Docs suggest creating a scraper in a directory that doesn't exist

[palewire]

[zstumgoren]

We haven't yet integrated a place-specific scraper (i.e. a court Site class that only works for one county), so this was intended as a prescriptive bit of documentation for when we do integrate that first non-platform site. But I could see us just putting all scrapers in platforms, even if it is only a site relevant to one county. That would simplify the workflow and the code base in spots like these:

• base/runner.py
• site.py

@palewire Thoughts?