Abstract:
This panel discussion will focus on the synergies and tensions
between open source and the standards process. Why and when should
open-source developers get involved in the standards development? How
do standards-developing organizations such as the Java Community
Process (JCP) need to change in order to accommodate individuals and
open-source groups? What lessons can each group learn from the other?
The participants in this discussion have direct and practical
experience in the open standards and the open source worlds.
Join us for a stimulating discussion, and bring your questions and comments.
Abstract:
The Web is fast approaching its 20th anniversary. However, although REST principles have certainly proven themselves, there have always been its detractors as far as "enterprise" scale distributed deployments are concerned.
The perceived lack of security, transactions, policy definitions etc., plus issues like the broken-link problem, made it easy for people to persuade themselves that the Web is only an infrastructure for documents. When Web Services came along they were developed for two principle reasons:
Internet-scale computing and interoperability of heterogeneous implementations.
WS-* has been designed from the start with capabilities such as transactions and security in mind and there has been a big push on standardisation throughout the industry (very important if you want to achieve interoperability). With the exception of SOAP/HTTP, WS-* ignores REST and owes much of its architecture to distributed systems such as CORBA and JEE.
Over recent years the REST versus WS-* debates have raged as advocates from both camps paint a black-or-white picture of systems development using only one or the other approach. As far as the vocal minority are concerned, there can be only one choice. However, there are important things that both sides can learn from one another, as well as from work that occurred before the Web was developed. We have seen some of this happen already, for example with the work on WS-Context or earlier work on REST-transactions from Hewlett Packard. Rarely does one size fit all in the world of computer science.
In this presentation we'll try to shed some light on the discussion and show that both REST and WS-* have their roles to play in any good architects toolkit.
We'll also look at where possible convergence could (should?) occur within the industry. If 2007 was the year of REST OR WS-*, maybe 2008 is the year of REST AND WS-*?