<h3>Save is a byproduct</h3>
**Save** is a byproduct of early hardware- and software design. It doesn't have a common equivalent in the real world.

Consider: If you take a pencil and make a mark on paper, that mark doesn't require an extra step in order to become permanent.

[![A pencil mark doesn't need to be made permanent][1]][1]

In other words, it does not need to be saved. The paper may need to be stored somewhere so it can later be found, or copied so it can be shared, but those are different task than the act of permanently marking the paper.

Old-fashioned **Save** became about more than saving. It was combined with tasks such as these:

 - Decide where to put the file.
 - Name the file, so it can be recognized.
 - Choose the file format, with an eye to sharing it.
 - Choose to make copies, by using **Save as**, perhaps as backups.
 - Choose the file format, with an eye to faster hardware performance.
 - Choose the file format, with an eye to forcing the software to reformat the file to "fix" a defect in the file.
 - More...?

The no-save approach reflects what happens in the real world, and so requires no learning—except unlearning to **Save**.

 - Stuff end up in default locations.
 - Stuff can be renamed as a separate task.
 - Stuff can be shared automatically, or shared manually as an explicit task.
 - The *SAME* stuff can be shared, and read and modified by many at the same time, removing the need for copies.
 - Stuff can be converted or exported as an explicit task.
 - Stuff versions itself, and users can roll back to an earlier version.
 - More...?

Google and others have taken the brave step of challenging the paradigm. Having users *save* when they really want to *back up*, *share*, *export*, *move*, or *identify* stuff requires the user to adapt to the software and puts the focus on the software rather than on the user's goal or the user's content.

What do you think?

  [1]: http://i.stack.imgur.com/cGyJ4.png