Anda di halaman 1dari 16

Content Repository Report

Survey Summary

Survey Results for the Survey Concerning Standards


for Enterprise Content Management
Author: Pascal Schrafl

Bachelor of Arts (Honours) Multimedia

Web: www.contentrepositoryreport.com
Email: info@contentrepositoryreport.com
Contents
1. Note to the Reader 1

2. Executive Summary 2

3. Survey Results 3
3.1. Usage of Content Management Systems . . . . . . . . . . . . . . . . . . . . . 3
3.2. Number of Content Management Systems . . . . . . . . . . . . . . . . . . . . 3
3.3. Solutions to Interoperability Issues . . . . . . . . . . . . . . . . . . . . . . . . 4
3.4. Important Functions of Content Management Systems . . . . . . . . . . . . . 5
3.5. Standards in General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.6. Importance of Java in the Enterprise Application Field . . . . . . . . . . . . . 8
3.7. API vs. Protocols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.8. JSR-170 vs. iECM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

4. Conclusion 12

A. List of Abbreviations 13

B. Copyright 14
1. Note to the Reader

Dear IT Decision-Maker,

This is a summary of the IT survey questionnaire you answered on the Internet sometime
during the first three weeks of March 2007. We are grateful for your participation. The
survey was one of three research tools used to evaluate the options available to IT decision-
makers for the purpose of managing the digital content that seems to be exploding in volume
for all of our enterprises. A special focus of the study was what is called unstructured
content, such as documents, reports, graphics, and website pages, which does not lend itself
to being kept in databases. It is this unstructured content that is growing exponentially.
What options are currently available to IT decision-makers for handling this type of content?
What parameters for handling it are important to you as a decision-maker? How important
are platforms, languages, licensing fees, and standards? Another focus of the research has to
do with accessing the various systems holding information in proprietary formats. How much
of a problem is this for you? What options are you aware of for dealing with it? How would
you evaluate these various options? Another of the research tools we used for this study was
an in-depth review of the information available on the five primary contenders for addressing
these problems. Essential information was extracted and organized for comparison purposes
for

1. Database Abstraction Layers

2. WebDAV

3. JSR-170 (Java Content Repository API)

4. Web Services

5. SOA and iECM

Our third approach was to do an in-depth interview with David Nüscheler, the leader of
the JSR-170 initiative (and CTO of Day Software AG in Basel), to gain his perspective
on the various options. Mr. Nüscheler gave us an objective and impartial overview, and
made predictions concerning roadblocks for the SOA and iECM standards process that have
already come dramatically true. Our intent with this research is to provide you, as an IT
decision-maker, with all of the perspectives and information you will need to make informed
choices concerning content management systems in the foreseeable future. The full report,
which will be released in November 2007, is available for a very reasonable cost. You can
place your reservation using the online reservation form on the Content Repository Report
Website at www.contentrepositoryreport.com. Here then is the summary of findings from the
Internet survey. Thank you again for your participation. If you have any questions, please
don’t hesitate to contact me at info@contentrepositoryreport.com

With best regards,

Pascal Schrafl

1
2. Executive Summary

The results of the survey, conducted during March of 2007, highlight that IT decision-makers
would most like to adopt standards if they are open and under free licensing. The standard-
ization process itself is perceived as an important factor in the acceptance of the standard;
and open, public and free standardization processes are preferred. Supporting the need for
standards is the fact that CMS clients with more than six CMS applications are trying to
keep the number of CMS vendors low, in order to achieve interoperability by either using
custom connectors or purchasing different CMS applications from the same vendor, hoping
to reduce the lack of interoperability by supporting proprietary standards. That this tactic is
only a temporary solution is shown by the fact that the vast majority of respondents would
prefer to use non-proprietary standards, if they are available.

Surprising is the result that the management of structured data is perceived as the most
important function of a CMS application, which can lead to the conclusion that IT decision-
makers perhaps see no way at present to effectively store or manage unstructured data in a
standardized manner.

The Java programming language is perceived as an important player in the enterprise ap-
plication field and its platform independence is an important factor in selecting Java. In
general, the programming language is not regarded as important for a standard, as long as
there are many implementations of the standard in different programming languages.

The broad support of standards by many different applications is rated as important, and
can be seen as an important factor in preferring older, but well established and supported
standards, as is the case with WebDAV. New standards are partially favored, as they are
more up-to-date, a trend that would speak against relying solely on WebDAV. A negative
influence on the decision concerning WebDAV is the fact that IT decision-makers appear to
believe that their programmers would have better success with an API standard than with
a protocol standard, which requires more programming skills and a deeper knowledge of the
internal commands of the protocol, as there is no API layer to capsulate and abstract the
protocol layer from the programmer.

There was a statistically nonsignificant preference for iECM/SOA over JCR, which we will
explore in the following summary, but the preference must be regarded as provisional, since
the iECM/SOA standard has not yet reached the draft state.

2
3. Survey Results

The purpose of the survey was to ascertain how IT decision-makers value content management
standards in general and how they would compare particular standards, such as JCR, Web-
DAV and iECM. David Nüscheler, one of the key figures in the CMS standardization process,
says that no research has yet been done in comparing IT decision-maker attitudes towards
JCR, WebDAV, and iECM. This survey is therefore the first of its kind. As participants in
the survey, you are the first group to learn the results.

3.1. Usage of Content Management Systems

The initial questions collected basic information about the systems already in place:

1. Does your organization have a content management system?

2. How many different applications from the content management systems field does your
organization have?

3. How many vendors do you have for your different content related applications?

The number of content management systems is addressed in question 1. As apparent in


Figure 1, 58% of the respondents have several systems in use, 33% have one system, only
3% have no CMS system at all, and 6% are planning a CMS system. Therefore, 91% of the
respondents have at least one CMS system currently in use.

2.7%
6.3%
Several CMS-Systems

One CMS-System
33.3%
57.7%
No CMS-System

Planning a CMS-System

Figure 1: Percentage values of the number CMS systems in use.


Total respondents: 112, Valid: 111, Missing: 1. Source: Pascal Schrafl, 2007

3.2. Number of Content Management Systems

The next two questions (Questions 2 and 3) assess the number of CMS applications and the
number of CMS application vendors.

3
Analyzing the results presented in Figure 2, it appears astonishing that there are more ven-
dors in the first two categories (one CMS application/one vendor and two to five CMS ap-
plications and the same range of vendors) than there are in companies with more than five
applications.
70
70
60
60
50
50
40 Applications
40
Percent

30 Vendors
30
20
20
10
10
0
0
1 2-5 6-10 11-15 16-20 > 20
Number of Applications / Vendors

Figure 2: Percentage values of CMS applications vs. CMS vendors.


Total respondents: 94, Valid: 94, Missing: 0. Source: Pascal Schrafl, 2007

It appears that organizations with 1-5 CMS applications prefer to have about the same
number of vendors for their applications, whereas organizations with 6 or more applications
prefer to minimize the number of vendors. This conclusion may be consistent with the
exponential growth of connectors or APIs necessary to combine different systems.

It also appears that interoperability among systems becomes a critical concern after about
five systems are in place, with the result that companies then tend to greatly reduce the
number of vendors with which they do CMS business. But is this the preferred solution to
the problem of interoperability, or merely a stopgap measure? Question 5 sought to find the
answer.

3.3. Solutions to Interoperability Issues

5. Please rank the following solutions to achieve interoperability between content manage-
ment systems by priority:

• Use content management systems from one vendor only.

• Use a custom connector for every individual content management system that needs to
be connected.

• Prefer content management systems that have adopted standards for interoperability.

As evident in Figure 3, as a group you prefer the support of standards for interoperability, as
80% ranked this solution with the highest priority. 53% of the participants rank the use of

4
custom connectors as the second best solution, and in the third and last position, with 47%,
is the tactic of purchasing all systems from a single vendor.
90
90
80
80
70
70 Purchase all CMS-Applications
60 from only one vendor
60
50 Custom Connector
Percent

50
40
40
30 Prefer CMS-Applications that
30
support Standards for
20 Interoperability
20
10
10
0
0
1 2 3
Ranking

Figure 3: Ranking of solutions to overcome the lack of interoperability.


Total respondents: 112, Valid: 112, Missing: 0. Source: Pascal Schrafl, 2007

This leads to the conclusion that IT decision-makers would prefer CMS applications that sup-
port standards for interoperability over implementing custom connectors or even purchasing
their CMS applications from only one vendor. CMS clients appear to have found a temporary
solution by reducing the number of CMS vendors, but would prefer standards for addressing
the interoperability issue.

3.4. Important Functions of Content Management Systems

A second major objective of the survey was to determine whether CMS is more important
for unstructured information (documents, web pages, e-mails, etc.) than for structured data
such as financial data and similar database-type information. Surprisingly, it was not.

The management of structured data is already solved by using ERP systems, but according
to many experts, only 10% to 20% of an organization’s data is structured, whereas 80% to
90% of the data is of the unstructured type, and needs to be managed in a different way
from structured data, by using ECM systems. As many organizations are doubling their
unstructured data every two months, the management of unstructured data should have
the highest priority in an organization and should reflect itself in being the most important
feature of a CMS application, especially as in many industries, ERP systems have become
standard practice and the management of structured data has become routine.

4. Please rank the following characteristics of a content management system by importance:

• Content Management Systems must be able to manage structured data.

• Content Management Systems must be able to manage unstructured data.

• Exchanging data between different content management systems must be supported.

• Legacy systems must be integrated.

5
• Interoperability between different content management systems must be supported.

• Being able to use data from different content management systems must be supported.
60
60
50
50
40
40
Percent

30
30
20
20
10
10
0
0
1 2 3 4 5 6
Ranking

Manage Structured Data Integration of Legacy Systems

Manage Unstructured Data Support Interoperability

Support for Data Exchange between different CMS-Applications Support usage of Data from different CMS-Systems

Figure 4: Ranking of the most important features of a CMS application.


Total respondents: 112, Valid: 112, Missing: 0. Source: Pascal Schrafl, 2007

Figure 4 reveals that the most important feature of a CMS application is the management of
structured data (52%). The second most important feature is the management of unstruc-
tured data (33%). Third place goes to supporting data exchange between different CMS
applications (28%), and the fourth most important features are integrating legacy systems
and supporting interoperability, both ranked in the same position (28%). The least important
feature is support for using data from different data sources (30%).

In contrast to expectation, respondents viewed the most important use of their CMS systems
not as the management of unstructured data but the management of structured data, a func-
tionality which, according to many ECM experts, should be already delivered by databases
and ERP systems. Nevertheless, the management of unstructured data ranks in the second
position, appearing to be important to IT decision-makers.

A second anomaly is the low ranking of support for data exchange and the support for using
data from different systems. The suggestion that a primary value of ECM is the aggregation
of data from different systems to broaden the organization’s knowledge base (unlocking the
enterprise content by breaking out of so-called content silos) can only be partially verified, as
this does not seem to be one of the most important functions attributed to a CMS system,
at least at this point in time.

It appears, then, that the management of structured data is regarded by participating IT


decision-makers as the most important function of their CMS systems. This result may be
interpreted in a couple of ways. First, IT decision-makers may perceive the management of
structured data as more important than the management of unstructured data because the

6
CMS applications in use in their organizations only work with structured data, as databases
are used to store the content. Alternatively, there could be a knowledge gap between the aims
of CMS vendors and CMS clients, which would be need to overcome by better communicating
the benefits of managing unstructured data.

3.5. Standards in General

Returning to the question of standards, IT decision-makers clearly favor open free standards
above all others (Question 6). Concerning the standardization process (Question 7), a free
and open process, involving all parties and allowing public access to the standardization
documents is preferred. Neither the JCP process nor the iECM process is completely open,
but this seems to be an important factor for IT decision-makers and should be considered by
the specification leads. Question 6, for example, asked you to:

Rank the following characteristics of standards according to your preference.

• Proprietary standards.

• Open standards without license fees.

• Open standards with Reasonable and Non Discriminatory Licensing (RAND) fees.

• Standards requiring license fees.


90
90
80
80
70 Proprietary Standards
70
60
60
Open Standards
50 without License Fees
Percent

50
40
40 Open Standards with
30 RAND License
30
20
20 Standards requiring
10 License Fees
10
0
0
1 2 3 4
Ranking

Figure 5: Frequency expressed in percentage values of four different standards and license
schemes.
Total respondents: 112, Valid: 112, Missing: 0. Source: Pascal Schrafl, 2007

Figure 5 presents the frequency of the four rankings of the different standards and licensing
schemes. Clearly visible with 80% is the open and license-free standards scheme. The second
most preferred is open standards with reasonable and non-discriminatory (RAND) licensing
(71%). The last two positions are not as evident as the previous two, but standards requiring
license fees and proprietary standards are the least preferred by participating IT decision-
makers.

7
3.6. Importance of Java in the Enterprise Application Field

We turn next to the question of individual standards, beginning with JCR. Question 8 asked
you to rate certain statements about Java as a programming language and its importance
in the enterprise application field. Figure 6 below graphs the frequency of the answers,
and highlights that the Java programming language has a strong acceptance in the field
of enterprise applications (30% absolutely agree, 47% agree, for a total of 77%). While
there was a firm conviction that a standard completely restricted to Java would not succeed,
especially according to respondents whose organizations use other programming languages,
the JCR standard has already been ported to many other programming languages. Also,
more than half of the IT decision-makers do not perceive the programming language of a
standard as important, so long as the standard is implemented in many different languages.
Figure 6 also displays that Java seems to be an important programming language in the
enterprise application field, as other programming languages do not find a high acceptance
in the Content Management field.

50
50
45
45
40 A Java-only standard cannot prevail,
40 as it is tied to the programming
35 language
35
30
30
Percent

25 Java is important, as it offers platform


25 independence
20
20
15
15
10 Java is not important, as we use other
10 programming languages
5
5
0
0
Partly agree
Absolutely agree

Agree

Disagree

Totally disagree

Programming language of a standard


is not important, so long as there are
many implementations in different
programming languages

Figure 6: Percentage values of the importance of the Java programming language in the
Enterprise Application field.
Total respondents: 112, Valid: 112, Missing: 0. Source: Pascal Schrafl, 2007

To conclude our findings in this area, the results clearly show that Java is an important
player in the field of enterprise applications, as it offers platform-independence. Neither the
Java-supporters nor the supporters of other programming languages appear to care much
about the programming language of a standard, as long as there are many implementations
of the standard in different programming languages, so that the standard can be used within
their own organization.

8
3.7. API vs. Protocols

A related question was raised concerning how much difference you as IT decision-makers
perceive between protocols and APIs. The reason for this question (see question 9) is that
some experts in the field perceive application programming interfaces (such as JSR-170) to
be stronger or easier to work with than protocols (such as WebDAV). As it turns out, the
final numbers showed only a very weak agreement (see Figure 7). Further statistical analysis,
however, revealed that there were actually two sets of much stronger preferences at work,
offsetting each other: one group that prefers APIs and another that prefers protocols. To
a certain extent, these preferences correlated with favoring older, well-established standards
(such as the WebDAV protocol) on the one hand, and favoring newer standards (such as the
JCR API) on the other.
60
60
Programmers only use APIs
50
50 Programmers use the protocol
40 Programmers use the API
40
Percent

30 WebDAV is supported by many applications


30 and should therefore be preferred

20
20 Older, established standards should be
preferred as their implementation is more
10 mature
10 New standards should be preferred, as they
are up to date
0
0
Absolutely agree Agree Partly agree Disagree Totally disagree

Figure 7: Percentage values for APIs vs. Protocols and new standards vs. older, established
standards.
Total respondents: 112, Valid: 112, Missing: 0. Source: Pascal Schrafl, 2007

A second important tendency can be spotted, as the broad support of a standard by many
applications seems to be an important factor in favoring a standard. This result is especially
important as it leads to the conclusion that the JSR-170 committee must promote their
standard and seek to have it supported by many applications, which in fact has happened.
For iECM, the other primary contender for ECM standard, this may be a major obstacle,
as the iECM committee, after more than two years of development time, has currently no
implementation at all.

3.8. JSR-170 vs. iECM

The next question of the survey asked you to evaluate several statements concerning the two
primary candidates for a new Content Management standard: JCR and iECM.

The full report provides detailed reasoning, but in brief, the other three original candidates
were either demoted or combined. First, DBALs were eliminated because database systems
are poorly suited for handling unstructured content such as documents and web pages. Web-
DAV is a very good adjunct to either JCR or iECM for Content Management, but was not

9
designed to convert all of an enterprise’s content from largely unavailable to readily acces-
sible. WebDAV was designed rather to handle individual documents on an as-needed basis.
Web Services, which are as compatible with JCR as with iECM, do not by themselves offer a
unified approach to content management. Instead, Web Services would be utilized by Service
Oriented Architecture (SOA) in order to access content from different systems. The SOA
concept, which is very popular right now, is the primary toolkit of iECM.

This leaves us with JCR and iECM as the primary contenders, and the process of elimination
just described provided the backdrop for asking you to evaluate several statements concerning
these two finalists.

While no strong preference appeared for either of these two standards, there was a slight
preference for SOA-based standards due to their language independence and the ability to use
the standard with service-enabled applications. Part of this preference may derive from the
fact that API standards involve programming, while SOA is being promoted as a technology
that involves minimal or no programming.

Statistical analysis showed that four of the responses best capture the very slight tendency
in favor of iECM/SOA. These four statements are:

1. A standard based on a service-oriented architecture can be used by all applications that


are service-enabled.

2. A standard should not depend on a programming language.

3. Standards based on a service-oriented architecture should be favored above all other


standards.

4. The top-down approach of a service-oriented architecture is to be favored above the


bottom-up approach of an API.

The above mentioned four items were combined into a new synthetic variable, called pro
iECM and were tested against the pro JCR items, leading to the slightly positive tendency
towards iECM as evident in Figure 8 below. The mean value of 2.42 for the pro iECM
variable is marked in yellow. As smaller numbers represent a higher level of agreement, the
slight trend towards iECM is visible.

10
18
18
16
16
14
14
12
12
Frequency

10
10
8
8
6
6
4
4
2
2
0
0
1,00
1,25
1,50
1,75
2,00
2,25
2,50
2,75
3,00
3,25
3,50
3,75
4,00
4,25
4,50

Figure 8: Histogram of the pro iECM variable. Mean: 2,4196. As smaller numbers represent
a higher level of agreement, the slight trend towards iECM is visible.
Total respondents: 112, Valid: 84, Missing: 28. Source: Pascal Schrafl, 2007

11
4. Conclusion

This report would be incomplete if we did not mention something not known by anyone at
the time of the survey, which is that the standards process for iECM/SOA has stalled. The
group’s deliberations have since been suspended, apparently because of the extreme difficulty
of determining a viable standard in this complex area. The suspension effectively leaves
JCR (as JSR-170 and its later version, JSR-283) as the only currently viable standard for
delivering on the promise of Content Management, which is to make unstructured content as
manageable and useful as structured database content is today.

12
A. List of Abbreviations

API Application Programming Interface


CMS Content Management System
DAV Distributed Authoring and Versioning
DBAL Database Abstraction Layer
ECM Enterprise Content Management
ERP Enterprise Resource Planning
iECM Interoperable ECM Standard
JCP Java Community Process
JCR Java Content Repository
JSR Java Specification Request
SOA Service-Oriented Architectures
WebDAV Web-based Distributed Authoring and Versioning

13
B. Copyright

Copyright
c 2007 by Pascal Schrafl.

All rights reserved.

This document, Results of the Internet Survey of IT Decision-makers, is copyrighted by the


author, Pascal Schrafl of Maur, Switzerland.

The recipient of this special report may download, copy, or print the contents for personal
use, provided that the recipient observes all copyright restrictions. Except for personal use,
no copying, storage, redistribution, selling, broadcasting, circulating, commercial exploit or
publication of this document, in part or in whole, is permitted without the express written
consent of the author, Pascal Schrafl.

In no event will the author or other persons transmitting the information of research reports
be liable to you or anyone else for any consequential, incidential, special or indirect damages
(including, but not limited to, lost profits, trading losses, and damages that may result from
the use of the information or the research reports or from inconvenience, delay or loss of
use of information or research reports or for omissions or inaccuracies in the information or
the research reports) even if advised the possibility of such damages and, as a condition to
accessing such information or research reports, you expressly waive any claim you may have
against the author or any other person with respect to any information or research report.

There is no warranty of merchantability, no warranty of fitness for a particular use, and no


warranty of non-infringement. There is no warranty of any kind, express or implied, regarding
the information of the research reports.

The author may be contacted by email at info@contentrepositoryreport.com The Content


Repository Report website can be found at http://www.contentrepositoryreport.com

14

Anda mungkin juga menyukai