I am not against CORBA or Object-orientation

Mark Little and I exchanged comments, which you may find interesting, on my last entry. We continued the exchange over instant messaging. Mark's understanding of distributed computing technologies is great. My discussion with him made me realise that some of my views may have been misunderstood. So, please allow me to clarify the following...

  • I am not against object-orientation as an architectural paradigm; I like O-O-based environments as development platforms. However, I try to make the separation between architectural concepts and implementation-related technologies.
  • I don't think that CORBA or any other existing distributed computing technology (object-oriented or not) will disappear overnight just because of Web Services technologies, which I never suggested were revolutionary. I advocate for service-orientation because I believe in the possibilities it offers. The industry will indeed have to find ways of building on its current investments and will not switch overnight to the new platform. Hence, technologies that bridge the gap will be important.
  • When I talk about technologies that attempt to facilitate the transition from the previous generation of middleware to the new order of things, I am not suggesting that they are unnecessary or that they are completely wrong but I believe that some (not everyone) will probably misuse them.

I hope that these clarify my position a bit.

8 responses to “I am not against CORBA or Object-orientation”

  1. Mark Little
    Some people will always misuse technology and unfortunately it’s in some vendor’s interests to encourage this. Vendor lockin is often a result (closely coupled systems). However, it doesn’t have to be the 80/20 case, so I don’t believe we should be preventing potentially useful technology like IDL-to-WSDL from being defined and deployed just because there’s a possibility some people will misuse it. If that were the case I don’t think we’d ever have made such progress with distributed computing in general, or Web services specifically (can you say WS-RF?)
  2. Mark, just to make sure i understand your last sentence… Are you suggesting that WS-RF is a good technology that is misused or that WS-RF is built on good technology but misuses that technology?
  3. Mark Little
    The latter.
  4. Mark Little
    Savas, the following sentence “I dont think that CORBA or any other existing distributed computing technology (object-oriented or not) will disappear overnight just because of Web Services technologies” implies that you think they will vanish eventually. If that is the case then I contend it’s wrong. SOA (ignoring the WS specific implementation) is just another tool in a good architect/programmer’s toolbelt, but IMO it’ll never be the only tool. Technologies *like* CORBA and DCE will continue to be used in environments were it makes sense to do so.
  5. No argument there Mark. The scope of my comments is internet-scale computing. Do you see CORBA and DCE deployed around the internet? Not behind service boundaries and inside service implementations but in cross-organisation scenarios?
  6. Mark Little
    I know of some used of CORBA/DCE that span multiple corporate intranets and for good reasons: they are more mature technologies than Web services, particularly in the area of security and messaging (strangely enough). However, certainly in some of the examples I know, to get agreement to do this kind of inter-organisation communication was difficult (politically more than technologically). Whether those organisations would do it the same way today is a different matter. I think that the answer now (i.e., 12:15 pm Amsterdam time) is probably they would because Web services is still evolving. But if we could fast forward a couple of years, I reckon it would be different.
  7. Which is exactly what I am talking about. I am talking about the future and not what is going on today. I am talking about what we’d be doing in the years to come and not how people are building system now.
  8. Mark Little
    Savas, as with most things I don’t think there’s ever going to be a single right solution, which I always get from you that there is 🙂 I certainly believe that given a couple of years, the WS approach to b2b intergration (for want of a better term) will be the 80/20 choice (probably even more). But there will always be other solutions, including CORBA, DCE or roll-your-own, that will be considered and used. Maybe we can agree to agree (!) on the toolbelt analogy: I have no idea which came first, but I’m sure years ago when cross-head screws came onto the market, the makers of flat-head screws had the same opinion – “flat-heads are the best”; as I’m also sure there were arguments that went “cross-heads are better; why would anyone want to keep using flat-heads” 😉 But as we know, they both co-exist happily in the same environment. (You can see that I’ve been doing a lot of DIY recently, and as a slight aside, I’d love to know why there are two types of screw!)