DIRECTORY Output processing lack of clarification

The standard doesn't clarify how should the outputs of `type` `DIRECTORY` be processed, whether output specified as `DIRECTORY` should be recursively uploaded file by file, or whether outputs can be treated and/or specified as `ARCHIVE` or similar to make the default choice for the TES server to `tar` the directory and curl the archive to the specified `url` , I am working on an implementation of the TES Server, [tesp-api](https://github.com/CESNET/tesp-api).

Start with the standard's definition of `DIRECTORY` outputs and compare it with the implementation context described for tesp-api. Resolve whether directories are uploaded recursively or packaged as archives, then update the standard so the required processing and URL behavior are explicit and implementers can verify compliance.

Written by the indexing model from the issue text.