Project meeting at Stanford

Date: 02/25/99

Attendees: Svetlozar Nestorov, Sebastien Brion and Yue Zhuge

Three of us studied the NoDoSE, Northwestern Document Structure Extractor
(available from http://shrike.cs.nwu.edu/nodose ), developed by Stanford
alumni Brad Adelberg and his students. The objective of this system is to
extract structure from semi-structured test using an interactive tool.
Svetlozar showed us the system trial version 1.0, using both the example
files provided and some of our own files. We found out that (for this
version) only very simple, regular files can be correctly parsed, for
example, files with records separated by blank lines. However, the
interactive idea is an interesting one to deal with semi-structured data
with inconsistencies.

(P.S., I talked with Brad later and he promised to send us a later and
better version of DoNoSE.)

Before the meeting, Sebastien and Yue tried to display XML documents in IE
5.0 Beta, here are a few things we found:

* IE5.0 Beta can display well formed XML documents. If a document is not
well formed, the browser will return an error and try to point out where
the
error occurred.

* IE5.0 Beta can display XML documents with DTDs, and check whether a
document confines its DTD.

* Without a style sheet, a document is displayed "as it is", that is,
everything including declarations, tags and angle brackets are all
displayed. Text is displayed in bold.

* Simple style sheet examples work fine, for example, one can change the
fonts or font color using a style sheet.

Yue also tried the XML editor from Microsoft: XML Notepad. This is a very
simple editor, one can use it to write well formed XML documents by
providing tags and element values. It returns an error if a document it
tries to open is not well formed, so it can not be used to convert (not
well formed) HTML documents to XML.