Best Practice For Updating Single Field In N Rows

Hi

I haven't included DDL etc as this is theoretical at this time. I can rustle up an illustrative example if required.
Following applies to disconnected environment.

In general, most edits are broad (i.e. n fields affected) against a single record. However, there are certain circumstances where a single field will be edited against a deep (n) set of records.

So business request:
User needs to access and edit n records at a time (for arguments sake n is unlikely to be enourmous - say 50 max) but only editing one single field, and always the same field. The values for this edited field will differ for each record. Initially retrieving data for the app is no problem, nor is identifying those records that have been edited. What is the best means of updating the table?

There are, to my mind, three ways of dealing with the latter senario -
1) Client calls the server n times editing a record at the time.
2) Client creates a csv string and passes to sproc. Sproc parses string using some UDF split function (probably chucking into table variable) and updates using a single set set based operation.
3) Client creates some flat file or other that is BULK inserted (or similar) by SQL Server.

3 - I think would only be an option if n was a very large number and or having the changes immediately reflected in the data is not priority.
1 - would create a chatty app and presumably put the most load on the server.

So I am left thinking 2 is (depending on circumstances) the best method. Is this fair? Are there any particular considerations gotchas I should be aware of? I know where to get hold of as many TSQL split functions as I could want so I'm not looking for code just opinions.
OR - am I just plain wrong? Is there a better alternative or am I dismissing the other two methods prematurely?

Thanks in advance