<li>New help system, browsable in the UI and via the HTTP server (<code>/responder/help</code>). In the UI listener, invoke <code>handbook</code> to read the documentation root, and invoke <code>\ foo help</code> to look at documentation for the word <code>foo</code>.</li>
<li><code>hash* ( key hash -- [[ key value ]] )</code> is now <code>hash* ( key hash -- value ? )</code></li>
<li><code>hash-clear</code> is now <code>clear-hash</code></li>
<li><code>hash-each</code>, <code>hash-each-with</code>, <code>hash-all?</code>, <code>hash-all-with?</code>, <code>hash-subset</code>, <code>hash-subset-with</code> now pass the key and value separately on the stack to the given quotation, instead of passing a cons cell</li>
<li>Literal syntax change: <code>H{ [[ key value ]] ... }</code> is now <code>H{ { key value } }</code></li>
</ul>
</li>
<li>Math:
<ul>
<li>The <code>sum</code> and <code>product</code> words have been moved to
<code>contrib/math/</code>.</li>
<li>The <code>mod</code> word is now supported for ratios and floating point numbers.</li>
<li>The <code>truncate</code>, <code>floor</code> and <code>ceiling</code> words are now supported for floating point numbers.</li>
<li>The NaN, positive infinity and negative infinity floating point numbers now parse and unparse as <code>0.0/0.0</code>, <code>1.0/0.0</code>, and <code>-1.0/0.0</code> respectively.</li>
<li>The NaN value is now equal to itself under <code>=</code>.</li>
<li>Negative and postive zero are no longer equal under <code>=</code>. However, the new <code>zero?</code> word tests if the top of the stack is a zero, and it tests for both positive and negative zero.</li>
<li><code>stream-format ( string style stream -- )</code> now takes a hashtable
rather than an association list for specifying style information.</li>
<li><code>stream-write</code> and <code>stream-terpri</code> are now generic words, and there is a new <code>with-nested-stream</code> generic word. You can wrap your output streams in a <code><plain-writer></code> to avoid implementing these.</li>
<li>The listener looks different now. An expandable top area is used for browsing objects, words and help, and the stack display has been shrunk to a single status line at the bottom of the window.</li>
<li>A left click on a presentation now invokes the default command. A right click
<li>Added expandable outliners. Used by the inspector, <code>.s</code>, <code>usage.</code>, <code>uses.</code>, <code>vocabs.</code>, and various other words.</li>
<li>Added word completion to the listener pane; press <code>TAB</code>.</li>
<li>Added word navigation shortcuts to the listener pane; press <code>C+LEFT</code> and <code>C+RIGHT</code> to move a word at a time, and <code>C+BACKSPACE</code> and <code>C+DELETE</code> to delete the previous and next word, respectively.</li>
<li>Some new sequence words that do not bounds check: <code>nth-unsafe</code> and <code>set-nth-unsafe</code>. These should only be used in special circumstances, such as inner loops (<code>each</code>, <code>map</code> and so on use them).</li>
<li>Easier exception handling. The <code>cleanup ( try cleanup -- )</code> word encapsulates the following idiom:
<pre>
[ A ] [ B rethrow ] catch
[ A ] [ B ] cleanup
</pre>
The <code>recover ( try recover -- )</code> word encapsulates the following idiom:
<pre>
[ A ] [ [ B ] when* ] catch
[ A ] [ B ] recover
</pre>
The <code>catch ( try -- error/f )</code> word no longer takes a quotation that receives the error caught; rather, it just pushes the error that was caught on the stack. That is, the old way:
<pre>
[ A ] [ B ] catch
</pre>
Becomes:
<pre>
[ A ] catch B
</pre>
However, most uses of <code>catch</code> can be replaced by <code>cleanup</code> and <code>recover</code>.</li>
<li>A new <code>with-server ( port ident quot -- )</code> combinator takes care of listening on a network socket, logging client connections, spawning client handler threads, and error handling. The quotation is called for each client, in a new scope with the client socket as the default stream.</li>
<li>New <code>1+</code>, <code>1-</code> words are a bit faster than <code>1 +</code> and <code>1 -</code> since there is one less generic dispatch involved.</li>
<li>Side-effect-free words that output immutable values are evaluated at compile time if all their inputs are literal. You can declare a word as having this condition by suffixing the definition with <code>foldable</code>, eg:
<pre>: cube dup dup * * ; foldable</pre></li>
<li>Various arithmetic identities such as <code>1 *</code> are optimized out.
<li><code>sort ( seq quot -- | quot: elt elt -- -1/0/1 )</code> combinator now works with any sequence, not just a list. The comparator also has to return a signed integer, not just a boolean. It is much faster than the old sorting algorithm.</li>
<li><code>binsearch ( elt seq quot -- i | quot: elt elt -- -1/0/1 )</code> and <code>binsearch ( elt seq quot -- elt | quot: elt elt -- -1/0/1 )</code> combinators perform a binary search on a sorted sequence.</li>
<li><code>join ( seq glue -- seq )</code> word. Takes a sequence of sequences, and constructs a new sequence with the glue in between each sequence. For example:
<li>Integers now support the sequence protocol. An integer is an increasing sequence of its predecessors. This means the <code>count ( n -- [ 0 ... n-1 ] )</code> word is gone; just use <code>>vector</code> instead. Also, <code>project</code> has been made redundant by <code>map</code>.</li>
<li>The matrices library has been greatly simplified. Matrices are now represented as vectors of vectors, and matrix words have been moved to the <code>math</code> vocabulary.</li>
<li>More descriptive "out of bounds" errors.</li>
<li>New <code>make-hash ( quot -- namespace )</code> combinator executes quotation in a new namespace, which is then pushed on the stack.</li>
<li>The <code><namespace></code> word is gone. It would create a hashtable with a default capacity. Now, just write <code>{{ }} clone</code>.</li>
<li>The <code>every?</code> word has been replaced with <code>monotonic? ( seq quot -- ? )</code>. Its behavior is a superset of <code>every?</code> -- it now accepts any transitive relation, and checks if the sequence is monotonic under this relation. For example,
<code>[ = ] monotonic?</code> checks if all elements in a sequence are equal, and <code>[ < ] monotonic?</code> checks for a strictly increasing sequence of integers.</li>
<li>The prettyprinter has been merged with the unparser. The <code>unparse ( object -- string )</code> word has been moved to the <code>prettyprint</code> vocabulary, and can now produce a parsable string for any class supported by the prettyprinter.</li>
<li>New <code>unparse-short ( object -- string )</code> returns a string no longer than a single line.</li>
<li>The prettyprinter now supports many more configuration variables. See the handbook for details.</li>
<li>New <code>profile ( word -- )</code> word. Causes the word's accumulative runtime to be stored in a global variable named by the word. This is done with the annotation facility, the word's definition is modified; use <code>reload ( word -- )</code> to get the old definition back from the source file.</li>
<li>Binary search is now used for spacial indexing where possible. This improves performance when there are a lot of lines of output in the listener.</li>
<li>Scroll bars now behave in a more intuitive manner, closer to conventional GUIs.</li>
<li>Menus now appear when the mouse button is pressed, not released, and dragging through the menu with the button held down behaves as one would expect.</li>
<li>The data stack and call stack are now shown. In the single-stepper, these two display the state of the program being stepped. In the inspector, the call stack display is replaced with an inspector history.</li>
<li>Pack layouts with gaps are now supported.</li>
<li>New <code>cond ( conditions -- )</code> combinator. It behaves like a set of nested <code>ifte</code>s, and compiles if each branch has the same stack effect. See its documentation comment for details.</li>
