Integration of External Data in Document Workﬂows via Web Services

Davide Gazz`e

, Alessio Bechini

, Marco Avvenuti

, Maurizio Tesconi

and Andrea Marchetti

IIT-CNR, Via Moruzzi, 1, I-56124, Pisa, Italy

Dep. of Information Engineering, University of Pisa, Largo L. Lazzarino, I-56122 Pisa, Italy

Keywords:

Document Workﬂows, Content Management Systems, Web Services, Ofﬁce Automation.

Abstract:

In recent years the adoption of systems to automatically manage information assets in networked, collaborative

ways has become vital for enterprises. In this work we focus on collaborative document management systems,

where actual document development relies on accessing structured data in remote repositories. Forms within

documents must be ﬁlled in with proper data items: An automated support to gather the required remote

data may substantially improve the overall document management process. The Web Service technology is

of fundamental help in this setting, because it lets remote data be exposed to potential clients according to

standard interfaces. On the other hand, document management systems must be modiﬁed to provide this new

kind of data access. This paper introduces a novel approach to support remote data integration in documents

handled by Content Management Systems. According to the reported experience, such a data integration can

be regarded as an enabling feature for a better exploitation of document workﬂow systems.

1 INTRODUCTION

Document management systems (DMS) are primary

tools for enterprises to deal with information involved

in business procedures (Baresi et al., 1999). Often the

enterprise knowledge base can be practically identi-

ﬁed with a collection of documents, to be handled

from creation up to ﬁnal archiving (Jones, 2007).

A signiﬁcant portion of business proceduresis car-

ried out by operating upon speciﬁc documents, whose

life cycle is determined by the progression of the busi-

ness task. The automation of this kind of activities

requires setting up both document templates and a

precise deﬁnition of document ﬂow (Rockley et al.,

2002). From this viewpoint, documents related to a

particular business process can be created as instances

of a procedure-speciﬁc template, and their content ex-

periences different and successive forms of compila-

tion. Thus, according to (Marchetti et al., 2005), in

the following we shall refer to Document Workﬂow

(DW) as to a particular workﬂow in which all activi-

ties turn out to documents compilation.

Several actions in a document workﬂow require

the end user to insert either non-structured or struc-

tured information. Although automated support to

the former can be hardly provided, the latter could be

more easily and proﬁciently dealt with, as in the case

of exploitation of ontological knowledge bases (Bec-

hini and Giannini, 2011). Most of the times the val-

ues to be inserted in a document are stored in an en-

terprise repository, or generally on a system that can

be accessed via Internet. In the following, such kind

of data will be referred to as external data.

In the last years, software tools dedicated to doc-

ument management and Web Information Systems

(WIS) have evolved side by side, and the advan-

tages of their merging have been recognized very

early (Balasubramanian and Bashian, 1998). A fur-

ther evolution has led to embed DMSs directly within

larger web application frameworks, namely Content

Management Systems (CMS) and Enterprise Content

Management (ECM) systems (Cameron, 2011). Sur-

prisingly enough, the web infrastructure has been ex-

tensively used for easily connecting users to the DMS,

but not to make external data directly available from

the DMS.

In CMS/ECM frameworks, access to external data

should be provided as a basic means to speed up doc-

ument compilation. This can be achieved by integrat-

ing a speciﬁc functionality in the framework, and by

accordingly modifying/enriching the standard docu-

ment handling approach. Examples on how to seam-

lessly operate on DMSs to obtain an integration of this

kind at the document repository level are reported in

(Bechini et al., 2008) and in (Bechini et al., 2007).

Anyway, external data are not always exposed to the

346

Gazzè D., Bechini A., Avvenuti M., Tesconi M. and Marchetti A..

Integration of External Data in Document Workﬂows via Web Services.

DOI: 10.5220/0004143203460351

In Proceedings of the International Conference on Knowledge Management and Information Sharing (RDBPM-2012), pages 346-351

ISBN: 978-989-8565-31-0

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

Internet or, if they actually are, not necessarily ac-

cording to a standard interface. For this reason, we

must specify how to implement data access on the

repository side as well. To this aim, Web Services

(WSs) provide a ﬂexible solution that keeps the sys-

tems completely decoupled. Moreover, past integra-

tion experiences on DMS, based on WSs, have shown

to be both effective and efﬁcient (Bechini et al., 2008).

In particular, (Chieu and Zeng, 2008) proposed to

build up a whole Content Management System ac-

cording to the SOA paradigm, using WSs as found-

ing components. In a slightly different perspective,

we advocate the use of WSs for data integration as

a means to provide ﬂexibility in the exploitation of

document workﬂows, with no stringent need to reor-

ganize the complete DMS according to a distributed,

SOA-compliant design. The outlined ideas led to the

implementation of a dedicated module in Drupal

one of the most popular CMS currently used: This

system will be described to show how the proposed

approach can be practically applied.

The paper is organized as follows: section 2

presents the typical way document workﬂows can be

speciﬁed and handled in CMSs, taking as a reference

the Drupal framework; in section 3, the proposed ap-

proach to integrate external data access in the standard

document management procedures is discussed; case

studies with practical workﬂows are shown in section

4. Conclusions are ﬁnally drawn.

2 DOCUMENTS AND

WORKFLOWS IN CMSs

Nowadays Content Management Systems let users

organize whole knowledge bases. To this aim, the

speciﬁc mechanisms put in place may be different

across different CMS platforms. We are interested

in investigating how document workﬂow can be ad-

dressed; in the panorama of open-source applications,

this particular feature is present in popular CMS/ECM

platforms like Drupal and Alfresco

, which adopt an

event-based workﬂow management mechanism.

Drupal is a very popular open-source content

management platform which has been used so far for

the development of millions of websites and applica-

tions (Miles and Miles, 2011). It makes use of an

abstraction of the document type, known as content

type, which can be enhanced to accommodate also the

access to external data. On the other hand, Alfresco

handles document workﬂow using content rules. In

http://drupal.org

http://www.alfresco.org

Figure 1: Access to structured information beyond the en-

terprise boundaries.

practice, the overall workﬂow can be considered as a

sort of a collection of possible paths along dedicated

spaces belonging to the involved actors. The move-

ments of a single document across spaces are auto-

matically driven by the system, according to rules de-

ﬁned at the space level.

In both platforms, the access to external data can

be implemented with no substantial modiﬁcation to

their core software, exploiting the respective standard

extension mechanisms and the underlying functional-

ities/modules. The Drupal’s content type abstraction

allows an elegant integration of remote data access,

which is independent from the workﬂow progression

mechanisms: This is the main rationale for focusing

on an actual implementation of the remote data access

system on this CMS platform.

In Drupal, each content item is represented as

a node, whose structure and behavior are deﬁned

through the speciﬁc content type it belongs to. Hence,

in a typical enterprise setting, it is a common prac-

tice to setup a content type for each document type

required in business processes. In particular, content

types might be deﬁned to formally support ordinary

administrative forms.

The overall framework is organized into modules,

a sort of functional plugins. Custom modules can

be developed and freely provided by members of

the Drupal community. Modules can both rely on

the CMS core functionalities, and exploit features of

other modules. They can also be shaped to permit

the customization of ﬁelds within node types. One

particular basic module is crucial for supporting re-

mote data integration: The Content Construction Kit

(CCK), which let the CMS administrator add custom

ﬁelds just operating via a web graphical interface.

Custom ﬁelds are intended to store additional struc-

tured information, respect to the system defaults.

Workﬂow management is available through the

workﬂow module. Arbitrary workﬂows can be cre-

ated and assigned to node types. Moreover, the rules

module provides the ability to deﬁne actions that are

conditionally triggered by occurring events. Thanks

to such modules, the DMS administrator can design

IntegrationofExternalDatainDocumentWorkflowsviaWebServices

347

RemoteCCK

Invocation

Interface

Remote Document

Manager

CCK

Access Control

CMS Core Services

User Interface

Figure 2: Architecture of the proposed solution. The Re-

moteCCK module is accommodated in a middle layer.

a document workﬂow for some of the content types

created with CCK.

3 SEAMLESS INTEGRATION OF

EXTERNAL DATA

The central problem tackled in this paper can be rep-

resented as shown in Figure 1. The access to struc-

tured data beyond the boundary of the enterprise in-

formation system must necessarily comply with con-

straints set by the external data server. Often the data

hosting organization is interested in exposing part of

its knowledge base via data access services. Further-

more, the easiest the access method, the larger the

beneﬁts that can be obtained. Thus it is not surprising

ﬁnding out that nowadays many organizations choose

to offer different services though the Web Services

(WS) technology (Chappell and Jewell, 2002).

The widespread adoption of the WS technology

is also determined by the modest coding effort to put

into operation a data access service. In fact, modern

IDEs for several languages are able to automate most

of the required coding steps. On the other hand, the

development of a WS “client” code is likewise really

simple. These characteristics make WS a key tech-

nology for decoupling a complex service into small,

manageable sub-services. WS have become pervasive

in information systems of any kind, and their adoption

in the proposed approach let us target the largest pos-

sible DMS users community.

Apart the basic choice of exploiting WS to con-

nect data access providers and a DMS/CMS client, it

is important to properly design the system architec-

ture to take advantage of the existing functionalities

in Drupal and to easily integrate the new modules in

the standard software platform.

3.1 System Architecture

The overall architecture for the proposed integration

system is organized according to a layered scheme, as

shown in Figure 2. In the depicted scheme, the two

lowest layers represent the core CMS and the Content

Construction Kit (CCK) module

. They offer the ba-

sic functionalities to create document templates, man-

age the life cycle of each single node, perform ver-

sioning, create and manage users, carry out logging

procedures, etc. CCK is devoted to handle the cus-

tomization of document templates.

The remote data access function is embedded in

RemoteCCK, the third layer of the system architec-

ture. It contains the basic services of our enhanced

data access system, which is able to exploit external

WSs. It encompasses two distinct functional blocks,

aimed at managing web services and users. In par-

ticular, the Remote Document Manager (RDM) block

extends the CMS core functionalities by supporting

the creation of document types, making them able to

access and process data via WSs. We designed the

RDM to let the user create a manual mapping between

an existing content type and an “operation” (i.e. a

data request towards WS), or a composition or WS

operations. The Invocation Interface block has been

expressly created to support the invocation of remote

WSs from the classical web interface of a document.

The uppermost layer, the User Interface, offers the

CMS administrator and ordinary users a web GUI to

properly interact with the system. In its basic form, it

is provided by the CMS, and it can be possibly modi-

ﬁed and enhanced according to the needs of new cus-

tom modules.

3.2 RemoteCCK

In the remote data access sytem, two different user

roles are deﬁned: The manager, who is in charge

of setting up the binding between content types and

WSs, and the document user, who must be presented

the outcomes of the WS invocations in a straightfor-

ward way.

Inside RemoteCCK, the RDM block supports:

• Analysis of WS speciﬁcations.

• Operation-driven content type creation and up-

date.

• Mapping of operations to content types.

• Creation of WS templates.

http://drupal.org/project/cck - In recent versions, the

CCK functionalities have been moved to the CMS core.

KMIS2012-InternationalConferenceonKnowledgeManagementandInformationSharing

348

Figure 3: WS are inspected via their WSDL speciﬁcation.

It is the Invocation Interface that actually calls WS

operations, according to three different modes: i) Be-

fore, ii) During, and iii) After. In the “Before” mode,

the WS is invoked at document creation time, i.e. be-

fore compilation start. In the “During” mode, the

invocation is manually triggered from the web in-

terface by the end user at compilation time. The

“After” mode must be used whenever the WS has

to be invoked at the end of the document compila-

tion/submission. Regardless of the chosen mode, the

operations to be carried out are: i) speciﬁcation of the

parameter values for the service request; ii) service in-

vocation; iii) insertion of the returned values into the

proper document ﬁelds. The WS operation parame-

ters must correspond to values in speciﬁc document

ﬁelds, and the returned values have to be inserted in

other speciﬁc document ﬁelds.

3.3 Enhanced Content Type

Speciﬁcation

The binding of external WSs to document ﬁelds is set

up by the manager following a two-step procedure:

i) registration and analysis of a new WS, and ii) cre-

ation of a link between a content type and operation(s)

of a registered WS.

In the ﬁrst step, the URL for the target WSDL

must be provided in input, and the corresponding WS

is then analyzed in terms of Operation, Message, ﬁeld

name and type as described in (Chappell and Jewell,

2002). During the WS registration, additional infor-

mation about the WS itself can be speciﬁed; typical

examples are the login username, the password, and

possible encodings. The graphical outcome of this

process is shown in Figure 3.

The second step involves the mapping of a WS

operation onto a content type ﬁeld. RemoteCCK pro-

vides three alternative procedures to accomplish this

task: 1) creation of a new content type, 2) update of

an existing content type, or 3) manual mapping.

The ﬁrst two procedures let the manager carry out

the mapping automatically. This way, the manager is

only asked to select a WS operation among the reg-

istered ones. In particular, the ﬁrst procedure creates

a new content type skeleton, whose structure mimics

the WS interface. The second one behaves in a sim-

ilar way, by adding new ﬁelds to an existing content

type.

Conversely, in the manual mapping point-to-point

links must be drawn on a graphical interface (see Fig-

ure 4). A list of the content type ﬁelds is depicted

against a similar representation of the parameters of

the WS operation; the manager can then easily draw

lines to connect endpoints of the opposite sides.

Figure 4: Mapping of operations onto content type ﬁelds.

3.4 Composition of WS Outcomes

To make the proposed data access and integration

mechanism as general and ﬂexible as possible, it must

be possible to combine the outcomes from multiple

(homogeneous) WSs. A straightforward way to per-

form this task can be based on the application of ag-

gregate functions, which semantically correspond to

the typical SQL-like ones like maximum, minimum,

average, count and so on. How to employ an ag-

gregate value depends on the content type semantics,

and it is up to the manager to carefully specify how

to compose and use it. In our system, the deﬁnition

of WS composition is supported by aggregation tem-

plates, which make available the calculation in the

form of a single reference, to be subsequently used

by RemoteCCK for content type updates.

The User Interface for managing an aggregation

template is quite complex and let the manager choose

different operations from many WSs. For the sake

of simplicity, the whole process can be divided in

two phases: A preliminary selection of the involved

WS/operations and the aggregate function, and the

subsequent mapping between ﬁelds and operation pa-

rameters. This last task can be graphically organized

as shown in Figure 5.

IntegrationofExternalDatainDocumentWorkflowsviaWebServices

349

Figure 5: Graphical speciﬁcation for aggregations.

4 CASE STUDIES

The validation of the proposed mechanism has been

organized in two parts: In the former, remote data ac-

cess and composition is tested; In the latter, a proﬁ-

cient integration within document workﬂow systems

must be achieved. In a preliminary experimentation,

WSs have been developed to access a master data set.

In such a data model, every employee has a personal

and unique ID. In document workﬂows each involved

employee, as document user, is asked to insert per-

sonal data. Thus, a speciﬁc WS can be developed

to obtain such information, given an employee ID.

On the DMS side, different content types can actu-

ally insert the required data into the proper document

ﬁelds, upon the related WS invocation (as shown in

Figure 6). We can complete the ﬁrst validation part

by deﬁning a report with the current cheapest price

for a product sold in a given list of malls. To this aim,

an aggregator can be used to determine the lowest

value out of the prices returned by different WSs; the

speedup due to the automated retrieval of distributed

information is here particularly evident.

In both the described cases, the proposed mecha-

nism has shown to be effective and ﬂexible.

4.1 Integration in Document Workﬂows

In the next example, we want to check the use of

RemoteCCK within the ordinary Drupal Document

Workﬂow System. Indeed, we can observe that Re-

moteCCK relies on CCK, whose functionality is typi-

Figure 6: WS-based access to a remote master data set.

cally exploited in a Document Workﬂow context, and

thus a straightforward integration is expected.

An integration test was performed by implement-

ing a typical example of use: A “vacation request

form” to be issued to the administration by an em-

ployee. In the ﬁrst place, we must underline that in

Drupal workﬂows a speciﬁc ﬁeld may be editable for

one user, only readable for another, or even hidden

for a third one. The workﬂow, which describes the

steps the application must undergo before being ap-

proved or rejected, involves multiple actors and can

be described as in Figure 7: The employee, using the

proper form, in the ﬁrst place applies to the Admin-

istration and to the Manager. Then, upon both ap-

provals, the document passes to the Director who will

forward the ﬁnal outcome to the Human Resource Of-

ﬁce for the ﬁnal notiﬁcation to the requester. Note that

in case of rejection at any level, the employee must be

always notiﬁed about it.

In this use case, the document progression has

been correctly managed, taking into account user

roles, assessment of conditions, and leveraging also

the “document creator” concept to determine the no-

tiﬁcation recipient.

Employee

Administration

Director

Manager

HR Office

Figure 7: Workﬂow for a Vacation Request Form.

KMIS2012-InternationalConferenceonKnowledgeManagementandInformationSharing

350

Figure 8: WorkMail can seamlessly integrate RemoteCCK.

4.2 Integration with WorkMail

WorkMail (Gazz`e et al., 2012) is a custom, fully con-

ﬁgurable Drupal module for documental workﬂow.

It is based on using emails to drive the compilation

ﬂow of shared documents. Users can access a docu-

ment according to their roles. A special content type,

namely WorkMail attachment, is dedicated to support

the workﬂow by including information about the re-

cipients, the subject, the body, and status ﬂags. The

WorkMail engine is in charge of regulating the users’

accesses to the document, according to their permis-

sions. One of the main strenghts of the WorkMail

system is its ability to resort to a delivering medium

(emails) that is particularly familiar to the end users,

thus making them easier collaborative jobs.

WorkMail is based on the CCK modules, and this

makes the RemoteCCK integration automatic. Users

that access the document via the WorkMail engine can

beneﬁt of all the RemoteCCK features, as shown in

Figure 8.

5 CONCLUSIONS

The efﬁcient employment of document workﬂows re-

quires an integrated access to data residing beyond

the boundaries of the enterprise information systems.

The proposed solution can be organized in a module,

called RemoteCCK

, to be seamlessly placed in the

Drupal framework; the new functionalities can be pro-

ﬁciently and ﬂexibly used both in the ordinary DMS

environment, and in additional custom modules. The

Available under GPL2 license at

https://drupal.org/sandbox/davide.gazze/1360152

possibility to graphically specify the inclusion in doc-

uments of plain/aggregated external data (accessed

via WSs) makes the module easy to use.

In its current implementation, at runtime Re-

moteCCK does not deal with possible modiﬁcations

of the registered WS operations; this problem could

be overcome by carrying out a WS validation step,

which may be included in future developments.

REFERENCES

Balasubramanian, V. and Bashian, A. (1998). Document

management and web technologies: Alice marries the

mad hatter. Commun. ACM, 41:107–115.

Baresi, L., Casati, F., Castano, S., Fugini, M. G., Mirbel, I.,

and Pernici, B. (1999). WIDE workﬂow development

methodology. SIGSOFT Softw. Eng. Notes, 24(2):19–

28.

Bechini, A. and Giannini, R. (2011). Management of

genotyping-related documents by integrated use of se-

mantic tagging. T. Large-Scale Data- and Knowledge-

Centered Systems, 4:15–39.

Bechini, A., Tomasi, A., and Viotto, J. (2007). Document

management for collaborative e-business: Integrating

ebxml environment and legacy dms. In ICE-B 2007 -

Proc. of the International Conference on e-Business,

pages 78–83. INSTICC Press.

Bechini, A., Tomasi, A., and Viotto, J. (2008). Collabora-

tive e-business and document management: Integra-

tion of legacy DMSs with the ebXML environment.

In Interdisciplinary Aspects of Information Systems

Studies, pages 287–293. Physica-Verlag HD.

Cameron, S. (2011). Enterprise Content Management: A

Business and Technical Guide. British Comp Society

Series. British Computer Society.

Chappell, D. A. and Jewell, T. (2002). Java Web Services -

using Java in service-oriented architectures. O

Reilly.

Chieu, T. and Zeng, L. (2008). Service-oriented approach

for implementing an extensible content management

system. In Congress on Services Part II, 2008.

SERVICES-2. IEEE, pages 96 –103.

Gazz`e, D., La Polla, M., Marchetti, A., Tesconi, M., and

Vivaldi, A. (2012). WorkMail: Collaborative docu-

ment workﬂow management by email. In Proc. of 9th

Int’l Conf. on Cooperative Design, Visualization, and

Engineering, LNCS. Springer.

Jones, J. I. (2007). The Document Methodology: for Enter-

prise Analysis, Second Ed. AuthorHouse.

Marchetti, A., Tesconi, M., and Minutoli, S. (2005). XFlow:

An XML-based document-centric workﬂow. In Proc.

of WISE 2005, volume 3806 of Lecture Notes in Com-

puter Science, pages 290–303. Springer.

Miles, E. and Miles, L. (2011). Drupal’s Building Blocks.

Addison-Wesley Professional.

Rockley, A., Kostur, P., and Manning, S. (2002). Manag-

ing Enterprise Content: A Uniﬁed Content Strategy.

Pearson Education.

IntegrationofExternalDatainDocumentWorkflowsviaWebServices

351