[Gzz-commits] manuscripts/xupdf SCRATCH article.rst

[Janne V. Kujala]

CVSROOT:        /cvsroot/gzz
Module name:    manuscripts
Changes by:     Janne V. Kujala <address@hidden>        03/02/12 09:14:25

Modified files:
        xupdf          : SCRATCH article.rst 

Log message:
        start structuring intro

CVSWeb URLs:
http://savannah.gnu.org/cgi-bin/viewcvs/gzz/manuscripts/xupdf/SCRATCH.diff?tr1=1.1&tr2=1.2&r1=text&r2=text
http://savannah.gnu.org/cgi-bin/viewcvs/gzz/manuscripts/xupdf/article.rst.diff?tr1=1.62&tr2=1.63&r1=text&r2=text

Patches:
Index: manuscripts/xupdf/SCRATCH
diff -u manuscripts/xupdf/SCRATCH:1.1 manuscripts/xupdf/SCRATCH:1.2
--- manuscripts/xupdf/SCRATCH:1.1       Mon Feb 10 03:23:24 2003
+++ manuscripts/xupdf/SCRATCH   Wed Feb 12 09:14:25 2003
@@ -1,3 +1,105 @@
+- `Link titles`_ (in the HTML standard at least since 2.0)
+  give additional information by the linking page's author
+  about a link's destination
+
+.. _Link titles: http://www.useit.com/alertbox/980111.html
+
+- Fluid links [zellweger98fluid]_ can be put in by the original author,
+  or by the user [XXX ht'01], 
+
+  - Fluid annotations can be rendered through a DOM: That is,
+    it suffices to modify the page's DOM and re-render, you don't
+    have to turn the browser's rendering engine inside out to support
+    lines cross-cutting the view hierarchy. This is why buoys would be
+    harder to adopt into an existing program. The fluid stuff supports
+    animation by modifying the DOM also (it simply grows a box).
+    [XXX Hypermedia in the Small: Fluid annotations through 
+    open hypermedia: using and extending emerging web standards, WWW2002]
+
+  - (IMHO) user interface problems: confuses using the mouse as eye or hand:
+    normal mouse use: mouse = hand. quake mouse: mouse = eye.
+    Here, neither paradigm applies fully as you would need to move both
+    the mouse AND your eyes.
+
+  - Also, the following the link that received the fluid annotation
+    still has the same disruptive change.
+
+
+- Focus+context views of the web as proposed by 
+  Mukherjea and Hara [mukharjea97fcwww]_ give a 1'000ft
+  view of the linking strucure (need to read & find out whether
+  they show thumbnails, or what they show).
+  Emphasis seems to be on the methods for generating the f+c structure,
+  i.e., finding landmarks, paths between nodes, what nodes to show, etc.
+  Some features used in the example views:
+
+        - nodes are shown as cubes (focus node as a sphere) with titles
+        - node size/height/color depicts importance/access frequency
+        - landmark nodes distinguished by color
+        - left/right direction for in/out links
+
+- In XLibris [XXX ht'02: going back in hypertext], recently visited
+  pages are shown as thumbnails at the bottom of the window
+- Free form 'digital ink' annotation [e.g. XLibris, iMarkup?] is currently
+  the closest thing to cross-cutting connections (because that's what
+  users do with them, at least sometimes)
+
+  - Point to make: This form of annotation (buoy-like) is nothing new...
+    Maybe start with 'old fashioned ink annotation' and then introduce
+    forms of representing linking in computers?
+
+        - also used a lot for images, see the NASA example image
+          in irregu
+
+  - XLibris [schilit98digital]_ also automatically searches for pages
+    matching stuff the user entered/circled on a page, and place
+    a thumbnail in the margin -> linking with visible targets
+    (no automatic placing as with buoys, though)
+    -- actually, not sure about stuff the user entered.
+    Certain about stuff the user circled/underlined/highlighted, though.
+    "Each annotation generates a query."
+
+  - Goal: "[A] modeless link *suggestion* mechanism" (emphasis mine).
+
+  - Bad assumption in XLibris: "Although hypertext was designed
+    to address some of these problems [quickly jumping to related docs],
+    authors cannot anticipate the information needs of all readers." 
[schilit98d
+igital]_
+    Therefore, *the readers must make the links*! (Or be able to.)
+    (Should be able to find this point somewhere.)
+
+- Hypercept: [milgram99hypercept]_
+
+- Nelson's as of yet unimplemented
+  transpointing windows [1979] are not about showing a set of links,
+  but also represent connections crosscutting the view hierarchy.
+  
+  Known prototype implementations to date:
+
+  - Southhampton transpointing windows demo
+  - Nile
+  - Cosmicbook?  (Has anybody tried this?)
+    - seems to be vapourware
+  - To some degree, Pyxi (which didn't solve the problem with the
+    existing view systems, but tricked its way around it)
+
+
+- pad++ browser: shows linked docs next to it
+  [fc-zooming]_
+  http://www.cs.umd.edu/hcil/pad++/papers/bookchap-98-webbrowser/index.html
+
+  - builds a zoomable tree of the visited web pages
+  - fluidly animates all view changes
+  - page size proportional to the distance from focus
+  - the layout depends on the order the user visits the pages
+  - ancestor / descendant pages shown on left/right sides of a page
+  - long pages get scrollbars
+
+
+
+
+
+
 ..  UML:: zzStructure
 
     class zzSpace
Index: manuscripts/xupdf/article.rst
diff -u manuscripts/xupdf/article.rst:1.62 manuscripts/xupdf/article.rst:1.63
--- manuscripts/xupdf/article.rst:1.62  Wed Feb 12 06:48:26 2003
+++ manuscripts/xupdf/article.rst       Wed Feb 12 09:14:25 2003
@@ -90,69 +90,49 @@
 Introduction
 ============
 
-Modern graphics accelerators enable, even on commodity hardware,
-several visual effects that 5 years ago were only possible on expensive
-graphics workstations. 
-It is time to re-examine the...
-
 In many hypertext systems today, following a link means a disruptive change
 in the user's workflow, replacing the current context with an
 entirely different one.
+This is possibly because most systems still use the paradigm developed in 70s 
+at Xerox PARC [XXXref] with
+overlapping, rectangular, *independent* viewports to paper-like documents ---
+A link operation can only create a new window or switch the contents of an
+existing window.
+Alternatives, e.g., replacing part of text (XXX antont's reference)
+are not very popular and ...
+
+There have been various attempts to improve user orientation.
+One approach is to embed information about the destination
+or the link relationship in the links.
+Link titles are now a standard HTML feature (at least since 2.0) and
+fluid links [zellweger98fluid]_ take the concept further by
+allowing a user to see gradually more and more target context 
+before following a link. 
+Hypercept [milgram99hypercept]_ provides a cue of local structure 
+by animating the transition from current document to a linked 
+document in different ways depending on the structural relationship
+exemplified by the link.
+
+.. XXX: behaviour, structural role
+
+Another approach is to show the local structure directly in the view layout.
+Focus+context views of the web as proposed by Mukherjea and Hara 
+[mukharjea97fcwww]_ provide overview diagrams of the linking strucure
+of web pages with important nodes emphasized.
+The Pad++ browser [fc-zooming]_ records history of the visited pages
+as a tree with nodes showing the complete pages.
+The focused page is shown at a larger scale and the user is free to
+pan and zoom the large virtual view with all layout changes
+fluidly animated.
 
-- possibly because most systems still use in the paradigm developed in 70s at 
Xerox PARC:
-
-    - overlapping, rectangular, *independent* viewports to paper-like documents
-
-       - link operation can only create new window or switch contents of window
-
-       - alternatives, e.g. replacing part of text (XXX antont's reference)
-         not very popular
-
-There have been various attempts to alleviate these problems:
-
-
-- `Link titles`_ (in the HTML standard at least since 2.0)
-  give additional information by the linking page's author
-  about a link's destination
-
-.. _Link titles: http://www.useit.com/alertbox/980111.html
+..  - long pages get scrollbars
 
-- Fluid links [zellweger98fluid]_ can be put in by the original author,
-  or by the user [XXX ht'01], 
+Buoy-like annotation, less PARC-like:
 
-  - Fluid annotations can be rendered through a DOM: That is,
-    it suffices to modify the page's DOM and re-render, you don't
-    have to turn the browser's rendering engine inside out to support
-    lines cross-cutting the view hierarchy. This is why buoys would be
-    harder to adopt into an existing program. The fluid stuff supports
-    animation by modifying the DOM also (it simply grows a box).
-    [XXX Hypermedia in the Small: Fluid annotations through 
-    open hypermedia: using and extending emerging web standards, WWW2002]
-
-  - (IMHO) user interface problems: confuses using the mouse as eye or hand:
-    normal mouse use: mouse = hand. quake mouse: mouse = eye.
-    Here, neither paradigm applies fully as you would need to move both
-    the mouse AND your eyes.
-
-  - Also, the following the link that received the fluid annotation
-    still has the same disruptive change.
-
-
-- Focus+context views of the web as proposed by 
-  Mukherjea and Hara [mukharjea97fcwww]_ give a 1'000ft
-  view of the linking strucure (need to read & find out whether
-  they show thumbnails, or what they show).
-  Emphasis seems to be on the methods for generating the f+c structure,
-  i.e., finding landmarks, paths between nodes, what nodes to show, etc.
-  Some features used in the example views:
-
-       - nodes are shown as cubes (focus node as a sphere) with titles
-       - node size/height/color depicts importance/access frequency
-       - landmark nodes distinguished by color
-       - left/right direction for in/out links
+In XLibris [XXX ht'02: going back in hypertext][schilit98digital]_, 
+recently visited pages are shown as thumbnails at the bottom of the window.
+Furthermore,
 
-- In XLibris [XXX ht'02: going back in hypertext], recently visited
-  pages are shown as thumbnails at the bottom of the window
 - Free form 'digital ink' annotation [e.g. XLibris, iMarkup?] is currently
   the closest thing to cross-cutting connections (because that's what
   users do with them, at least sometimes)
@@ -180,7 +160,6 @@
     Therefore, *the readers must make the links*! (Or be able to.)
     (Should be able to find this point somewhere.)
 
-- Hypercept: [milgram99hypercept]_
 
 - Nelson's as of yet unimplemented
   transpointing windows [1979] are not about showing a set of links,
@@ -196,17 +175,6 @@
     existing view systems, but tricked its way around it)
 
 
-- pad++ browser: shows linked docs next to it
-  [fc-zooming]_
-  http://www.cs.umd.edu/hcil/pad++/papers/bookchap-98-webbrowser/index.html
-
-  - builds a zoomable tree of the visited web pages
-  - fluidly animates all view changes
-  - page size proportional to the distance from focus
-  - the layout depends on the order the user visits the pages
-  - ancestor / descendant pages shown on left/right sides of a page
-  - long pages get scrollbars
-
 In this article, we take the ideas seen in the above references
 a step further, by
 
@@ -218,6 +186,11 @@
 we present some user interface technologies 
 developed for 
 
+Modern graphics accelerators enable, even on commodity hardware,
+several visual effects that 5 years ago were only possible on expensive
+graphics workstations. 
+It is time to re-examine the...
+
 
 xupdf:
   - less rigid structure of the view:
@@ -517,6 +490,7 @@
 ..  UML:: xupdfStructure
     :caption: An UML diagram of the structure XXX. XXX orthogonal...
     :width: 8cm
+    :alt: figuml
 
     class xuFluidMediaUnit
        fields