I'd be interested to know exactly what you think Csound1 is "plain wrong" on.
Here's the overview of CalDAV from calconnect.org:
CalDAV
CalDAV is a calendar access protocol and is defined in RFC 4791 [6]. The protocol is based on WebDAV which is an extension to HTTP that provides enhanced capabilities for document management on web servers.
CalDAV is used in a variety of different environments, ranging from very large internet service providers, to large and small corporations or institutions, and to small businesses and individuals.
CalDAV clients include desktop applications, mobile devices and browser-based solutions. It can also be used by "applets", for example, a web page panel that displays a user's upcoming events.
One of the key aspects of CalDAV is its data model. Simply put, it defines a "calendar home" for each calendar user, within which any number of "calendars" can be created. Each "calendar" can contain any number of iCalendar objects representing individual events, tasks or journal entries. This data model ensures that clients and servers can interoperate well.
In addition to providing simple operations to read, write and delete calendar data, CalDAV provides a querying mechanism to allow clients to fetch calendar data matching specific criteria. This is commonly used by clients to do "time-range" queries, i.e., find the set of events that occur within a given start/end time period.
CalDAV also supports access control allowing for features such as delegated calendars and calendar sharing.
CalDAV also specifies how scheduling operations can be done using the protocol, rather than, for example, sending scheduling messages via email. Whilst it uses the semantics of the iTIP protocol, it simplifies the process by allowing simple calendar data write operations to trigger the sending of scheduling messages, and it has the server automatically process the receipt of scheduling messages. Scheduling can be done with other users on the CalDAV server or with calendar users on other systems (via some form of "gateway").
How Csound1 describes it, is how it works. It reads and writes data to and from the server. No comparison is done of the data - which is what "synchronizing" would do.