Lore help sought: how might my players change a deity’s domains?

The short setup: my players believe they’ve angered a major deity and want to atone for their actions. Since the deity in question is Desna, who wants to reclaim the Beasts domain from Lamashtu, one of the "obvious" options is to help her achieve that goal.

I’ve been looking for lore about how PCs (and mythic ones at that) might be able to even nudge the cosmic balance such that one deity might take a domain from another.

I know the ultimate answer is "whatever the GM says works works". I’m hoping that there’s some obscure bit of lore that I’ve overlooked that might point in a helpful direction.

So: what, if any, precedent/lore/… is there in Golarion for mortals to change – or aid in a change to – a deity’s domains when there’s no chance of a negotiated transfer?

Feedback Sought on Java Reflection Factory Example

I set out to create an example or proof of concept of an idea I had and would be interested in receiving constructive feedback on my idea. I needed a Java factory pattern that allowed the end-user of the application to extend it (think add-ons, plug-ins, user contributed content, etc.). So I created a Java factory that uses reflection. While I am doubtful that I am the first to have done this, I have not found [any] examples of it. In my example, I created an abstract class called Vehicle. I then extended Vehicle a number of times (i.e. car, train, boat, airplane, etc.) and placed the subclasses in a separate package. In my example, my factory has a class scanner (does not currently support JAR files) that registers each subclass. Then the factory can be easily called using some class data (i.e. ID, name, whatever) to create a new instance of the subclass. I have included a link to the repo in GitHub (I hope that makes it easier to browse/review). What are your thoughts? Is this practical? Useful? Are there obvious improvements? Easily adapted to support JAR files?

I do want to note that I am aware of Google Guava and the Reflections Library. I wanted to create an example without dependencies on other libraries.

To see the underlying code, please follow link to GitHub repo. https://github.com/MannyPeterson/ReflectionFactory

USAGE

/*   * Create a new instance of the reflection factory by   * passing the package name to the constructor. The class  * scanner will look in this package for classes. All of the  * classes in this package must share the same superclass -  * Vehicle in this case.  */ VehicleFactory vehicleFactory = new VehicleFactory("org.codehamster.vehicles");  /*   * Call the getVehicle() method with the identifier ("name")  * of the class to create a new instance of.  */ Vehicle vehicleOne = vehicleFactory.getVehicle("boat"); Vehicle vehicleTwo = vehicleFactory.getVehicle("train"); Vehicle vehicleThree = vehicleFactory.getVehicle("truck"); Vehicle vehicleFour = vehicleFactory.getVehicle("airplane"); Vehicle vehicleFive = vehicleFactory.getVehicle("car");  /*   * Call each of the new classes with the println() method  * to invoke the toString() method and verify the output  * is correct.  */ System.out.println(vehicleOne); System.out.println(vehicleTwo); System.out.println(vehicleThree); System.out.println(vehicleFour); System.out.println(vehicleFive); 

OUTPUT

Vehicle is a boat Vehicle is a train Vehicle is a truck Vehicle is a airplane Vehicle is a car 

Reference sought: spaces with a 1-symmetric basis form a closed subset of the Banach–Mazur compactum

I am looking for a citable reference to the following statement.

Let $ {\rm BM}_n$ be the metric space of all norms on an $ n$ -dimensional space (modulo an isometry) with the Banach–Mazur distance. Then the set of norms that have a 1-symmetric basis is a closed (and proper) subset of $ {\rm BM}_n$ .

I am pretty sure that I have seen a similar statement somewhere in the literature… Many thanks.

Crawl History Explanation sought

With reference to an example Crawl History, below, please could someone explain what the figures actually report:

enter image description here

In particular, the last 2 Incremental crawls show large drops in the number of errors reported but not similar increases in the Successes. Assuming that there aren’t large numbers of deleted items, during this period, there thus seems to be a notable variation in the number of items found by the crawl! Thanks for any clarification you can give