Working Together, Apart

rangaleclickSoftware and s/w Development

Nov 4, 2013 (3 years and 7 months ago)

56 views

http://www.intranetjournal.com/features/idm0398
-
pm1.shtml


Working Together, Apart

The Web as Project Infrastructure



by Gordon Benett, Managing Editor, IDM


O
ne of the more effe
ctive methods of
organizing and tracking work

is management by
walking around: the simple act

of popping your head in a team member's
office and shouting, "Hey James! How's it going?" Do this across the team a few
times a month and your project will either

perform, or you'll know you need a new
team.

Trouble is, today's "teams" often have upwards of the half
-
dozen people it's feasible
to visit in a stroll, and these folks may well be scattered across your city, state or
continent. The solution is to visit
project resources
virtually

rather than
physically
, a
variation that might be called
management by flying around

--

except that travel
takes place in cyberspace, and you need not leave your office to do it.

Virtual project management (VPM) is the Informat
ion Age equivalent of management
by walking around. Most recently, the rise to dominance within organizations of
Internet
-
based collaboration tools offers new possibilities for
web
-
based project
management
.

In this article I'll discuss the value propositi
on for web
-
based VPM, focusing on a
couple of software products that illustrate both the potential and current limitations
of the medium. Interestingly, some of the best
real
-
world Java applications

implemented today support collaboration.

Doing, sharing,

measuring

"A virtual project is a collaborative effort towards a specific goal or accomplishment
which is based on 'collective yet remote' performance," according to a seminal paper
on the topic [
1
]. Another source sets "working together, apart" as the goal of
enterprise networking [
2
]. These modes of work share a need for
mana
gement
tools

that enable communication and coordination at a distance.

In addition, many projects require the concerted effort of several individuals sharing
a common set of tools. For example, an engineering team might use a computer
-
aided design (CAD) p
rogram to develop and compare design alternatives without
holding a physical meeting. Look at the kind of project management activities
associated with this process:



A
task

must be defined to develop alternative designs based on
project requirements



Reso
urces

--

people, time, expenses if any
--

must be allocated to
the task



The team members involved must
communicate

before and during
the task, both with each other (design issues) and with the PM
(administrative status reporting)



The PM needs to
track

th
e task, and based on performance relative to
allocated resources, administer course corrections.

There is room in these activities for several layers of information system support:



Communication

implies e
-
mail, phone calls, memoranda (hopefully as
e
-
mail

attachments but possibly as paper), and other media.



Collaboration

goes beyond basic communicating to sharing design
information, which for all but the simplest projects will reside in
specialized repositories such as CAD programs, CASE tools, simulation

software, etc.



Tracking and leveling resources

are functions performed by traditional
PM products.

In addition, some organizations are required by regulation or committment to a
specific methodology (e.g., ISO 9000) to maintain
complete configuration co
ntrol

over project artifacts. A pharmaceutical firm, for instance, might need to store not
only the final specification for a new drug but all alternatives and iterations leading
up to it. In such cases the dimension of
process management

can consume more
resources than the projects themselves.

Now do it virtually

D
rug manufacturers and bridge builders weren't exactly on hold until the advent of
distributed computing. All of the activities mentioned so far have been handled for
years
--

since the start of
the Industrial Revolution, in a sense
--

with pencil and
paper and human ingenuity. How do electronic information systems change this, and
how do web
-
based applications in particular add value?

Communication

This is networking's quick hit. E
-
mail
allows i
deas to flow asynchronously

(i.e.,
without parties online at both ends), enabling work to flow across holidays and time
zones. E
-
mail also creates with no incremental labor a
searchable audit trail
, key
to many formal processes.

Taking this idea a step fu
rther, consider how products like Microsoft® Outlook™
support and extend project communication. Outlook integrates a
multi
-
protocol e
-
mail client

with directory, scheduling, and journaling functions. Through
journaling
, the process of keeping a record of w
ork performed, Outlook extends the
concept of
automatic audit trail creation

to include phone calls, faxes and other
non
-
integrated communications.

Unfortunately, Outlook97 happens to be painful to use and miserably insular when it
comes to sharing projec
t info. But the integration potential is there, and hopefully
will be better realized in future releases. Meanwhile, similar functionality is on
tap in
competing Web
-
based messaging clients, such as Netscape Communicator™ and
Lotus Development Corp.'s Notes 5.0 client.

Remote collaboration

The ability to "work together, apart" is hardly possible without shared storage and
concurrency control
--

so
lved problems, thanks to client/server database technology.
A good example of a product that enables web
-
based collaboration is
NetObjects
TeamFusion
, reviewed in this issue of IDM
[
3
]. TeamFusion facilitates the
construction and maintenance of complex web sites by multi
-
developer teams.

In addition, TeamFusion allows contributors not directly involved

in web development
to add content through a form
-
based Java applet, accessible with any browser.
Browser
-
based project collaboration

is going to be one of Java's "killer apps,"
you betcha, and a theme you'll see more of in Part II of this article.

Of cou
rse, Java isn't the only distributed computing platform on the block. Microsoft's
Component Object Model (COM), lately folded into a new marketecture called DNA
(Distributed interNet Architecture), holds out much the same prospect as Java for
providing rem
ote access to shared workspaces and applications.

CORBA, on the other hand, does not. While the Common Object Request Broker
Architecture standard is arguably more mature and better integrated than DNA, it
lacks the the visual control elements present in
both Microsoft's ActiveX and Java's
AWT. This relegates CORBA to adding value behind the scenes while Java and DNA
duke it out on your desktop.

Project management

Integrated messaging tools like Lotus Notes and design collaboration tools like
TeamFusion d
o a good job within their respective domains. When a project manager
needs to lay out tasks, assign resources and track performance, however, she must
look elsewhere
--

specifically, to project management software. PM software adds
value by facilitating th
e administrative chores associated with teamwork, from
schedule production and cost estimation to critical path analysis.

"Is the project on target?" PM software's agenda is to answer this question, and as
anyone familiar with tools like Microsoft Project
, Primavera Systems SureTrak Project
Manager or Scitor Inc.'s PS6 can attest, they do little else. These products are
intended for use by professional managers, not by the members of a project team.
They add no value as task collaboration tools because the
y
don't understand the
vertical knowledge

of specific problem domains.

This kind of tool is, of course, indispensable within a narrow administrative domain.
Some products, like SureTrak Project Manager from Primavera Systems Inc. and
Microsoft Project 98
, can publish current project data to a web server, making status
information and associated files available to all comers through a standard issue
browser.
Web publishing

is much more efficient
from both a cost and client
configuration standpoint than the per
-
seat licensing model of older PM products,
which required every user who might conceivably need access to install a full copy of
the client software.

Both SureTrak and Project 98 also featu
re extensive
e
-
mail integration
. In
SureTrak's case, users can send messages about project data, screen captures, and
selected activities through a gateway called Primavera Post Office to team members
who can then review, approve and merge updates back int
o the project schedule.
Microsoft Project 98 goes these workgroup capabilities one better by giving users a
choice between e
-
mail and web
-
based communications [
4
].

These fea
tures bring aspects of "management by walking around" to the virtual
project realm. It's important to remember, though, that collaboration in PM software
remains
strictly limited to project management functions
. Even best
-
in
-
class
products like Project 98
and SureTrak Project Manager can't "reach out and touch"
vertical applications, for instance a drafting tool such as AutoCAD for construction
projects, or Rational Rose, Rational Software Corp.'s system modeling tool, for
software development projects.

As

a result, teams have had to turn to a
hodgepodge of non
-
integrated tools
,
each of which supports a facet of virtual work
--

project/process management,
project communication, or collaboration on project tasks. Only recently have tools
capable of providing

a complete process management framework for virtual work
begun to appear.

Web
-
based project infrastructures

L
et's recap. Virtual projects
--

"working together, apart"
--

require communication,
collaboration and project management. The present generation
of software aims to
support work
within

each of these domains by leveraging client/server technologies
such as shared data access, standards
-
based messaging and browser economies.

But not until now, with the impending rise of web application technologies
such as
Java, ActiveX and XML, have project teams had access to integrated environments
that bridge project domains. The goal of
integrated process management

through
a suite of cooperating tools seems at hand.

Next issue

I'll describe how corporate webs are becoming project infrastructure. In
particular, I'll profile two web
-
based products that support virtual project teams. One
is
Mesa/Vista Project Manager

by Mesa Systems Guild, Inc., a
high
-
end offering
that provides comprehensive process management through a rich mix of Java,
JavaScript and XML technologies. The other,
WebProject

by WebProject, Inc., offers
a more traditional PM feature set enhanced by an all
-
Java implementation.

Citations

1.

Terry Krile & Dr. Paul Juell, Proceedings of the Small College
Computing Symposium (SCCS'97), "Virtual Project Management,"
North Dakota State University
, March 1997.

2.

R. Grenier and G. Metes,
Enterprise Networking: Working Together
Apart
. Digital Press, 1992.

3.

In late 1998 NetObjects replaced TeamFusion 1.0 with Authoring
Server 3.0, reviewed
here
.

4.

Microsoft Project 98
Product Enhancements Guide
, article on "Web
-
based Workgroup Features."

B
usiness processes are becoming increasingly
virtual
--

staffed on the fly by transient, dispersed teams.

Achieving to
potential under these conditions requires a
distributed project
infrastructure

that helps people communicate, collaborate and manage shared
tasks in an integrated way.
Part I

of this articl
e outlined requirements for such an
infrastructure, pointing the way to solutions that leverage intranet technology.

In Part II we ask how well current product offerings are meeting these requirements.
The answer, in a word, is "imperfectly," but the prog
nosis has never been better.

To gauge the state of the practice I'll focus on
two web
-
based products
, both
available now, that support virtual project teams. One is Mesa/Vista Project Manager
by Mesa Systems Guild, Inc. [
1
], a high
-
end offering that provides comprehensive
process management through a rich mix of Java, JavaScript and XML technologies.
The other, WebProject by WebProject, Inc. [
2
], offers a more traditional PM feature
set enhanced by an all
-
Java implementation.

The process is the solution

C
alling Mesa/Vista project management software is a bit like calling Microsoft a PC
sof
tware vendor: it conveys a duplicitously thin sliver of truth. Mesa Systems Guild
was formed in 1990 by a management team from Cadre Technologies, Inc. (now
Cayenne Software, Inc.), maker of Team
work
TM
, an end
-
to
-
end software engineering
tool. From this ba
ckground comes Mesa/Vista,
a project environment

built upon
web technology. Mesa/Vista adds value through
process integration

rather than
project management.

The product's agenda includes web
-
based communication, which is where the
e
-
mail
or HTML publishi
ng

capabilities

of most PM software leaves off. Yet Mesa/Vista has
no native ability to produce Gantt charts or histograms, to level resources, or to
track budgetary estimates. Instead, Mesa's product interoperates with a wide range
of design and project m
anagement tools, including (as of this writing): Cayenne's
Team
work
, Rational Rose 98, TD Technologies' SLATE and Microsoft Project. Mesa
also sells a Developer's Kit for creating custom interfaces to other analysis and
design applications.

I said in Part

One of this essay that virtual projects entail
communication,
collaboration and management
as a multi
-
dimensional whole. Tools that service
only one or two of these needs offer limited gains to distributed team members,
forcing them to manually integrate
and update across applications.

Mesa/Vista ambitiously aims to cover the virtual project waterfront. It does so not
through a surfeit of features, but by acting as a
rich services environment

that
understands a variety of vertical project applications cal
led
plug
-
ins
.

Some of the services Mesa/Vista provides are similar to those available in Lotus
Notes or other groupware products. For instance, Mesa/Vista gives users the ability
to define roles, and to filter large volumes of project data on that basis.
Fine
-
grained access control

based on user profiles is also available. In a nod to
Microsoft Outlook's journaling capability, Mesa/Vista can log action items and project
events. Documents such as meeting minutes and other supporting files can be
moved to th
e web server (using HTTP upload) for secure access by team members.

Mesa/Vista's
document management features

are a subset of the full
-
strength
library services found in Lotus Notes or Open Text, Inc.'s
Livelink Intranet
. One
powerful plus is Mesa/Vista's ability to put any file under version control. Businesses
with well
-
defined regulatory or in
-
house methodologies will welcome this feature,
easily invoked by clicking a button. Access to managed docu
ments is subject to
check
-
in/check
-
out

and automatically generates an audit trail.

Mesa's wise inclusion of versioning makes the absence of indexed search from
Mesa/Vista all the more mysterious. According to the company, full
-
text search will
be offered
by mid
-
year. I'd like to see them embed something like Verity Inc.'s OEM
software, which can read and index office document formats such as Microsoft Word.
Notes and Microsoft's Index Server incorporate the Verity engine (
reviewed in IDM

last September).

Browser
-
based groupware features give Mesa/Vista its
process orientation
; the
tool assumes that projects take place in a managed context, and is tailored to
facilitate the activities and
record
-
keeping that context entails. But Mesa/Vista is
much more than a groupware wannabe.

Keep in mind that the features I've described to this point form a
horizontal
service layer
, on top of which Mesa/Vista supports supports tight integration with
des
ign and project management tools. Groupware and traditional workflow products
offer nothing like this
--

one reason, no doubt, why Mesa's customers include
pharmaceutical giant Abbott Laboratories and the Naval Air Warfare Center (NAWC).
Mesa/Vista is uniq
uely suited for managing multiple distributed projects that take
place in a formal process context.

To understand just how unique Mesa's product is, I suggest you visit their web site
and take the
Unguided Tour
. (The Guided Tour is cute but limited in scope.) Here I'll
give an example illustrating Mesa/Vista's
plug
-
in approach
, which differentiates it
from every other project environment I'm aware of.


Example: Mesa/Vista Tool Integration

I
magine that you're the team leader of a software design project code
-
named
Corker
, and you
need to review the status of a change order (CO) that resulted from a fix during acceptance
testing. Since Project Corker is being managed with Mesa/Vista, you can d
o this from anywhere
using a web browser. Once you log in, you query the change order and click on the resulting
hyperlink to bring up the Microsoft Project Gantt chart.

Mesa/Vista displays third
-
party application data like MS Project files through
Java
a
pplets

called
plug
-
ins

(no relation to Netscape's browser API of the same name).
As a result, you're looking at the latest Project export data, which indicates that the
CO of interest is in testing.

By right
-
clicking tasks in the Gantt chart you can pop u
p menus of relevant actions.
Note that this is much richer functionality than the graphic hot spots available with
static Gantt chart depictions made, for example, with Net
-
It Software's jDoc
technology. Mesa/Vista
understands the task semantics

and gives
you
appropriate options. For a testing task, these might include
Show Test Plan
, which
you click to bring up the underlying documentation.

This could be any type of information displayable in a web browser, but since Project
Corker was designed using Rati
onal Rose you have
direct, lightweight access

via
Mesa/Vista's Rose plug
-
in to the full UML specification. Action items can be initiated,
and e
-
mail linking project documents sent, directly from within the Vista
environment.


Mesa/Vista is hard to catego
rize because it stretches the categories into which we
traditionally place aspects of project work. This is just the kind of "stretch" we need
to envision tomorrow's fluid collaborations, however, for which Mesa/Vista is
conceptually well prepared. With th
e addition of
LDAP directory services

and a
solid search engine
--

both promised by mid
-
year
--

Mesa's product will set the
project infrastructure standard.

WebProject: the future of web
-
based PM

L
ike Mesa Systems, startup WebProject Inc. was formed by an

entrepreneurial
management team. But in contrast with Mesa's systems engineering pedigree,
WebProject founder Chris Ritke came from
Scitor Corporation
, maker of the
estimable Project Scheduler series of project management software. As a result,
WebProject

hews closer to the traditional PM feature set, which it powerfully extends
with an
all
-
Java implementation

and intelligent web
-
based design.

WebProject targets the "collaboration and communication" aspects of teamwork, but
its collaborative features are c
loser to those of Microsoft Project 98, which puts
project status data on the web, than of a shared working environment like
NetObjects TeamFusion or Mesa/Vista.

On the other hand, WebProject does an excellent job of facilitating
distributed
decision
-
maki
ng

through an integrated tool suite that includes real
-
time chat, a
voting/survey application, e
-
mail notification, and threaded discussion groups.
WebProject's use of multiple windows lets teams hold
virtual status meetings

with
concurrent access to proje
ct data. Furthermore, to help WebProject users create
effective virtual workspaces, the vendor offers an
online meeting facilitation

service through its consulting partner CONSENSA.

Alongside these communication features WebProject offers a unique Java
-
ba
sed
project management system. The product's
PlanTrack

module allows authorized
users to define projects, tasks and resources from any web seat. Web Project does
an impressive job of emulating familiar PM tools such as timelines and histograms in
Java.

Pr
ojects can be imported from or exported to the major PM software formats,
including Microsoft Project and Scitor PS7. This is a
healthy nod to legacy
projects
, allowing WebProject to pick up existing resource lists. A bigger step in this
direction would be

access to LDAP directories. No distributed project management
tool should be without this key service interface.

The product's commitment to Java has advantages and disadvantages. On the
upside, WebProject achieves its goal of
platform and location indep
endence

--

the only PM software of comparable sophistication that does so. Shared project
information held in relational databases can be accessed from anywhere using JDBC.
But on the server
-
side, where WebProject's resource management and query engines
li
ve, Java's performance is not especially scalable. Also problematic is the
inability
to print Java
-
generated displays

such as Gantt charts and resource loading
histograms.

These are limitations of the Java language that will fade as the platform matures.
What WebProject offers in spite of these caveats is a distributed project
communication environment that makes unprecedented use of client/server web
technologies.

For a more detailed look at WebProject, check out the
online help system

or
download a 30
-
day trial version

of the software.

Thinking ahead

T
his article began with a description of the challenges facing
virtual projec
t teams

--

resources convened rapidly to collaborate remotely in pursuit of shared goals.
Meeting these challenges takes more than a set of discrete messaging and project
management tools.

Most of today's software takes a discrete approach. The second par
t of this article
focused on two products that aim at, and achieve, higher integration.

Mesa/Vista emphasizes teamwork in a formal process context. WebProject distributes
complete project information using Java. Both products facilitate communications,
ma
nage access and run in a standard issue web browser. Both products point the
way to a generation of collaboration software that can meet virtual project
challenges.

The rise of web integration frameworks such as LDAP, JavaBeans, DNA and XML is
laying the foundation for work that bridges geography, time and project domains. As
a result, I am confident that
integrated process management

will become the
state of the prac
tice within a few years, and features such as those described here
will enter the software mainstream.