[rkward-devel] Import preview? (CSV)

[d_jan]

Am 30.12.2015 um 11:37 schrieb Thomas Friedrichsmeier:
[SNIP]
>> Like this:
>>
>> Preview__________________↓__
>> Code_____________________↓__
> 
> [SNIP] What bugs me in this approach, is that
> 
> 1) […] > essentially a bit of a waste of space.

Yes. One could restructure other parts of the GUI etc. but if we keep
changes to a minimum (for good reasons) it is a waste of space.

> my thoughts keep returning to a button-like solution…

Yes, I think thats o.k., it is not really elegant, but a step forward (a
big one, if it enables us to have a preview) , and that’s what is
important.

Along these lines, I thought of
- having checkboxes instead of buttons
- Showing code and/or preview in a tabbed area. Having tabs has the
advantage that the tabs ("place-where-we-show-stuff") have titles, so
the user knows that "show code" matches to the "code" tab etc. So even
if only one thing (e.g. code) is shown, the tab performs a useful function.

What do you think of that?


> We also have the  concept of "Wizards" (e.g. there is one for
> Analysis->Means->t-Test).
> Rarely used, and perhaps not often useful

I like them actually – from the perspective of the beginner, they are
probably useful. In this case, the preview would be the last step, if
possible, or would be shown at the last step. Does that make sense?

  Happy new year!

Jan

> 
> 2) A second problem is more subtle: If both previews can be visible at
> once, we need some way to adjust their relative sizes. The obvious
> solution is a "splitter" (as is used to allow resizing the tool views
> in the main window, for example). In this setup, it is not clear to me
> (technically and conceptually) how the two would be combined,
> meaningfully. Esp. preventing states like preview X is expanded, but
> only sized header + 2 px (i.e. barely visible), etc.
> 
> Problem 2 would mostly go away, if only one preview can be visible at a
> time. That would not be _too_ bad, of course, but still leave problem
> 1. For problem 1, my thoughts keep returning to a button-like solution,
> rather similar to what we have, now (only enhanced a bit with
> collapse-expand indicators, and perhaps a different button style).
> 
> Your thoughts? Am I just too stubborn, or do you have any further tricks
> up your sleeves? (To increase complexity some more: We also have the
> concept of "Wizards" (e.g. there is one for Analysis->Means->t-Test).
> Rarely used, and perhaps not often useful, but something I would not
> want to give up, easily. How would we squeeze in a preview, there?)
> 
> Regards
> Thomas
>