You’re probably surprised that I’ve kept this up, aren’t you. Here we are, five interviews into this series and still going strong. This month, we chat with the one Flanders that Homer Simpson actually appreciates: Jon Flanders. Jon is a blogger, MVP, thought leader in the SOA space, and is comfortable wearing a skirt. Jon has recently released his book RESTful .NET to critical acclaim and has taken a break from his whirlwind book tour (and the thousands of screaming ladies) to engage in a little Q&A with us.
Q: Tell us why a developer who has always built SOAP-based web services should care about REST. Why is it worth it to them to learn a different paradigm and what benefit does this paradigm offer to enterprise services that typically are built in a SOAP/RPC fashion?
A: What I typically tell people here is two things.
1) REST has some significant advantages over traditional RPC styles (which most SOAP-based services are). GET results can be cached, REST services are *more* interoperable than SOAP and WS-*, and the statelessness constraint encourages more scalable implementations, and the uniform interface (GET, POST, PUT, DELETE) make building and using services much simpler than custom APIs (which SOAP-based services are because each one is a custom interface). If you use all of the constraints of REST (specifically the hypermedia constraint), you also get a highly decoupled implementation.
2) Because of these advantages, most of the non-Microsoft parts of the computer industry have moved towards a RESTful approach already, and Microsoft is currently moving that way. When you look at ADO.NET Data Services, Windows Azure, you see a lot of Microsoft’s effort going into building RESTful services. Because of this, even if you aren’t planning on implementing all your services using REST, you probably will be consuming one or more RESTful services in the near future.
In the end, I don’t advocate moving away from SOAP/WS-* where it makes sense or is necessary (for things like transactional calls between .NET and Java for example), but I think more services than people think could benefit from using a RESTful approach.
Q: Outside of the plethora of WCF related things you inevitably learned during the writing of your latest book, what more general “service design” concepts/principles/pitfalls have you picked up as a result of authoring this book?
A: Nothing really new. The concept/principle I believe in most is Keep it Simple Stupid (KISS).
Q: In addition to being an author, blogger, instructor, and part-time samurai, you also do consulting work. Tell us about the most complicated BizTalk Server project you ever worked on and how you solved the business problem.
A: Honestly, I’ve never been involved in a “real” BizTalk Server project (what do they say “those who can’t teach” ;-)). I have built a number of fairly complex demos for Microsoft using BizTalk, probably the most complicated demo involved using BizTalk Server with BizTalk Services (now .NET Services).
Q [stupid question]: You regularly make the rounds in the conference circuit and naturally meet folks who only know you by your online presence. What’s the oddest thing someone has remarked to you upon meeting you in person for the first time? For me, on multiple occasions, I got a “oh, I thought you were taller.” Apparently I have the writing style of a 7 footer.
A: Where’s the kilt?
Hope you all are enjoying this series, and if you have interest in being added to my “interview queue”, do let me know.
Richard Seroter's Architecture Musings 