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

[Jean-Francois Dockes]

Mikkel Kamstrup Erlandsen writes:
 > Let me gauge the current vibes to try an focus on where we are heading.

Hey, I have given no recent vibes. Let me vibe a little. I'm still much
with your original project.

 > ?1) There is general consensus on forgetting about
 > org.freedesktop.search.simple as it stands and focus on an interface with
 > toolkit bindings only/mainly?
 > ?2) The query as passed over dbus will be XML.

The simple interface that you first described is very valuable for naive
apps in my opinion. It's so simple that it probably doesn't warrant its own
binary lib, but sample code for each important platform would be a good
idea.

What's important is that we define well the semantics of the search, the
exact nature and definition of the different elements: what attributes, in
what format, what is an uri, a date, a file reference, etc.. If all these
are well defined, they can be reused in a different interface. The simple
interface being, well, simple, should not be a big additional coding burden
to have on a server that will also supports the other, more sophisticated
one.

 > ?3) We don't know about any known XML Query language suitable for text index
 > queries.

Yes, we'd have to roll our own, which shouldn't be that problematic. But
let's roll the simple one first, to validate the data elements and produce
a toy implementation to play with before going sophisticated.

 > ?4) Any management service/library (and related specs) can (and will) be
 > punted for now.

No opinion.

 > ?5) Given ?1, the api/framework should be easily extensible to accommodate a
 > metadata service  framework.

Yes given 1-2, but I'd do this in a second phase. 

You really convinced me that the trivial approach has its uses. I'm not at
all convinced by the performance arguments. This *is* going to be less
responsive than other approaches might be, but I don't think that this
would be prohibitive in many usage scenarios. We're mainly talking about
personal workstations, session bus, and personal search process here right?
So if the user wants to DoS his own search service, in the words of Joe,
I can't see that this is a major issue. Bad apps will be weeded out.

Regards,
JF