diff --git a/extra/io/files/tmp/tmp-tests.factor b/extra/io/files/tmp/tmp-tests.factor
new file mode 100644
index 0000000000..ba2ff7046c
--- /dev/null
+++ b/extra/io/files/tmp/tmp-tests.factor
@@ -0,0 +1,5 @@
+USING: io.files io.files.tmp kernel strings tools.test ;
+IN: temporary
+
+[ t ] [ tmpdir string? ] unit-test
+[ t f ] [ ".tmp" [ dup exists? swap ] with-tmpfile exists? ] unit-test
diff --git a/extra/io/files/tmp/tmp.factor b/extra/io/files/tmp/tmp.factor
new file mode 100644
index 0000000000..da1deec9a7
--- /dev/null
+++ b/extra/io/files/tmp/tmp.factor
@@ -0,0 +1,22 @@
+USING: continuations io io.files kernel sequences strings.lib ;
+IN: io.files.tmp
+
+: tmpdir ( -- dirname )
+    #! ensure that a tmp dir exists and return its name
+    #! I'm using a sub-directory of factor for crossplatconformity (windows doesn't have /tmp)
+    "tmp" resource-path dup directory? [ dup make-directory ] unless ;
+
+: touch ( filename -- )
+    <file-writer> stream-close ;
+
+: tmpfile ( extension -- filename )
+    16 random-alphanumeric-string over append
+    tmpdir swap path+ dup exists? [
+        drop tmpfile
+    ] [
+        nip dup touch
+    ] if ;
+
+: with-tmpfile ( extension quot -- )
+    #! quot should have stack effect ( filename -- )
+    swap tmpfile tuck swap curry swap [ delete-file ] curry [ ] cleanup ;
diff --git a/extra/semantic-db/context/context.factor b/extra/semantic-db/context/context.factor
new file mode 100644
index 0000000000..f4d5834665
--- /dev/null
+++ b/extra/semantic-db/context/context.factor
@@ -0,0 +1,14 @@
+! Copyright (C) 2008 Alex Chapman
+! See http://factorcode.org/license.txt for BSD license.
+USING: semantic-db.db ;
+IN: semantic-db.context
+
+: all-contexts ( -- contexts )
+    has-type-relation context-type relation-object-subjects ;
+
+: context-relations ( context -- relations )
+    has-context-relation swap relation-object-subjects ;
+    
+: get-context ( name -- context )
+    context-type swap ensure-node-of-type ;
+
diff --git a/extra/semantic-db/db/db-tests.factor b/extra/semantic-db/db/db-tests.factor
new file mode 100644
index 0000000000..303ec658a0
--- /dev/null
+++ b/extra/semantic-db/db/db-tests.factor
@@ -0,0 +1,26 @@
+USING: io.files kernel namespaces semantic-db.db semantic-db.db.private sqlite tools.test ;
+IN: temporary
+
+[ "n.id" ] [ "id" "n" [ 0 column-text ] <field> field-sql ] unit-test
+[ "select n.id from nodes n where n.content = :content" ] [
+    <query>
+    "id" "n" [ 0 column-text ] <field> over add-field 
+    "nodes n" over add-table
+    "n.content = :content" over add-condition
+    query-sql
+] unit-test
+
+[
+    create-node-table create-arc-table 
+    [ 1 ] [ "first node" create-node ] unit-test
+    [ 2 ] [ "second node" create-node ] unit-test
+    [ 3 ] [ "third node" create-node ] unit-test
+    [ 4 ] [ f create-node ] unit-test
+    [ "first node" ] [ 1 node-content ] unit-test
+    [ 5 ] [ 1 2 3 create-arc ] unit-test
+    [ { { 1 2 3 } } ] [ 2 node-arcs ] unit-test
+    [ { { 1 2 3 } } ] [ 3 node-arcs ] unit-test
+    [ { { 3 1 } } ] [ 2 node-subject-arcs ] unit-test
+    [ { { 2 1 } } ] [ 3 node-object-arcs ] unit-test
+]
+with-tmp-db
diff --git a/extra/semantic-db/db/db.factor b/extra/semantic-db/db/db.factor
new file mode 100644
index 0000000000..5616f07a1c
--- /dev/null
+++ b/extra/semantic-db/db/db.factor
@@ -0,0 +1,287 @@
+! Copyright (C) 2007, 2008 Alex Chapman
+! See http://factorcode.org/license.txt for BSD license.
+USING: arrays assocs kernel math namespaces sequences sqlite ;
+IN: semantic-db.db
+
+! sqlite utils
+: prepare ( string -- statement )
+    db get swap sqlite-prepare ;
+
+: binding ( statement key val -- statement )
+    >r dup integer? [ 1+ ] when dupd r> sqlite-bind-by-name-or-index ;
+
+GENERIC# bindings 1 ( bindings statement -- statement )
+
+M: assoc bindings
+    swap [ binding ] assoc-each ;
+
+M: sequence bindings
+    swap dup length swap [ binding ] 2each ;
+
+: prepare-with-bindings ( bindings string -- statement )
+    prepare bindings ;
+
+: select-with-bindings ( bindings string quot -- results )
+    >r prepare-with-bindings dup r> sqlite-map swap sqlite-finalize ;
+
+: ignore-and-finalize ( statement -- )
+    dup [ drop ] sqlite-each sqlite-finalize ;
+
+: sql-update ( string -- )
+    prepare ignore-and-finalize ;
+
+: update-with-bindings ( bindings string -- )
+    prepare-with-bindings ignore-and-finalize ;
+
+: 1result ( array -- result )
+    #! return the first (and hopefully only) element of the array, or f
+    dup length 0 > [ first ] [ drop f ] if ;
+
+: (collect-int-columns) ( statement n -- )
+    [ dupd column-int , ] each drop ;
+
+: collect-int-columns ( statement n -- columns )
+    [ (collect-int-columns) ] { } make ;
+
+! queries
+TUPLE: field name table retriever ;
+C: <field> field
+
+TUPLE: query fields tables conditions args statement results ;
+
+: call-field-retrievers ( query 
+
+: <query> ( -- query )
+    V{ } clone V{ } clone V{ } clone H{ } clone f f
+    query construct-boa ;
+
+: invalidate-query ( query -- query )
+    f over set-query-results ;
+
+: add-field ( field query -- ) invalidate-query query-fields push ;
+: ,field ( name table retriever -- ) <field> query get add-field ;
+
+: add-table ( table query -- ) invalidate-query query-tables push ;
+: ,table ( table -- ) query get add-table ;
+
+: add-condition ( condition query -- ) invalidate-query query-conditions push ;
+: ,condition ( condition -- ) query get add-condition ;
+
+: add-arg ( arg key query -- ) invalidate-query query-args set-at ;
+: ,arg ( arg key -- ) query get add-arg ;
+
+<PRIVATE
+
+: field-sql ( field -- sql )
+    [ dup field-table % CHAR: . , field-name % ] "" make ;
+
+: fields-sql ( query -- sql )
+    query-fields dup length [
+        [ field-sql ] map ", " join
+    ] [
+        drop "*"
+    ] if ;
+
+: tables-sql ( query -- sql )
+    query-tables ", " join ;
+
+: conditions-sql ( query -- sql )
+    query-conditions dup length [
+        " and " join "where " swap append
+    ] [
+        drop ""
+    ] if ;
+
+: query-sql ( query -- sql )
+    [
+        "select" , dup fields-sql , dup "from" , tables-sql , conditions-sql ,
+    ] { } make " " join ;
+
+: prepare-query ( query -- )
+    [ query-sql prepare ] keep set-query-statement ;
+
+: bind-query ( query -- )
+    dup query-args over query-statement bindings swap set-query-statement ;
+
+: (retrieve) ( statement query -- result )
+    query-fields swap [ field-retriever call ] curry each ;
+
+: retrieve ( query -- )
+    dup query-statement over [ (retrieve) ] curry sqlite-map
+    swap set-query-results ;
+   ! dup query-statement over query-retriever sqlite-map swap set-query-results ;
+
+: finalize-query ( query -- )
+    query-statement dup sqlite-finalize f swap set-query-statement ;
+
+PRIVATE>
+    
+: run-query ( query -- )
+    dup prepare-query dup bind-query dup retrieve finalize-query ;
+
+: get-results ( query -- results )
+    dup query-results [ nip ] [ dup run-query query-results ] if* ;
+
+: with-query ( quot -- results )
+    [
+        <query> query set
+        call
+        query get get-results
+    ] with-scope ;
+
+! nodes and arcs
+
+! maybe merge nodes and arcs table, so arcs can be nodes too:
+! create table nodes (id integer primary key autoincrement, value none, type integer, subject integer, object integer)
+! nodes:
+!   value: node content
+!   type: nid of node type
+!   subject: null
+!   object: null
+!
+! arcs:
+!   value: ordinality, or null
+!   type: nid of relation
+!   subject: nid of arc subject
+!   object: nid of arc object
+!
+! An alternative layout:
+!
+! nodes:
+!   id
+!   type
+!
+! content:
+!   id
+!   content
+!
+! arcs:
+!   id
+!   relation
+!   subject
+!   object
+!   ordinal
+!
+! A third alternative. In this, all arcs have an entry in the nodes table, but
+! their content is null. No node that isn't an arc can have null content. If an
+! arc needs an ordinal, then it can be created as another arc.
+!
+! nodes:
+!   id
+!   content
+!
+! arcs:
+!   id
+!   relation
+!   subject
+!   object
+
+: create-node-table ( -- )
+    "create table nodes (id integer primary key autoincrement, content none);" sql-update ;
+
+: create-arc-table ( -- )
+    "create table arcs (id integer, relation integer, subject integer, object integer);" sql-update ;
+
+: create-node ( content -- id )
+    #! if content is f then it is inserted as NULL
+    [ 1array ] [ drop { } clone ] if*
+    "insert into nodes (content) values (?);"
+    update-with-bindings db get sqlite-last-insert-rowid ;
+
+: create-bootstrap-nodes ( -- )
+    { "context" "relation" "is of type" "semantic-db" "is in context" }
+    [ create-node drop ] each ;
+
+: context-type 1 ; inline
+: relation-type 2 ; inline
+: has-type-relation 3 ; inline
+: semantic-db-context 4 ; inline
+: has-context-relation 5 ; inline
+
+: create-arc ( relation subject object -- id )
+    f create-node -roll 4array
+    "insert into arcs (id, relation, subject, object) values (?, ?, ?, ?);"
+    update-with-bindings ;
+
+: create-bootstrap-arcs ( -- )
+    has-type-relation has-type-relation relation-type create-arc drop
+    has-type-relation semantic-db-context context-type create-arc drop
+    has-context-relation has-type-relation semantic-db-context create-arc drop
+    has-type-relation has-context-relation relation-type create-arc drop
+    has-context-relation has-context-relation semantic-db-context create-arc drop ;
+
+: init-semantic-db ( -- )
+    create-node-table create-arc-table create-bootstrap-nodes create-bootstrap-arcs ;
+
+: node-content ( id -- content )
+    1array "select content from nodes where id = ?" [ 0 column-text ] select-with-bindings 1result ;
+
+: node-arcs ( node-id -- arcs )
+    1array "select id, relation, subject, object from arcs where subject = ?1 or object = ?1;"
+    [ 4 collect-int-columns ] select-with-bindings ;
+
+: node-subject-arcs ( node-id -- arcs )
+    1array "select object, relation from arcs where subject = ?;"
+    [ 2 collect-int-columns ] select-with-bindings ;
+
+: node-object-arcs ( node-id -- arcs )
+    1array "select subject, relation from arcs where object = ?;"
+    [ 2 collect-int-columns ] select-with-bindings ;
+
+: relation-subject-objects ( relation subject -- objects )
+    2array "select object from arcs where relation = ? and subject = ?;"
+    [ 0 column-int ] select-with-bindings ;
+
+: relation-object-subjects ( relation object -- subjects )
+    2array "select subject from arcs where relation = ? and object = ?;"
+    [ 0 column-int ] select-with-bindings ;
+
+: subject-object-relations ( subject object -- relations )
+    2array "select relation from arcs where subject = ? and object = ?"
+    [ 0 column-int ] select-with-bindings ;
+
+: type-and-name-node ( type name -- node )
+    has-type-relation 3array
+    "select n.id from arcs a, nodes n where a.subject = n.id and a.object = ? and n.name = ? and a.relation = ?"
+    [ 0 column-int ] select-with-bindings 1result ;
+
+: create-node-of-type ( type name -- node )
+    create-node [ has-type-relation -rot create-arc drop ] keep ;
+
+: ensure-node-of-type ( type name -- node )
+    2dup type-and-name-node [ 2nip ] [ create-node-of-type ] if* ;
+
+: type-and-name-in-context-node ( context type name -- node )
+    [
+        "id" "n" [ 0 column-int ] ,field
+        "nodes n" ,table
+        "n.name = :name" ,condition
+        ":name" ,arg
+        "arcs a" ,table
+        "a.relation = :has_type" ,condition
+        has-type-relation ":has_type" ,arg
+        "a.subject = n.id" ,condition
+        "a.object = :type" ,condition
+        ":type" ,arg
+        "arcs b" ,table
+        "b.subject = a.relation" ,condition
+        "b.relation = :has_context" ,condition
+        has-context-relation ":has_context" ,arg
+        "b.object = :context" ,condition
+        ":context" ,arg
+    ] with-query 1result ;
+
+! ideas for an api:
+! this would work something like jquery, where arcs can be selected according
+! to parameters, and the contents of nodes and arcs are retrieved on demand, or
+! at the program's convenience. It may be better to do this as a query language.
+! TUPLE: node id content ;
+! : node-text ( node -- text )
+!     dup node-content [
+!         nip
+!     ] [
+!         node-id ! now get content from database, save it in node-content, and return it
+!     ] if* ;
+! TUPLE: arc id relation subject object ;
+! 
+! TUPLE: arcs ids relation subject object ;
diff --git a/extra/sqlite/sqlite.factor b/extra/sqlite/sqlite.factor
index d651ad916c..63d9d64237 100644
--- a/extra/sqlite/sqlite.factor
+++ b/extra/sqlite/sqlite.factor
@@ -7,8 +7,8 @@
 ! executing SQL calls and obtaining results.
 !
 IN: sqlite
-USING: alien compiler kernel namespaces sequences strings sqlite.lib
-    alien.c-types continuations ;
+USING: alien compiler io.files.tmp kernel math namespaces sequences strings
+    sqlite.lib alien.c-types continuations ;
 
 TUPLE: sqlite-error n message ;
 SYMBOL: db
@@ -50,12 +50,34 @@ SYMBOL: db
   #! Bind the text to the parameterized value in the statement.  
   dup length SQLITE_TRANSIENT sqlite3_bind_text sqlite-check-result ;
 
+: sqlite-bind-int ( statement index int -- )
+    sqlite3_bind_int sqlite-check-result ;
+
+GENERIC: sqlite-bind ( statement index obj -- )
+
+M: object sqlite-bind ( statement index obj --  )
+    sqlite-bind-text ;
+
+M: integer sqlite-bind ( statement index int -- )
+    sqlite-bind-int ;
+
 : sqlite-bind-parameter-index ( statement name -- index )
   sqlite3_bind_parameter_index ;
 
 : sqlite-bind-text-by-name ( statement name text -- )
   >r dupd sqlite-bind-parameter-index r> sqlite-bind-text ;
 
+: sqlite-bind-by-name ( statement name obj -- )
+    >r dupd sqlite-bind-parameter-index r> sqlite-bind ;
+
+GENERIC# sqlite-bind-by-name-or-index 1 ( statement key val -- )
+
+M: object sqlite-bind-by-name-or-index ( statement object val -- )
+    sqlite-bind-by-name ;
+
+M: integer sqlite-bind-by-name-or-index ( statement integer val -- )
+    sqlite-bind ;
+
 : sqlite-finalize ( statement -- )
   #! Clean up all resources related to a statement. Once called
   #! the statement cannot be used. All statements must be finalized
@@ -77,6 +99,9 @@ SYMBOL: db
   #! from zero, as a string.
   sqlite3_column_text ;
 
+: column-int ( statement index -- int )
+  sqlite3_column_int ;
+
 : step-complete? ( step-result -- bool )
   #! Return true if the result of a sqlite3_step is
   #! such that the iteration has completed (ie. it is
@@ -125,3 +150,7 @@ DEFER: (sqlite-map)
         [ db get sqlite-close ] [ ] cleanup
     ] with-scope ;
 
+: with-tmp-db ( quot -- )
+  ".db" [
+      swap with-sqlite
+  ] with-tmpfile ;
diff --git a/extra/strings/lib/lib-tests.factor b/extra/strings/lib/lib-tests.factor
new file mode 100644
index 0000000000..2779e190c9
--- /dev/null
+++ b/extra/strings/lib/lib-tests.factor
@@ -0,0 +1,8 @@
+USING: kernel sequences strings.lib tools.test ;
+IN: temporary
+
+[ "abcdefghijklmnopqrstuvwxyz" ] [ lower-alpha-chars "" like ] unit-test
+[ "ABCDEFGHIJKLMNOPQRSTUVWXYZ" ] [ upper-alpha-chars "" like ] unit-test
+[ "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ" ] [ alpha-chars "" like ] unit-test
+[ "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789" ] [ alphanumeric-chars "" like ] unit-test
+[ t ] [ 100 [ drop random-alphanumeric-char ] map alphanumeric-chars [ member? ] curry all? ] unit-test
diff --git a/extra/strings/lib/lib.factor b/extra/strings/lib/lib.factor
index 223fdb2090..6affe067fd 100644
--- a/extra/strings/lib/lib.factor
+++ b/extra/strings/lib/lib.factor
@@ -1,4 +1,4 @@
-USING: math arrays sequences kernel splitting strings ;
+USING: math arrays sequences kernel random splitting strings ;
 IN: strings.lib
 
 : char>digit ( c -- i ) 48 - ;
@@ -12,3 +12,28 @@ IN: strings.lib
 
 : >Upper-dashes ( str -- str )
     "-" split [ >Upper ] map "-" join ;
+
+: lower-alpha-chars ( -- seq )
+    26 [ CHAR: a + ] map ;
+
+: upper-alpha-chars ( -- seq )
+    26 [ CHAR: A + ] map ;
+
+: numeric-chars ( -- seq )
+    10 [ CHAR: 0 + ] map ;
+
+: alpha-chars ( -- seq )
+    lower-alpha-chars upper-alpha-chars append ;
+
+: alphanumeric-chars ( -- seq )
+    alpha-chars numeric-chars append ;
+
+: random-alpha-char ( -- ch )
+    alpha-chars random ;
+
+: random-alphanumeric-char ( -- ch )
+    alphanumeric-chars random ;
+
+: random-alphanumeric-string ( length -- str )
+    [ drop random-alphanumeric-char ] map "" like ;
+