|
The following page contains a series of screen shots taken
from some of the sub-projects making up the Abora project. The
author recognises the value of such pages to get an overview of
a project from a very practical point of view before expending
the significant amount of effort needed to download, install,
and prod and poke the actual software. If you have any ideas of
missing views then please contact the author.
Dolphin Demo
The Dolphin Demo was developed during the Summer of 2002
to investigate the Udanax-Gold design and explorations into
xanalogical UIs. The result is very rough, but starts to hint at
some of the interesting features that come out of the basic
design.
|
The demo is centred around a simple text editor that
supports a number of xanalogical features. The central pane of
the UI is the text editor, to the left the library of known
works and the revisions of the currently selected work. Text can
be modified just as in a normal editor. Some of the standard
Windows mechanisms such as a copy-and-paste have been extended
to support transclusions, it is also possible to create new
works based on the current text again using a transclusion
mechanism. |
|
It is possible to compare the contents of the central text
pane with any other edition. The base edition that the central
text is compared against is shown in the pane to the right of
the window. The mechanism of this comparison is interesting as
it enables accurate fine grained differences to be recognised at
the users level. This is documented in the Tech
Report, but comes down to finding the similarities rather
than the differences between texts. The rendering shown here is
based on highlighting contiguous sequences of any texts that
exist in both editions. Each sequence is assigned a colour and
the background of the matching text in both editions is then
painted with this colour. The splitter between the texts
connects together matching sequences by drawing filled regions
between them. The value of the rendering is that it helps to
visualise when blocks of text have been copied, moved or even
duplicated. Any text that is only found in one of the editions
is rendered with a grey background. A further embellishment is
that as the text cursor moves around the main text, the matching
point in the compared text is brought into focus. |
|
All of the links onto the text comprising the edition can be
viewed from this pane. Links are fine grained down to the level
of characters and may also have a type associated with them. The
server implementation of links is interesting as it is built out
of the transclusion mechanisms. One of the useful benefits of
this is that links can survive revisions of documents. A simple
typing scheme is used, so that the text tagged as being of type
Title can be used for the name of the work as shown in the
top-left pane. Another aspect of the link is whether it is
defined as part of this work, or defined elsewhere and is just
onto some of this edition's content. |
|
It is possible to back-follow any of the content that this
edition transcludes to all the other works that include the same
content. This can be useful to find if anyone is quoting your
own work, or to find other quotes or comments onto any content
that you are interested in or quoting in you document. The
extent of the content that is considered can be trimmed down to
just the current selection for finer grain research. |
|
Developer tool to investigate the actual ent-node structure
that is being used under the covers to represent the current
work. The top pane of the inspector shows each of the roots of
the ent tree, with all the children branching off towards the
right. The leaf nodes represent one or more characters of the
content. Details of each node, such as the revision it was added
to, and any displacement values are shown in line. The nodes of
the ent tree can actually be shared by multiple edition
revisions, which results in some nodes having multiple parents.
This can be investigated in the lower pane which navigates
root-ward from the node selected in the top pane. Each of the
leafs of this lower pane should correspond to a root in the top
pane. |
|
Developer tool to list some summary statistics on the make
up of the nodes making each revision of the ent tree. |
Udanax-Gold2Java
Udanax-Gold2Java is a small throw-away command line
application to auto-translate the Udanax-Gold Smalltalk source
into approximate Java source.
|
The translation process is driven by an ant script. You feed
in the original Smalltalk source files from Udanax-Gold,
supplied with the release, and something approximating Java
source files will be spit out, one file per class. There are
about 500 or so classes in the release. These classes are
generated under a src-gen directory to also allow some
manually pre-created classes to patch up some of the wholes left
by the unreleased Smalltalk source code. The whole process takes
less than a minute on a 800mhz P3 machine. |
|
