Hans Michel wrote: > Please see the thread from 8/10/2010 " Frontend suggestions". You > are not the only one thinking of this.
It's good to know that I am not alone. :-) And yes, your suggestion of Compartment (or Page, or Minipage, or SuperCell or Gizmo) is exactly what I had in mind. Even if WRI implemented a non-flowing version, it would be a giant leap forward.
> Anyone who has taught or sat in a class room with a blackboard when you run > out of vertical space you move horizontally to the next blackboard segment. > The analogy fails for Mathematica because you should be using it to solve > any initially stated problem, not to write out the steps. So if the > application is working as designed you should get an answer in the next > Out. But sometimes I want the Out in another column next to the In; > not vertically below it. > > We need compelling reasons other than I want my paper to look like journal > X's 2 column format. > > Why would we need alterative composition in the FrontEnd?
Well, from my point of view there are plenty of reasons. So many I am not sure I can name them all.
First, as you wrote, a two-dimensional layout allows for a better exploitation of the screen real estate. This means that instead of spreading text and formulas on a long ribbon that forces me to scroll down and lose sight of the first part of a paragraph, I could have everything I need to veiculate a concept on the same page (be it a window pane or an actual printed page). Forcing a reader to turn back and forth the pages to grasp a concept was not a good practice even in 19th and 20th century book-writing.
Being able to put the output of a cell right next to the input that produced it (and the subsequent comment/explanation/text) is one big step toward this goal. It's not always desirable or necessary, granted, but when the input and output are short a lot of space (should I call that 'field of view) get wasted (especially after the DVD-oriented craze for computer monitors).
But a two-dimensional layout is also useful to better veiculate concepts: it might be used to put side by side competing or analog descriptions. (before/after, right/wrong, predicted/measured, simplified/actual, electrical/mechanical RL-circuit/RC-circuit, resistance/reluctance... you name it). It's easier and faster to grasp differences or analogies in this way that it is when the competing descriptions are one under the other. (Alignment on the same row is the key to have it 'all at a glance' without losing time to find the row we were reading before jumping to the the alternative version)
Moreover, two-dimensional layout is the natural choice for _real-world_ notebooks. A diagram, a sketch, even formulas end up side by side with text, there. Same reasons as before: not wasting paper, yes, but also having related pieces of information at the same 'eye-level'. It's easier to associate one to the other, it does not make the reader struggle to find the 'jump point' because it's right next to where he/she is looking now.
There also is an aestethical reason: a 'full' page looks much better than a half full page. This might not be a compelling reason (and there could obviouly be exceptions), but seen the progress made in making Mathematica's output so professional looking, it strikes me that this *basic* aspect has not been pursued any further.
And yes, one might also want to have his/her notebook look like an article from journal X. I don't consider this to be relevant as the other reasons, but it's a further reason nonetheless.