Service-Oriented Computing: a new programming paradigm?
The point of this post is to introduce a paper that explains service-oriented computing that can be read by anyone, without requiring too much previous knowledge. The link is in the last paragraph of the post. If you were interested in knowing more about what service-oriented computing is, it could be a good starting point.
The concepts of online service and distributed programming exist since quite some time now, to the point that they evolved into a huge research topic both for academia and the industry. Huge as it is, a lot of different results from different research areas appeared; conferences on Service-Oriented Computing (SOC) were opened; a lot of investments were made.
In the last decade or so it became apparent that new foundations were need in order to build truly inter-operable distributed applications and that ad-hoc technologies were not good enough for the job. Moreover, it was clear how much SOC could benefit from theoretical research for handling its complex inherent problems. New concepts were forged and new languages were made, the most notable result being the set of languages, specifications and technologies that falls under the name of "Web Services".
Today, we reached the point in which we can look back and see that "services" follow common principles (e.g. public interface, protocol descriptions...) and must necessarily address certain problems (e.g. session handling, structured data description and manipulation, interoperability...). The whole thing grew up so much in detail that we may be looking at a new programming paradigm: the service-oriented one. A paradigm that has the concept of "service" as a native element, much like object-oriented programming has the native concept of "object" or functional programming of "function".
This is not something that we just imagined. We see this every day at italianaSoftware, programming solutions for our customers with the JOLIE language. Having a language that allows you to make lightweight services in a matter of minutes and to exploit local communications between them in order to maintain performance can change your point of view: services are not mere bridges between different applications anymore, they can be used (and reused) as the building blocks of a single application.
We tried to express our definition and experience on this new paradigm on a concept paper at YR-SOC'09, which is publicly available online. Here is the link (on the upper right area of the linked page you can find links for the pdf or postscript versions). It is an introduction to service-oriented computing that considers what has been made so far, so you will find links to a lot of existing works inside it. Some concepts may be pretty new to some readers (after all, it was originally meant for an academic audience and contains references to theoretical works), but we are currently interested in enhancing the paper to make it more readable by the casual developer (probably with less references to theoretical works and more comparisons with existing languages and technologies) and put the new version online in the JOLIE website. So, if you read the paper and find some concepts too obscure or you have some doubts or comments (apart what I have already said) or some arguments that you wish would be addressed, please let us know!
The concepts of online service and distributed programming exist since quite some time now, to the point that they evolved into a huge research topic both for academia and the industry. Huge as it is, a lot of different results from different research areas appeared; conferences on Service-Oriented Computing (SOC) were opened; a lot of investments were made.
In the last decade or so it became apparent that new foundations were need in order to build truly inter-operable distributed applications and that ad-hoc technologies were not good enough for the job. Moreover, it was clear how much SOC could benefit from theoretical research for handling its complex inherent problems. New concepts were forged and new languages were made, the most notable result being the set of languages, specifications and technologies that falls under the name of "Web Services".
Today, we reached the point in which we can look back and see that "services" follow common principles (e.g. public interface, protocol descriptions...) and must necessarily address certain problems (e.g. session handling, structured data description and manipulation, interoperability...). The whole thing grew up so much in detail that we may be looking at a new programming paradigm: the service-oriented one. A paradigm that has the concept of "service" as a native element, much like object-oriented programming has the native concept of "object" or functional programming of "function".
This is not something that we just imagined. We see this every day at italianaSoftware, programming solutions for our customers with the JOLIE language. Having a language that allows you to make lightweight services in a matter of minutes and to exploit local communications between them in order to maintain performance can change your point of view: services are not mere bridges between different applications anymore, they can be used (and reused) as the building blocks of a single application.
We tried to express our definition and experience on this new paradigm on a concept paper at YR-SOC'09, which is publicly available online. Here is the link (on the upper right area of the linked page you can find links for the pdf or postscript versions). It is an introduction to service-oriented computing that considers what has been made so far, so you will find links to a lot of existing works inside it. Some concepts may be pretty new to some readers (after all, it was originally meant for an academic audience and contains references to theoretical works), but we are currently interested in enhancing the paper to make it more readable by the casual developer (probably with less references to theoretical works and more comparisons with existing languages and technologies) and put the new version online in the JOLIE website. So, if you read the paper and find some concepts too obscure or you have some doubts or comments (apart what I have already said) or some arguments that you wish would be addressed, please let us know!