<\body> content model> All documents or document fragments can be thought of as , as explained in more detail in the chapter about the document format|../../format/basics/basics.en.tm>. Inside programs, there are two main ways to represent such trees, depending on whether one manipulates active or passive documents: trees> Passive documents, like those which are processed by aconversion tool, are usually represented by . For instance, the fraction <\equation*> |b+c> is typically represented by <\scm-code> (frac (concat "a" (rsup "2")) "b+c") This representation is convenient in the sense that they can be manipulated directly using standard routines on lists. Active documents, like ones which are visible in one of the editors windows, are rather represented using the internal C++ type , which has been exported to via the glue. When a tree is part of a real document inside the editor, the tree is aware about its position inside the document. Using routines from the tree API, you may then make changes in the document simply by assigning new values to the tree. For instance, consider the following experiment: open two windows and start a session in each window. In the second window, enter the lines <\session|scheme|default> <\input|scheme] > (use-modules (utils library tree)) <\input|scheme] > (define t (buffer-tree)) In the first window, you may now modify the document in the second window using commands like <\session|scheme|default> <\input|scheme] > (tree-set! t (tree 'document (string-\tree "First line.") \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ (string-\tree "Second line."))) <\input|scheme] > (tree-set t 1 (string-\tree "New second line.")) <\input|scheme] > (tree-set t 0 (tree 'strong (tree-ref t 0))) From the last three lines in above experiment, it becomes apparent that it is quite cumbersome to manipulate trees using the standard tree constructors. For this reason, provides a hybrid type for manipulating scheme trees and C++ trees in a common framework. For instance, the last three lines in the above experiment may be replaced by <\session|scheme|default> <\input|scheme] > (tree-set! t '(document "First line." "Second line.")) <\input|scheme] > (tree-set t 1 "New second line.") <\input|scheme] > (tree-set t 0 `(strong ,(tree-ref t 0))) More precisely, a scheme expression of the type is either a string, a tree or a list whose first element is a symbol and whose remaining elements are other expressions of type . provides several routines (usually prefixed by ) for basic operations on content, like , , list>, , etc. Most higher level routines are built on top of these routines, so as to accept arguments of type whenever appropriate. Besides the fact that trees remember their inside the global edit tree, it is also possible to create cursor positions inside the global edit tree, which are naturally updated when modifications take place. This technique is useful when you want to write an editing routine which does not act locally at the cursor position. For instance, the following routine can be used to insert content at the start of the current buffer in a reliable way: <\scm-code> (define (insert-at-buffer-start t) \ \ (with-cursor (path-start (root-tree) (buffer-path)) \ \ \ \ (insert t))) The macro temporarily changes the cursor position, while storing the old cursor position in such a way that it will be updated during changes of the document. The user may also use the more explicit routines , , and to manage persistent positions. >