You may have to dig beneath the hype a little, but at any gathering of 40
Java vendors there's bound to be some treasure buried in there somewhere.
It's just waiting for you to find it.
If you have children, you tend to measure time by their needs. The first
day of summer camp I was at the Web Services Edge 2002 East Conference (hope
you found my "Show Report" helpful [JDJ, Vol. 7, issue 8]), and the first
day back to school found me at Wall Street IT - The Next Generation.
Held September 4 and 5, 2002, at the Metropolitan Pavilion in New York
City, Wall Street IT - tng gathered almost 40 vendors on the show floor and
backed them up with 15 seminars and presentations by some high-level players
from technology companies and major Wall Street firms. Produced by
Lighthouse Partners with show management by Flagg Management Inc., this
conference provided a good look at the state of technology.
Don't worry if you're not on Wall Street. The show had plenty of good
information and, more important, good news for the Java world. Sure the
focus was on using technology in the financial services industry, but many
of the vendors work in other spaces as well and these conferences presented
a great opportunity to read between the lines. While everyone was waving the
Web services banner, by listening closely, I realized something rather
valuable: despite how dank and miserable it may be out there, Java is not
going away, maybe even can't go away.
Many of the vendors on the floor were Java shops. In and of itself this
is neither surprising nor very inspiring. People still need technology, so
it's inevitable that some Java work will get done. It's when you look closer
that things become interesting.
California-based DevelopMentor provides technical training in Java, XML, Web services, and .NET. They were telling me that at an XML show the prior week, around 7080% of the show attendees they spoke with were asking about courses on how to do Web
services with Java.
Ah. Now we're getting somewhere. If you think about it, trainers have a
unique view of the market, especially in tight times like these. Budgets
allow for only the most necessary of items. Training courses have to be
directly relevant to existing projects.
With this thought in mind, I spent about 30 minutes with Greg Brill,
president and founder of Infusion Development Corp. and the editor and driving force behind the CodeNotes book series. Infusion develops custom training courses for corporate clients. Apparently, they're no longer called on for basic Java courses. It's all
Java and ___. Clients want a blend of technologies. Infusion recently did a
14-day training course for Lehman in Tokyo a Java, EJB with WebLogic, and
XML/XSL course. Demand for Infusion's courses has remained steady and
perhaps even grown a little. This is because their courses are
custom-developed to suit the needs of particular projects and, as mentioned
earlier, those projects still exist.
What Brill is seeing is that Java is the facilitator of integration;
this shows Java's maturity and penetration. In the boom days, Java's hype
machine was working full-time: everyone was putting out J2EE servers and
everyone else was using them. Now Java is everywhere. It's a fact of life.
No matter what happens in the market, no matter what happens with .NET or
anything else, Java is not going away. As Brill put it, Java colonized tech
companies much as the Spaniards colonized the New World. Spanish is spoken
everywhere south of the U.S.
Web services is pretty much guaranteed to play a big part in any IT
conference these days. Wall Street IT - tng did not disappoint. Day one
provided us with a seminar given by Anne Thomas-Manes of Systinet entitled
"Web Services for Technology Managers," in which she gave a fairly
high-level view of the technologies behind Web services and laid out how Web
services work. She also discussed which companies are providing Web services
solutions and gave some guidance on how to pick from among this field.
As you know, aside from SOAP, WSDL, and UDDI, most of the Web ser-
vices standards are still evolving. The industry is getting closer to having
standards that deal with security, workflow, and transactions, but we're not
there yet. This would be a good place to note that the only thing that
doesn't seem to move on Internet time is the development of the standards
underlying the whole mess.
There seems to be a burgeoning awareness that the exposure of
functionality via Web services, what is basically cross-language RPC, is not
all that sexy. According to Thomas-Manes, the best a SOAP message can do is
two to three times slower than RMI, and that's top speed on a small, simple
message. You'll really need something more to make this take hold after the
hype dies down.
There was a panel discussion called "Future Technology Platforms for
Deploying Web Services." The moderator was Frank Greco of Crossroads
Technologies and the panelists were Jim Bole of Infravio, Dmiitri
Tchikatilov of Microsoft, Ed Schwarz of Sun, Adam Greissman of UDICo, and JP
Morgenthal of Software AG. It was during this discussion that the "something
more" came out. The panel reiterated that Web services are too slow and not
secure and that simply exposing old functionality was inadequate.
According to this panel, the real power of Web services is in enabling
new semantic definitions of data held in disparate systems. The ability to
look at your data in combination, to bring it together, allows you to
leverage it in ways not available before. It allows you to create new
functionality centered around data access, data management, and business
process monitoring. In a presentation immediately preceding this panel
discussion, John Stone and JP Morgenthal declared that it won't be mere
exposure of existing functionality but rather a massive demand for semantic
standards that will ultimately drive interoperability.
This view of Web services is penetrating the marketplace. Based on their
Web services management system, Infravio recently deployed a series of
applications at an industrial supplies company. Using Java-based Web
services, Infravio exposed existing business functionality but then wrote
caching code and combined the two to provide real-time access to the
company's back-end SAP system.
In talking to Jim Bole, Infravio's VP of engineering, I learned that for
his client, the true value of this project was not the exposure of existing
functionality but rather the ability to manage the hundreds of services
being exposed in this way. While initially this will be an internal facing
system only, the company intends to publish it to its suppliers as well.
Other values these Web services provide include short time-to-market and a
new semantic architecture that allows them to decouple back-end lock in.
Let's not forget that it was Java that solved the performance problems by
enabling caching. This was a pure Java addition to Web services.
It is this sort of usage that will expand the adoption of Web services
and that adoption will spur the standards.
Back in October 2000, Forrester Research coined a new and rather silly
term: X Internet. There were two presentations that fell under the title
"Financial Application Deployment with Web Services and the X Internet." It
was during the first of these that Michael Baresich, CEO of CoKinetic,
suggested mostly tongue-in-cheek that Forrester came up with this term to
sell more reports. Then he defined it: the X Internet is a sloppy term for a
thin-client technology that adds more functionality on the client side than
is possible with just HTML.
A much better term for this concept is rich client. I have Michael
Curry, Altio's director of product and services, to thank for that phrase.
Such applications bridge the gap between 100% thin-client/HTML solutions and
more traditional client/server applications. In your basic browser-based
application, what you present to your user is HTML. Doesn't matter how fancy
you are on the server side servlets, JSPs, and STRUTs the result is a
static HTML document. At best, this is a compromise. Rich-client
applications offer a more robust interface. Curry gave the second X Internet
presentation and demonstrated that such applications can offer improved
visualization, real-time data (subsecond data updates as opposed to hitting
the "refresh" button), and client-side data manipulation.
There are a number of vendors in this product space and they take a
variety of approaches. CoKinetic has its CoKinetic Player, which is
appropriate for institutional uses. Written in C++, the player is a 2MB
download and has a 5MB footprint. However, once installed, ideally as part
of the corporate desktop image, it's launched transparently when a user
clicks on a link for a CoKinetic application. Synchrony Systems offers
Sizzle, a Swing-based rich-client system. Altio wrote its own Java client
for the client side of its Altio Live suite of products. All three offer a
developer's toolkit for building the user interfaces, and each of these
products accepts XML messages to define and update the UI as well as to
carry the data to the application.
Altio presents a true Java success story. Its entire product is written
in 100% Java. There is the AltioLive Presentation Server that receives
messages from your server-side applications and then sends the output to the
browser where the AltioLive Smart Client sits. The Smart Client has a 200Kb
footprint and handles all the rendering in the browser. Altio chose this
approach rather than a more traditional applet for a variety of reasons.
First, size. That 200Kb application contains all the UI controls, all the
rendering, a DOM and JAX Pack interpreter, as well as the logic for
communicating with the server. By building their own client, Altio is able
to have better browser compatibility as well. Using the 1.1 JVM, AltioLive
Smart Client can work with 4.0 browsers and forward. This way, no one has to
download a plug-in to use an Altio application.
I challenged Curry during his presentation, asking how Altio is guarding against backsliding toward the old fat-client problems. His response was that Altio is dedicated to a very strict Model-View-Controller paradigm and the only functionality they put on
the client side is the View. Granted, that View is richer than you can get
with just HTML, but it is strictly a View. In addition, after the initial
client load, the only data that comes down the pipe is XML, so it's easily
compressed, and the presentation server needs to send only the data.
Now, being a Java shop, even a good Java shop, is not what makes Altio a
success. It's their client base. They've recently installed an
AltioLive-based system at The Hartford, a large insurance company. The
Hartford has replaced the front end of their old claims-handling system with
a rich UI-based system in Altio. The Hartford's primary reason is that they
wanted a thin, Web-based client but they needed more functionality than was
available in HTML. And the development time with Altio's development
environment was much quicker than they would have been able to achieve with
a more traditional JSP-based application. Altio is installing systems for
seven or eight additional clients who all have similar reasons for choosing
Altio better functionality and faster development.
. . .
Conferences present you with a great opportunity to discover what is
happening in the industry. This is why SYS-CON was pleased to see how many
developers, IT managers, and vendors had attended their Web Services Edge
2002 West Conference & Expo, October 13, in San Jose, CA, where as in New
York City Java-based Web services were very prominent.
Steven Berkowitz has done development and project management for Fortune 100 companies, startups, and nonprofit organizations. He recently started techniCrafters to provide Web
development services to small business and Municipalities.