Sometimes, you just have to pull the trigger.
The thing about writing is, once you pull the trigger, you can edit the gun out entirely.
Researching historical fiction set in Late Antiquity has some tricky bits attached. Scholars love writing about Rome, and though, say, the world of Theodoric the Great gets attention along the way, the details about his furniture and sleeping habits - and, say, the schooling of his daughter - are less attended upon. Ironic given these are royals and all.
So, you go in for, say, an attempt to name a slave in the royal household at Ravenna, and you get all sorts of information: about Rome. Very quickly, you begin to note that highly similar tidbits repeat in different sources, none of them *quite* addressing exactly what you need, and yet all of them reflecting one another. This alone can be instructive, even if it's not to the point you wanted to drive to.
Many later Roman slave names were Greek. Not all of the holders of these names were Greek, by, apparently, a long enough shot to mention it.
Slaves' origins were a noted point in buying or assigning them. There were stereotypes of Egyptian and Briton slaves, there were expectations about types of work and types of workers. The concept of "wish-names" - slave names indicating desirable traits or accomplishments - is especially intriguing. "Hedone" is a poignantly telling sobriquet for a woman available for sale.
And then your question becomes: how much does this Roman research apply to my only semi-Roman setting?
How much can I USE, when discussion of place-settings (guitarists please note, this anachronism is an intentional joke) and sleeping habits for the Ostrogoths is less than ubiquitous?
And then the question becomes: how long before I stop thinking about the guitarists, trust myself and my research, and focus on the story ... ?
Research is a wonderful way not to write, sometimes. It's a great excuse, believing "I have to get it right, before I write."
And it's so easy to forget: anything I write, I can edit. The fat lady doesn't sing until you have a contract; even at the query stage, you are still allowed to correct yourself, if you find you actually did write a firearm into a scene starring Theodoric the Great's only daughter, in the year 535. Even when you have an agent - if you're lucky and open to it, an editorial one - the book's not done until the publisher sticks a fork in and it's tender.
Sometimes, you have to pull the trigger.
I've been writing. How about you?