What does a license actually allow when a database or data file can be copied onto devices? That question sounds simple. In practice, the answer often sets the whole shape of use.
I look first at the license grant. Here, the right is narrow and tied to one user. The data may be used, and a file containing it may be installed, but only on a specific set of personal devices: one desktop computer, one laptop, and one mobile device such as a phone or e-reader. That is a device rule, but it is also a user rule. The permission belongs to a single person, not to a household, a lab, or a department.
This matters because many people assume that “downloadable” means “freely portable.” It does not. A license can permit local installation and still keep tight control over who may use the material and where it may sit. The file may live on several devices, but the license still belongs to one user only. If another person opens the same data under that license, the use is outside the stated terms.
For libraries, the limit is different but just as strict. A professionally managed academic or public library may place the data in an ebook-management system for view-only access by one reader at a time. That is a useful allowance, but it is not a broad campus license. It resembles a single-seat arrangement, not a shared pool. The phrase “one reader at a time” does the real work here. It tells us that access is controlled at the point of use, not simply at the point of purchase.
The document also points to a second layer of limits. A user may create static publications that include some of the data or data derived from it, but only in a narrow setting. The output must support third-party analysis inside an established research platform. It is not a general right to extract, remix, and redistribute the data. The license draws a bright line between limited analytical reuse and broader reuse for other purposes.
That line becomes clearer in the rules on export. Only an insubstantial portion may be taken out. In plain terms, that means a small amount with no real standalone value and no ability to substitute for the source itself. It also cannot be something that is marketed on its own or pulled out by regular batch jobs. This is a familiar licensing idea in a sharper form. The publisher is saying that the source may inform analysis, but it may not become a shadow copy of the original.
Here the phrase “derived data” needs plain language. It means data that a user has transformed from the source by applying calculations, analytics, or another method. Even then, the license keeps control. If the derived material is exported, it still must stay inside added-value reports or presentations that are part of the user’s normal services. It cannot be turned into a stand-alone product, a mass mailing, or content for external websites. The license allows a narrow path for analysis. It does not open a broad path for redistribution.
A small example makes the limit easy to see. Suppose a researcher uses the data to build a short chart for a client report. The chart is part of a larger analysis package, and the excerpt is tiny. That fits the licensed pattern more easily than posting the same chart on a public webpage or placing the underlying dataset in a newsletter sent to many readers. The difference is not cosmetic. It is the difference between embedded analysis and public reuse.
For librarians and digital-humanities workers, the practical lesson is simple. Device access and reuse rights are separate questions. A file may be installable on several personal devices and still remain tightly locked to one user. A library may be able to host access for one reader at a time and still be barred from wider sharing. The terms tell us what can be stored, who can see it, and how much can be moved outward.
I also notice how much of the license is built around subtraction. It tells the user what cannot be done more clearly than what can. That is normal in electronic licensing. But it means the reader has to track three things at once: the number of devices, the number of users, and the amount of data taken outside the source. Miss one of those, and the use may stop being licensed.
So the key idea is this: installation rights are not ownership. They are a defined permission with a user limit, a device limit, and a reuse limit. Once that is clear, the rest of the license reads less like a blur of legal language and more like a map of boundaries.
After this lesson, the reader can tell the difference between installation rights, shared access, and reuse rights. That is the real value of being specific. The Source List aims for one digital source worth knowing, one search tip, and one honest limitation, and this kind of license is exactly the sort of limitation that deserves plain words.