RE: [xml-dev] Iceberg development (was Re: [xml-dev] maps)

["Mark Seaborne" <MSeaborne@origoservices.com>]

>-----Original Message-----
>From: John F Schlesinger [mailto:jschlesinger@computer.org]
>Sent: 08 August 2002 16:26
>To: Mark Seaborne
>Cc: Jeff Lowery; Xml-Dev (E-mail)
>Subject: Re: [xml-dev] Iceberg development (was Re: [xml-dev] maps)


>Mark Seaborne wrote:

>>I suppose I am wondering not so much whether formal and precise agreement is necessary, so much as when things get 
>>validated. 
>>
>I think you have to validate as soon as something comes into your domain 
>for the first time.

Well, presumably an XML parser checks for well-formedness before moving on to attempt to establish validity. 

I am used to having my own validation/typing mechanisms that work on my own, internal physical structure, and obviously these cannot work until the physical structure used to express a message has been verified and stripped away. So I am happy for schemas just to handle structure, because I don't really care much about what they have to say about anything else, at least with regards to validation (as opposed to filling in gaps).

One useful by-product of this approach is that message transformation becomes much simpler. Data and logical structure are always transformed in my application, where I am in complete control of physical representation. Once this has been done, the message can be output in an appropriate physical structure, without me having to worry, or even know, what the original physical structure was. Separation makes mappings simpler and more manageable, especially if one has to deal with what is essentially the same logical model, expressed in many different physical formats.

All the best

Mark