xml-dev - Re: updategrams

[james anderson <james.anderson@mecomnet.de>]

Michael Rys wrote:
> 
> Dear James
> 
> > 2. anyone can give a rundown on why they went for a diff technique
> > rather than a direct pat specification with a replacement value?
> 
> The scenarios that we tried to address were mid-tier caches that want to
> synchronize with the database. Direct pat[h] specifications with replacement
> values would not fit that well into these scenarios and would go more
> towards an update language that would require a query language.

that sounds more reasonable. i suppose i'd have to have read the
scenariour to know how these caches were expected to behave, but i'm
still surprised that one would want to work out a yet another addressing mechanism.

>                                                                 While XPath
> could cover some aspects, we did not want to invent something in that space
> before XML Query and its relationship to update languages is clear.

well, sometimes it's hard not to.

> 
> Please note that people can design their own XML update language in
> conjunction with the OpenXML XML rowset provider inside T-SQL stored
> procedures.
> 

i was wondering 'cause i had done an update mechanism against an object
database and found (for the large object case which im mentioned in the
previous message) ended up using path-based adressing...