| 18537 |      1 | 
 | 
|  |      2 | text {*
 | 
| 20429 |      3 |   In practice, super-contexts emerge either by merging existing ones,
 | 
|  |      4 |   or by adding explicit declarations.  For example, new theories are
 | 
|  |      5 |   usually derived by importing existing theories from the library
 | 
|  |      6 |   @{text "\<Theta> = \<Theta>\<^sub>1 + \<dots> + \<Theta>\<^isub>n"}, or 
 | 
|  |      7 | 
 | 
|  |      8 | 
 | 
|  |      9 | 
 | 
| 18537 |     10 |   The Isar toplevel works differently for interactive developments
 | 
|  |     11 |   vs.\ batch processing of theory sources.  For example, diagnostic
 | 
|  |     12 |   commands produce a warning batch mode, because they are considered
 | 
|  |     13 |   alien to the final theory document being produced eventually.
 | 
|  |     14 |   Moreover, full @{text undo} with intermediate checkpoints to protect
 | 
|  |     15 |   against destroying theories accidentally are limited to interactive
 | 
|  |     16 |   mode.  In batch mode there is only a single strictly linear stream
 | 
|  |     17 |   of potentially desctructive theory transformations.
 | 
|  |     18 | 
 | 
|  |     19 |   \item @{ML Toplevel.empty} is an empty transition; the Isar command
 | 
|  |     20 |   dispatcher internally applies @{ML Toplevel.name} (for the command)
 | 
|  |     21 |   name and @{ML Toplevel.position} for the source position.
 | 
|  |     22 | 
 | 
| 20429 |     23 | *}
 | 
|  |     24 | 
 |