Processing of HTML FORMS from XML Documents

Using

XML

Metadata

to

Enable-the

Automatic

Generation

and

Processing

of

HTML

FORMS

from

XML

Documents

Anil

K.

Dubey,

M.D.

and

Henry

C.

Chueh,

M.D.,

M.S.

Laboratory

of

Computer

Science,

Massachusetts

General

Hospital

Boston,

Massachusetts

INTRODUCTION

A

common

scenario

in

web-based

database

applications

is

the

following:

obtaining

data

from

a

database,

presenting

it

to

the

user

(using

a

HTML

FORM),

allowing

the

user

to

update

the

information,

and

finally

submitting

the

updated

data

back

to

the

database.

XML

can

be

used

as

a

transport

medium

to

shuttle

data

between

the

database

and

the

supporting

web

application

logic.

A

significant

part

of

this

logic

is

often

devoted

to

creating

a

HTML

presentation

of

the

XML

data

and

reintegrating

the

HTML

FORM

submission

data

with

the

original

XML

data

(for

insertion

back

into

the

database).

While

the

XMIJXSL

transformation

paradigm

works

well

in

presenting

static

data,

its

use

in

generating

HTML

FORMS

from

XML

documents

is

less

clear.

XSL

stylesheets

used

to

present

static

(i.e.

non-FORM)

data

focus

primarily

on

presentation

and

layout.

However,

in

the

case

of

presenting

HTML

FORM

data,

one

not

only

needs

to

know

about

layout,

but

also

which

HTML

FORM

element

(i.e.

radio

buttons,

check

boxes,

etc)

to

use

for

each

data

item.

Such

details

relate

more

to

the

schema

of

the

XML

document

than

the

presentation

information

contained

in

the

XSL

stylesheet.

The

task

of

reintegrating

the

HTML

FORM

submission

data

(i.e.

HTTP

POST

or

GET

request)

into

the

original

XML

document

is

complicated

by

the

fact

that

the

structure

of

XML

documents

is

hierarchical;

the

structure

of

the

name-

value

pairs

in

HTTiP

GET

or

POST

requests

is

essentially

flat.

This

mapping

difficulty

basically

amounts

to

a

problem

with

naming;

in

order

for

the

mapping

to

be

performed

correctly,

each

name

in

the

name-value

pair

collection

of

the

HTTP

FORM

submission

must

be

unique.

In

addition,

the

naming

scheme

must

respect

and

acknowledge

the

inherent

structure

of

the

XML

document,

so

that

child

data

elements

with

a

common

parent

are

grouped

appropriately.

METHODS

AND

APPROACH

What

is

needed

to

automatically

associate

each

data

element

with

its

corresponding

HTML

FORM

element

is

metadata.

In

order

to

create

a

metadata

scheme,

we

first

had

to

develop

a

taxonomy

of

basic

data

types.

This

taxonomy

described

how

data

elements

relate

to

their

HTML

FORM

element

countexparts.

Basic

data

types

(numeric

data,

textual

data,

categorical

data)

map

to

the

following

intermediate

types:

single

line

text,

multi-line

text,

1

of

n

selection,

m

of

n

selection.

Each

of

these

intermediate

types

can

be

rendered

by

one

or

more

HTML

FORM

element

types

(i.e.

"1-of-n

selection"

being

represented

by

radio

buttons

or

a

select

box).

For

each

HTML

FORM

element

type

we

created

a

corresponding

METADATA

node.

This

METADATA

node

contains

information

needed

to

properly

render

the

data

within

the

FORM

element.

We

addressed

the

naming

problem

by

using

a

"mapping

document"

that

assigned

unique

identifiers

for

each

data

node.

We

then

made

the

document

"persistent"

so

that

it

can

be

used

to

process

the

submitted

FORM

data.

DESIGN

AND

IMPLEMENTATION

We

implemented

a

framework

capable

of

generating

HTML

FORMS

from

XML

documents.

The

framework

we

developed

is

also

capable

of

receiving

the

HTTP

FORM

submission

data

and

automatically

reintegrating

this

data

back

into

the

original

XML

document.

Our

framework

is

also

capable

of

performing

some

data

validation.

This

framework

was

used

in

implementing

a

chart

review

data

entry

application.

This

application

is

in

production,

and

is

in

active

use.

DISCUSSION

Our

framework

allows

us

to

use

XML

as

an

input/output

layer

to

the

HTML

FORM

presentation/submission

process.

In

their

effort

known

as

XFORMS,

the

W3C

is

developing

standards

for

the

next

generation

of

web

forms.

The

basic

idea

behind

the

new

standard

is

to

increase

functionality

and

interoperability

by

keeping

the

data

a

form

captures

("instance

data"),

its

functional

definition

(XFORMS

model),

and

details

regarding

its

ultimate

rendering/presentation

(XFORMS

user

interface)

separate

and

encapsulated

from

each

other.

There

are

some

similarities

between

the

design

of

our

framework

and

that

of

the

XFORMS

model.

In

our

framework

we

are

able

to

implement

some

of

the

"additional

functionality"

of

the

XFORMS

model,

using

existing

protocols.

CONCLUSION

The

use

of

a

METADATA

XML

document

can

enable

the

data-driven

generation

of

HTML

FORMS

from

XML

documents.

One

can

automate

the

incorporation

of

data

from

an

HTTP

FORM

submission

into

an

XML

document

by

mapping

each

XML

data

node

to

a

unique

identifier,

using

these

unique

identifiers

to

name

FORM

elements,

and

making

the

mapping

document

persistent.

This

approach

has

been

used

successfuilly

in

a

real-world

application,

and

can

potentially

lessen

some

of

the

overhead

involved

in

building

web-based

data

entry

applications.

1067-5027/01/$5.00

C)

2001

AMIA,

Inc.

894