simple search api (was Re: mimetype standardisation by testsets)

[Jos van den Oever]

2006/12/3, Mikkel Kamstrup Erlandsen <mikkel.kamstrup at gmail.com>:
> 2006/12/3, Jos van den Oever <jvdoever at gmail.com>:
>
> > 2006/12/3, Joe Shaw <joeshaw at novell.com>:
> > > Mikkel Kamstrup Erlandsen wrote:
> > > > Just a quick idea,
> > > >
> > > > How about using JSON ( www.json.org <http://www.json.org>) as to
> > > > represent the query objects (instead of xml)?
> > >
> > > I'm reluctant to do this because I'm not sure how comprehensive the
> > > stacks are for C#.  The XML stacks we have -- the built in one in Mono,
> > > or libxml2 for C -- are very mature parts of the platforms.
> > >
> > > > It is light, easily readable, widespread, and there is possibility to
> > > > write extremely fast parsers.
> > >
> > > Maybe, but we should be focusing our efforts on building the search
> > > software, not on
> hot-serialization-technology-of-the-day parsers.
> >
> > I agree. Lets just say that the simple interface cannot have nested
> > queries, in other works, no brackets. Then it is possible to use dbus
> > for specifying the query object.
>
>
> I'm not sure I follow you. What do you mean by "use dbus for specifying the
> query object" ?

if we dont allow for nested queries, a query can be defined as a(iss)
which is a list of query arguments.
i -> query operator: +, -, <, or >
s -> name of the field to search in
s -> value to look for

Then we dont need a query language or a structured string format and
client an server code is much easier to write.

For the simple query this is very desirable.

Cheers,
Jos