METS and ALTO – popularly, METS/ALTO – are terms that are thrown around when you start talking about digitizing and hosting your collections. Not all digital asset management programs, or DAMs, require METS or ALTO files, but a handful do (Veridian being a notable example), and if you’re considering article level segmentation, or ALS, then you will need to provide METS/ALTO files alongside your digitized collection.
Whoa, now – that’s a lot of acronyms. Let’s break these terms down a bit before we start to explain just what a METS file is doing, how ALTO files tie in, and whether generating METS/ALTO data is the right choice for you.
- METS: Metadata Encoding and Transmission Standard
- DAM: Digital Asset Management
- ALTO: Analyzed Layout and Text Object
- ALS: Article Level Segmentation
This blog was created with the assistance of Paul Boyko, one of our Backstage programmers who specializes in these XML files. Thank you, Paul!
METS – What is it, and what does it do?
Starting from the top, the best way to describe your METS file is to call it your ringleader, or maybe even a conductor. Your METS file knows everything there is to know about each page or bound collection of images that you’ve digitized. It’s going to help your DAM interpret where data is, how it’s been organized, and even monitor for data degradation. It provides descriptive, administrative, and structural metadata to describe your digital objects, one METS file per singular item. This item could be a poster or a single issue of a newspaper consisting of multiple pages. This is up to you, and typically, the organization follows the same way in which your materials were first archived or bound.
What’s in a METS file?
Your METS file, structured in XML, will have several sections.
- METS Header – This section introduces information about the file, such as who created it and when.
- DMD: Descriptive Metadata Section – An outline of one or more metadata files included in your digitized asset, the DMD section will explain how your Dublin Core, MARC, MODS, or other metadata fit into the asset, either by pointing to them externally or including them internally – or both.
- AMD: Administrative Metadata Section – AMD is divided into four subsections, describing the technical information behind your asset, the copyright and licensing information, the source, and its digital provenance.
- File: File Section – This speaks to the relative position (within the directory) of all files based on how your DAM interprets data. If METS is the ringleader, File Section is the translator and crossing guard, indicating to the DAM where the TIFFs, JPEG2000s, the single or bound PDFs, and/or ALTO files (or other OCR related derivatives), etc., are stored within the asset. It’s also able to keep tabs on the integrity of those contained files by using “checksum” actions.
- Structural Sections: Maps and Links – As you might infer from the blanket name of these two sections, mapping and linking elements in your METS file provides organization for what the end-user will see, and links within those pages.
- Behavior Section – This can be used to create behaviors between different content elements in the asset.
For more specific descriptions of these sections, please visit the Library of Congress’ standards page for METS, found here.
The Library of Congress hosts all the latest information and resources on METS on their website . METS is a standardized schema, so some of the guesswork of “how do I structure my XML?” is answered for you. Even still, sorting through the rules and validations required can be daunting, which is why Backstage offers this service to its digitization clients.
How does ALTO fit into this?
An ALTO file provides word contents, style, and the layout of a digitized page containing text including each character’s spatial coordinates. Typically, the ALTO file is per image/per page, and it contains all of the data necessary to provide OCR within the DAM. OCR accuracy depends upon the quality of the original printed material (unless manual correction is performed). With good results, the patron will be able to search by keywords and see those terms highlighted on the image itself. The content provided can then be viewed or even copied for use in their research.
Do I need ALTO files if I’m digitizing my collection?
If your collection is image heavy, you might decide that having text transcription and OCR is unnecessary. Certainly, if your collection consists of images only, then there’s no need to generate ALTO! Despite the fact that METS and ALTO are frequently tied together in conversation, a METS file does not require ALTO.
That said, if you have a text-heavy collection, adding ALTO files generates an additional, and very valuable, layer of access to your materials. Some DAMS are able to generate OCR on the server, but this may not be as reliable or accurate as a stable-state series of ALTO files.
Do I need METS files if I’m digitizing my collection?
This depends on your DAM, your end goal for the collection, and even just preference. You SHOULD create METS if:
- Your DAM requires the addition of a METS file. As mentioned before, Veridian is one that needs METS to host your collection. Do some research beforehand or reach out to your DAM provider to check for this and other requirements.
- You’re already planning to generate ALTO files, the same mechanisms used to generate ALTOs tend to also create a “standard” METS. More-customized METS may require an additional process.
- You’d like to receive some of the other benefits of METS beyond its being a vehicle towards uploading data to a DAM. The METS file and the checksums within the XML record a snapshot of everything as it is at the time of its creation. Since everything interlocks, it’s important that everything within the asset be the same as when it was created. So, let’s say a file has been edited accidentally, or an image within the asset begins to degrade (it’s strange, but it can happen!). Your checksum can tell you if something has changed. It’s a matter of opinion, but without METS, “…The information might not be as rich or in your control,” explains Paul. “The METS file is an agreed upon schema/standard. What you put in there, how you put it in there, is in your control. “Intentional are the wise.””
Lastly, you SHOULD create METS and ALTO files if: you intend to provide ALS. The default for OCR collection and display is PLS, or page level segmentation. This means that all of the information is divided by page. Let’s say you wanted your patrons to be able to see each article on its own within a newspaper. Since articles tend to jump around, sometimes starting on the cover of a newspaper and ending several pages inside, you’d have to define where an article begins, where it ends – and that takes a lot of technical organization. As a result, if you think you’re going to do ALS in the future, you definitely should generate METS/ALTO.
Backstage is able to assist with all end-processing needs for your digital collection above and beyond just METS/ALTO mapping. To learn more about pricing or our workflows, you can call us at 1.800.288.1265, send us an email at moc.wlsb@ofni, or fill out our Contact Us form.