Transcript Validation - Berkeley Database Research

XML, Java, and the future of the
Web
Jon Bosak, Sun Microsystems
Last revised 1997.03.10
Presented and Annotated by
Byung-Hoon Kang
([email protected])
RoadMap :

Part1: Why XML ?




Part2: Example applications






HTML and SGML
SGML and XML
SGML vs. XML vs. HTML
Example applications envisioned in 1997
Database interchange problem
Why solution is not feasible in HTML
XML solution envisioned in 1997
Part3: Why CSS/XSL ?

What are style sheet and why?

Style sheet languages

Why two style sheets language for XML ?
Conclusion: XML,DTD,CSS and XSL
HTML and SGML

SGML allows documents to describe their own grammar

HTML hardwires a small set of tags: a single SGML
specification
HTML



Extensibility: new tag ?
application change?
Yes
Needed
SGML
Yes
Independent

Structure: deep nested ?
No
Yes

Validation: structural validity ?
No
Yes
SGML and XML


SGML :

Contains many optional features not needed for Web applications

Cost/benefit ratio unattractive to current vendors of Web
browsers

DTD is required all the time
XML :

A simplified subset of SGML for Web applications

Retain SGML’s extensibility, structure, and validation

DTD is optionally needed for validation
SGML vs. XML vs. HTML
HTML

Yes/No
SGML
XML
Yes
Yes

Extensibility: new tag ?

Structure: deep nested ?
No
Yes
Yes

Validation: structural validity ?
No
Yes
Yes

DTD is required for parsing ?
No
Yes
Opt.

Browser dependency ?
High
Low
Low

Cost/Benefit ?
Good
Poor
Med

When were you born ?
‘92
‘86
‘98
RoadMap :



Part2: Example applications

Example applications envisioned in 1997

Database interchange problem

Why solution is not feasible in HTML

XML solution envisioned in 1997
Part3: Why CSS/XSL ?

What are style sheet and why?

Style sheet languages

Why two style sheets language for XML ?
Conclusion: XML,DTD,CSS and XSL
XML applications : envisioned in 1997

to mediate between two or more heterogeneous databases.

to distribute a significant proportion of the processing load
from the Web server to the Web client.

to present different views of the same data to different
users.

to tailor information discovery to the needs of individual
users.
Database interchange: the universal hub

The information tracking system for a home health care
agency

Need to key-in all of patient medical histories and billing
data from a variety of doctors, hospitals, pharmacies,
and insurance companies.

1. Log into the hospital's Web site.

2. Become an authorized user.

3. Access the patient's medical records using a Web
browser.

4. Print out the records from the browser.

5. Manually key in the data from the printouts.
Database interchange: the universal hub

A real solution would look more like this:

1. Log into the hospital's Web site.

2. Become an authorized user.

3. Access the patient's medical records in a Web-based
interface that represents the records for that patient with
a folder icon.

4. Drag the folder from the Web application over to the
internal database application.

5. Drop it into the database.
Why solution is not feasible in HTML ?

HTML tag set is too limited

to represent or differentiate between the multitude of database
fields

hard to automate the process

HTML is incapable of representing the variety of structures
in those documents.

HTML lacks mechanism in checking the data for structural
validity
XML solution envisioned in 1997

1. Form a standards consortium for Health Care Markup
Language and its HCML.DTD

2. Adopt a single industry-wide interchange format that
serves as the single output format.

3. Use any standard application to validate and process
HCML marked-up data with HCML.DTD

Similar to the steps in traditional SGML usage

Still have the same problem of standardization.

Is XML advantageous than SGML in doing steps 1, 2 or 3
?
RoadMap :


Part3: Why CSS/XSL ?

What are style sheet and why?

Style sheet languages

Why two style sheets language for XML ?
Conclusion: XML,DTD,CSS and XSL
What are style sheets and Why?


Style sheets

Describe how documents are presented on screens, in
print or in audio.

Detach the formatting information from data source in
document
Why: want to influence the presentation of documents

without sacrificing device-independence or

adding new HTML tags.
Stylesheets

DSSSL (Document Style Semantics and Specification
Language) for SGML

DSSSL-online : the subset of DSSSL for SGML online.

CSS (cascading style sheets) :


provides a style mechanism well suited to the relatively low-level
demands of HTML

currently used a little bit
XSL (extensible style language):

needed to present XML document on the browser.

transformational functionality
Why two Style Sheet languages?
CSS XSL


Can be used with HTML? yes
no

Can be used with XML ? yes
yes

Transformation language? no
yes

Syntax ?
XML

CSS need to be used to style HTML documents

XSL, on the other hand, is able to transform documents


CSS
For example, XSL can be used to transform XML data
into HTML/CSS documents
Both languages can be used to style XML documents
Conclusion: XML, DTD, CSS and
XSL

To validate XML structural input : DTD

To exchange data using XML: None or DTD for structural
validation

To standardize the data exchange model: Still need to form
a consortium

To render XML in XML-support-browser : CSS or XSL

To render XML in HTML-support-browser : Use XSL to
transform XML into HTML/CSS