loose coupling, was Re: [xml-dev] Early Draft Review: XQuery for Java (JSR 225)

Michael Kay wrote:

>>The issue is relevant to ad hoc queries but it's moot for 
>>repetitive queries,
>>such as for generating a monthly publication. 
>>    
>>
>
>That statement confuses me, and not only because Americans appear to use the
>word "moot" in the opposite of its original English sense of "arguable".
>  
>
I wonder whether to which extent the interpretation of words like moot 
influences the quality of API designs. I always though argueing (with 
the goal of agreement) was a good thing, but by now I come to think that 
I might be too idealist.

>The benefit of a string interface is that it gives looser coupling between
>systems. The benefit of a custom syntax/protocol is that it gives earlier
>validation. The experience of the last few years is that loose coupling
>tends to win, across most application scenarios. Hence XML.
>
>  
>
I am not sure of the "loose coupling" argument in the context of 
large-scale software development.

You put element names, variables names in your query strings. You might 
rely on the schema of the data to process the query result.

Now the schema evolves, you have to change your queries - is it easier 
to change (=>loosely coupled) if you have your queries in objects, or if 
you have them in strings ?

cheers,
Burak