The Archiving Problem

It's not necessarily clear what is the best way to archiving information for the long haul. Carving into stone, incising into ceramic tablets, or writing on parchments using stable inks made of iron-salts are all proven approaches that can last centuries. Photographs, especially silver-halide prints, apparently can last at least a hundred and fifty years, and perhaps much longer.

When it comes to digital files, however, there's a lot less certainty. There's much greater resilience to physical destruction because it's so much easier to copy the information — and you do keep backups, don't you? Still, I have Kodak digital camera files from 1998 that can only be read by installing legacy programs into a Windows 95 virtual machine. While it will almost always be possible (with enough effort and reverse engineering) to convert historical data formats, it may become prohibitively difficult for rare or obscure formats.

The JPEG image format is arguably the most popular digital image format in history. Because of its wide use, it is a reasonably safe format to use for preserving archival images. There will likely be programs that will understand how to read these files for a very, very long time. There are arguments to be made that the lossy aspects fo JPEG make it less desirable, but for most purposes, I believe the format is a safe bet.

labeled image

Metadata Fields

The JPEG file standard contains a special area for application-specific fields. These fields can contain any number of records. Of interest to us, however, are the Exif, IPTC, and XMP fields:

The Exchangeable image file format (Exif) was established in the late 1990s to contain mostly technical information about images and recordings. It has very limited support for "user" metadata.

The International Press Telecommunications Council's Information Interchange Model headers (IPTC) were originally designed to store photographic metadata for news media, but have been widely used by photographers of all kinds. Because of this wide use, most image viewing software will recognize the key IPTC fields embedded in a JPEG image file. However, in recent years, Adobe's Extensible Metadata Platform (XMP) has gained in prominence. Based in part on the IPTC core, the XMP record can contain more diverse data. Adobe's dominant position in the digital imaging world has helped assure that the XMP standard is supported by most modern imaging software.

Technical Details

The Legacy Labeler stores your information according to both the IPTC and XMP standards. The information you provide in the form gets mapped into relevant fields (based largely on the author's interpretation of the field purposes).

FieldEXIF fieldsIPTC FieldsXMP Fields
TitleObject Name, SubjectDublin Core Title
CaptionCaptionDublin Core Description
HeadlineHeadlineAdobe Photoshop Properties Headline
AuthorImage.ArtistByline,CreditDublin Core creator, Adobe Photoshop Properties Credit
CopyrightImage.CopyrightCopyrightDublin Core rights
SourceSourceAdobe Photoshop Properties Source
Contact EmailIPTC Core CreatorContactInfo/Iptc4xmpCore:CiEmailWork
Contact PhoneIPTC Core CreatorContactInfo/Iptc4xmpCore:CiTelWork
KeywordsKeywordsDublin Core subject
URLIPTC Core CreatorContactInfo/Iptc4xmpCore:CiUrlWork, XMP Rights WebStatement
CityCityAdobe Photoshop Properties City
StateProvinceStateAdobe Photoshop Properties State
CountryCountryNameAdobe Photoshop Properties Country
Country CodeCountryCode
DatePhoto.DateTimeOriginalDateCreatedAdobe Photoshop Properties DateCreated, Basic Properties CreateDate

What About Other Image Formats?

Unfortunately, the standards are not as well-defined for metadata in many other file formats. Legacy Labeler will gamely try to do the right thing if you upload Portable Network Graphics (PNG) files or Graphics Interchange Format (GIF) files. If there was existing metadata in a file of one of those formats, Legacy Labeler may not be able to extract that to preload the form. Similarly, when it writes the metadata into images of those formats, it may or may not be readable by a given image viewer.

So for non-JPEG formats, I'm afraid our lack of guarantee goes double.