One of the last minute changes in Bucket Explorer 2.0, which delayed its beta release was a change in the way we upload empty folders.

Amazon S3 does not have any concept for folders and files within a Bucket.  When we upload files to a bucket, we don’t need to create a “folder” on S3 to maintain a folder & sub-folder structure. We just use a “/” in file name and when displaying that file, we treat the first part of the file name as folder.

To upload empty folders, which don’t have any files in them, we used a naming convention which was used by a few other S3 client UIs at the time when we released Bucket Explorer 1.0.  We uploaded an empty file with extension _$folder$ to represent an empty folder.

In Bucket Explorer 2.0, we have changed that extension. Now empty folders will be uploaded using “/” at the end instead of “_$folder$”. We have also eliminated the need for these empty files for listing folder structure. You only need these files on S3, if there is an empty folder.

The result is that now you can use Bucket Explorer to explore files uploaded to S3 using virtually any S3 client, including support for Jungle Disk 2.0’s compatibility buckets.