not only the validation of the content that is added via Inscribe, but also whether a given user can see, add, delete, or change data. This granularity of rights is often useful in managing controlled vocabularies, by allowing an administrator to only enable specific users to add new entries.
NOTE:To enable this feature, you will need to use the Administrator Tools.

Field Groups and Record Types

Field Groups and Record Types represent the substructures of a given record, with fields as the base components. Field Groups and Record Types enable you to organize common content for display and data entry. Record Types represent groupings of fields within your metadata schema. Field Groups represent groups of fields organized for display. In many cases, both Record Types and Field Groups contain the same fields; when building data models within Insight Studio, for example, creating a Record Type will create a corresponding Field Group. Some topics to consider:
What are the logical pieces of your metadata model? Think about how you organize your records. Are fields grouped? Do they repeat? What's special about them? Does a set of fields pertain to a creator? What do they do? How should they work?

Independent Records (an authority record)

Independent Records should be used when data fields are related more to each other than the main (object) record. For example, a single creator record may be created to describe an individual artist (eg. Pablo Picasso, 1881-1973), but it will be linked to multiple object records representing works by Picasso. The independent record enables catalogers to access and manage this information separately from the object record (accessible from the File | Open Record menu option in Inscribe).
Figure 1: Independent Record Diagram
Work 1 Les Demoiselles d'AvignonWork 2: Blue BoyCreator Name: Pablo Picasso
Creator Dates: 1881-1973
In the example above, both Work Records (Work 1 and Work 2) link to the same creator information (Creator 1). If the Creator Record is updated, both records will share the updated data.

Dependent Records

Dependent Records duplicate data and are merged into the parent record (in many cases, this is the base record). A metadata architect might use dependent records for something like a Work Title where a unique title needs to be created for each object instead of linking each record to a unique title record. Dependent Records can also be used to allow a field to repeat, by placing only that field within the record. Dependent records are not displayed in the form selection list in the File menu of Inscribe.
Figure 2: Dependent Record Diagram Main Work Record*Work Title:* Les Demoiselles d'Avignon
Work Title Type: Primary*Work Title:* Young ladies of Avignon
Work Title Type: Title Translation

Controlled Vocabularies

Controlled vocabularies are a special type of independent record. Like independent records, they maintain unique values, but can only contain one field. Like dependent records, within Inscribe, they do not show up in the File | Open Record menu.

Object Records

An Object Record represents one complete record in the data schema. If you were cataloging slides, it would describe the data for one slide. Object records are paired with images, or are associated to create Multi-Page Documents.

Search & Data Display Properties

By default, all fields in Insight are searchable and displayed within the data. However, some fields may not be useful for searching, and some data fields should be visible only to the catalogers. Sometimes data fields should be displayed, but not searched, or searched but not displayed.
Some examples might include:

• Fields that shouldn't be seen or searched (a cataloging Notes field)
• Fields that should be seen but not searched (a type qualifier such as a measurement unit)
• Fields that should be searched but not seen (the numeric versions of dates or OCR text)

When creating your Catalog Template, you have the ability to specify whether a field is searchable or not.
NOTE:You can always change this at a later date, using the Administrator Tools.

Considering Common Search Fields (Quick-Search Fields)

Quick-Search fields are offered as an easy way to search a collection. They are intended to provide guidance to users about useful ways to search a collection for specific records (i.e. the who, what, where, why, and how of your collection). Some topics to consider:
How do you expect people to search for images or data within your collection? What are the common fields that people will want to search by? Users may not want to search on repository name—especially if the repository is always the name of your institution.

Pulling it Together – Creating a Metadata Map

There is a lot to consider when creating a Catalog Template or metadata schema. Take the time to map out your template before going any further, and test your template by cataloging a few records by hand.
Table 1: Catalog Template

Cataloging a Few Records by Hand

Before implementing a data model, test it by cataloging a few records by hand. Testing a template has two main benefits: it can provide better guidelines to catalogers and will ensure fewer changes to the data model in the future.
If you have existing data: Select a few records. Find one or two "simple" records, one or two "complex" records, and one or two "normal" records that represent the collection. Complex records may contain many more fields than usual, repeating fields, longer text, or special characters, such as accents (è, ç, ?) or currencies (£, ¥). Finally, find one or two minimal records that represent the base of your system. The depth and breadth of the records will ensure that the Catalog Template you have created is a good fit.
If you don't have existing data: Create a few records and catalog them by hand to test out your schema.
This exercise is intended to see how well your data is represented in the metadata schema and to see if the schema can manage the data you've cataloged. If necessary, adjust the metadata schema to support the sample data.

Workflow and Metadata Creation:

Workflow is central to all tasks. However simple, all institutions and projects have a workflow. Workflow encompasses image processing, cataloging, data review, image review, and use. Workflow is the outgrowth of institutional policy, efficiency, availability of resources, as well as controllable factors.
When thinking about workflow, it's useful to consider:
How is your content created (metadata and media)? Is there a review process? Are there many individual catalogers and a supervisor? Are your images scanned or imported before entry of metadata or after? Do you generally process content in batches or catalog one item at a time? If the content is cataloged ahead of time, what format is it in?

Choosing the Best Tools to Fit your Workflow

Workflow is also important when understanding how to make best use of the tools provided. When planning to import content into Insight, it is important to understand how the tools manage the content and how to make best use of those tools. Within Insight, workflow is broken into three different stages: Data Import, Media Processing, and Record-to-Media Linking.

Creating a Base Record Template

One of the fastest ways to build a collection and have it available within Insight is to create a Base Record Template. A Base Record Template contains the set of fields with your data which are common to all data records, plus any information to map the media to the data record. This could include the image filename, copyright statement, unique ID, institution name, notes, processed date, or other information. See example below.

Working with Small Batches

If you plan to catalog records, import data, and import media in small batches (under 50 items), then you can:

• Import media in Insight Studio first, cataloging and linking records later in Inscribe.
• Import data first, importing and linking the media in bulk later using Insight Studio.

Working with Large Batches

If you have large batches (100+ items) of content to import (eg. delimited text files or XML), then it is best to import the metadata and then the media. After importing the data, it will be possible to review it within Inscribe. If the data is correct, use Insight Studio's media processing tool to process and link the data using a mapping file. The batch-linking section of the media processing tool will allow you to map the Object Records and Media Files and review the mappings before they are published to the collection.

More Complex Workflow Situations

Insight also supports more complex workflow situations. You may have a database that is used for both Insight and other applications. You may have procedures in which one group of users catalogs data that is subject to review by another group of users before publication.

If You Have Content in an Existing Database

If you have content in an existing database, then you could potentially make use of the Administrator Tools indexer. The indexer allows Insight to draw content from a set of external data tables and represent that content within Insight.
Cases where you would use the Indexer:

• If you have a collection with data published from another database, but don't need Inscribe to update.

• If you have a collection with data published from another database which you would like to use Inscribe to update.

NOTE: This is an advanced configuration that may be complex in some situations. Please contact Luna for more information on this configuration.
NOTE: It is STRONGLY suggested that you install a special Collection Manager just for the collection you will use with this source data.

If you Have Different Types of Users Importing and Managing Content

If you have different types of users with different responsibilities, then you can use field-level rights and permissions associated with the Approval Levels in the Administrator Tools. Approval levels allow administrators to restrict end-users from creating, editing, deleting, or linking records – enabling a more dynamic workflow. Common uses include establishing permissions for entry-level users to enter data which is then held for review by administrator level users before publishing to the collection.

Images and Multimedia

Insight is a visual tool, emphasizing the detail and quality of your media content while allowing you to organize and use it. Insight will take advantage of high resolution, archival quality media and facilitates the use of that content through the Java Client and BrowserInsight. Before creating your collection, think about the format of your source images and multimedia content and how to best facilitate the use of that content within Insight.
Many Insight customers find other uses for the derivative images that Insight creates for websites or other projects. Insight Studio creates up to six derivative files, depending on the size of the source image. Derivative sizes 0-4 (each doubling the size of its predecessor) are commonly created in JPEG format and stored in the Media Manager. Sizes 5 and higher are commonly created dynamically from the wavelet media file (JPEG2000 or MrSID). For a full list of image sizes, please see Table 14: Image Sizes in Insight on page .
NOTE:Using Administrator Tools, media processing Profiles may be adjusted to create MrSID, JPEG2000, or JPEG files at the requisite Insight sizes.

Considering Virtual Collections

The Insight Java Client, BrowserInsight, and XML Gateway support the concept of Virtual Collections – the ability to separate a subset of an existing collection into its own collection. Virtual Collections enable institutions to select content from private repositories for public use, or to use a single metadata structure to support many sub-collections.
Figure 3: A Visual Representation of a Virtual Collection Institutional Collection Virtual Collection (public use, subset of an Institutional Collection's content)
NOTE:Virtual Collections can be accessed from the Insight Java Client, BrowserInsight and the Insight XML Gateway as independent collections. Though Inscribe can be used to view and maintain data in collections that contain Virtual Collections, Inscribe does not recognize the Virtual Collections themselves.