Re: [service-orientated-architecture] Re: Legacy into SOA (was Vandersluis on a Data Abstraction Layer's Benefits)
- Michael Poulin wrote:If you mean Web Service when talking about service, I understand what you said.I definitely didn't mean to limit my statement to Web Services. The plethora of Web Service standards is just one, not particularly elegant, way of implementing services. I was speaking in terms of the general concept of a service, not one implementation technology.Nevertheless, I stay with my statement. I tried to explane this in one of my latest posts.
Personally, I am in favor of aggregated services rather than composite applications.Is the difference between those two deeper than mere semantics?If the functionality provided by the applications is wrapped in one or more services, the two become synonymous. In many cases, this is the best practice.Regards,Patrick
- How do you know what to want?
You have a repository of services with its descriptions - functionality, RWE and, BTW, interfaces (possibly, more than one per service). You need a service which fulfill your task. Services are announced by somebody you do not know. Even if you read the descriptions, how do you know that the publisher meant the same thing unfer particular word as you do? Even if you use a word-matching searchers like Lucene or Google's indexing, maximum you can get is a match of the word (pray if it is in English and not in other languages that modify the words depending on the context...), not its meaning.
Jini is nice for programmers, when the consumer and provider are in the same culture and language dialect. This is oversimplified world, unfortunately (with all respect to Jini).