Languages English Español Register/Login   Welcome   How to Use   Search   Depositors    - Formats    - Access    - Metadata   Rights   Languages   Help   Links   Contact   AILLA papers   Events

Metadata - cataloging archive materials

Metadata is the information that describes each archive resource. Some of it helps us to process a new deposit correctly, like the title of each resource and the depositor's contact information. Other metadata is used by the search tools to sort through the collection, by speech genre or geographical region, for example. Metadata also gives archive users a summary of the contents of an archive resource. Metadata is very much like the information about books that is contained in a library catalog, but with more detail.

Depositors provide us with information about each item, including the names of the creators, the title of the work, a description of the intellectual content of the work, such as its genre and the context in which it was created, the date the work was created, and some additional information about other participants and the project for which the work was produced. AILLA staff add information about the original media that we receive, and about the digital file that we produce, such as its size in bytes and its length in minutes (for recordings). The metadata thus provides a complete record of the history and content of the resource. Each piece of information is entered into our relational database, and can be used in searches or examined in the results of a search.

AILLA's metadata schema conforms to the standards for language resources defined by both the International Standards for Language Engineering and the Open Language Archives Community. This means that you can use the same search terms at AILLA that you will find at any other conforming archive, and that you will be able to search for AILLA resources along with resources in many other language archives through the global search portal at the LINGUIST list. Standard metadata schemas make it easier for users to find what they want, and to make use of what they find.

The structure of AILLA's metadata

Archive resources usually come in groups of related items, which we call bundles. A typical bundle in the archive will include a recording (an audio file) with a transcription, translation, and perhaps other forms of annotation (text files). All of these items have essentially the same intellectual content; that is, they are all reflections of the same event in some sense. They were most likely produced by the same group of people for the same project, and should probably have the same title. We group these items together in the metadata description to save the depositors from having to enter the same information over and over again for each item, and most importantly, to make sure that users can easily find all the items in a bundle. Finally, AILLA produces additional items for each bundle, because we archive most items in two formats.

Some bundles may be very large, with dozens of items in them. For example, a book based on several performances of verbal art would be digitized as a bundle with these elements:

1. a text file for the book,
2. audio files in two formats (probably with 1-minute sample files) for each performance,
3. additional text files with the transcriptions and translations that the book was based on,
4. some photographs of the performers.

Other bundles may actually consist of a single item, such as an article or a dictionary in text format. (Although it would be more interesting to create a new version of this text dictionary, with links to lots of little audio files that contain examples of each word in context.)

Depositors decide how to organize their materials into bundles, and provide descriptions of how the items in each bundle are related. Users can then use this information to reassemble the bundle correctly.

Entering metadata

Filling out the metadata forms is work, but detailed descriptions of resources make the archive more useful for everyone. You may find that this metadata is also a useful tool for organizing your own corpus for your own purposes.

If you are unable to provide a complete metadata description at this time, you may fill out the Metadata Short Form, online or on paper. You can update the metadata information for any resource that you have deposited whenever you wish: we keep track of what you have deposited in your user account record.

We offer a variety of methods for defining your metadata:

1. Online Metadata Editor. (Not yet available.) This interface allows you to fill out either the Metadata Short Form or the complete set. Your information is saved in the appropriate format at AILLA. Go to the AILLA Online Metadata Editor.
2. Offline Metadata Editor. The Technical Group at the Max-Planck Institute for Psycholinguistics has built a suite of tools for managing language data, including a very nice Metadata Editor that you can use on your own PC. The ME will generate a file in the format that AILLA requires (XML). You can either email this file, or print it and send it with your package. Our addresses are on the Contacts page. Download the MPI Metadata Editor.
3. Paper Metadata Forms. You may find it more convenient to fill out paper forms and mail them to us along with the materials you are depositing. AILLA staff will then enter the information using the Online editor. There are two sets of forms with instructions: the Metadata Short Form and the Metadata Complete Form. Download Metadata Forms

Remember that metadata describes archive resources - stories, poems, films, dictionaries - NOT storage media, like tapes and diskettes. Please label everything in your deposit clearly so we can match metadata with the individual resources on the storage media correctly. It is especially important to give a title to each item on each of the media. AILLA will assign a unique identifier to each item that we archive, but this is a coded form that is not designed for readability. Titles are more meaningful for users in search results, they look better when the item is cited, and they will help us be sure that we have grouped items into bundles correctly.

Example: Let's say you're depositing two tapes and three diskettes. You could label the tapes T1 and T2, and the diskettes D1, D2, and D3. Then you could identify each item on each medium with a title, using the same title for everything that belongs in the same bundle. You can write a complete title on the metadata form and use an abbreviation (short name) on the media, if you wish. So, T1 might contain the recrodings of Sato, Dos Pie, and Rebelde, while D1 contains the text files with the same names. T2 might contain a long recording on both sides labelled Entrevista, while D2 and D3 contain files labelled Entrevista trans and Entrevista pic1, Entrevista pic2, etc. This deposit thus contains a total of 4 bundles: Sato, Dos Pie, Rebelde and Entrevista. The metadata for each bundle will include the full titles and an explanation of the contents of each item, as well as a description of how the items are related.