Refactor the chapter on publishing

[adswa]

This work-in-progress PR starts a larger effort on restructuring and extending the handbooks content on publishing datasets to third party services.

In particular, it

• separated service-specific walk-throughs into standalone sections. This affected the "intro" section in particular, from which content on configuring dropbox as a special remote, exporting dataset contents to figshare, and using Git-lfs have each been transformed into standalone sections in the chapter. At the moment, there are walk-throughts for dropbox, s3, figshare, git lfs and gin. This new structure may also make it easier to add new services as walk-throughs.
• improved the introduction into the topic of dataset publication. I have split it into different usecases depending on the capabilities of the employed third party services. The intro lists them, and links to the relevant walk-throughs later in the chapter.
• added a standalone section on data privacy
• extended the example on exporting datasets to figshare with screenshots and code

There is still missing content and room for improvement:

• The placeholder section on publishing datasets to git repository hosting services needs to be filled with the current commands and procedures to do this
• The placeholder section on publishing datasets to git repository hosting services needs to document the new create-sibling-{github/gogs/gitlab...} command that will come in 0.16.0
• The walkthroughs might benefit from a more generic setup, such as the one that @jsheunis has created in the S3 workflow, where the section is a complete stand-alone piece including the generation of a dataset to publish. If we adopt this approach for the other walk-throughs, too, their content can become more accessible and would not require readers to e.g., have a "DataLad-101" dataset from the previous chapters to publish.
• the section on data privacy could be extended with more strategies and ideas or examples - at the moment, its a lose and hastily collected set of ideas. @mslw maybe has more content based on his work on the RDM module on this topic