2008/5/6 Philip Van Hoof <<a href="mailto:spam@pvanhoof.be">spam@pvanhoof.be</a>>:<br><div class="gmail_quote"><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class="Ih2E3d"><br>
On Tue, 2008-05-06 at 20:09 +0200, Mikkel Kamstrup Erlandsen wrote:<br>
> 2008/5/6 Philip Van Hoof <<a href="mailto:spam@pvanhoof.be">spam@pvanhoof.be</a>>:<br>
<br>
</div><div class="Ih2E3d">> We can revisit it if we ever want to make API breaking changes (ie<br>
> Xesam Search 2.0).<br>
<br>
</div>If we know how exactly it'll look now, than we can make a remote API in<br>
our own DBus namespace that looks exactly the same.<br>
<br>
Perhaps start a draft of Xesam Search 2.0 already?<br>
<br>
This will make our migration in future much more easy, and it'll give<br>
Xesam a free test platform to verify the usefulness and applicability of<br>
the added remote API.<br>
<div><div></div><div class="Wj3C7c"><br>
</div></div></blockquote><div><br>Well you can always just start hashing out the stateless-xesam-search-spec on the wiki and see what people say. It will not be part of Xesam 1.0, but if we put it in another dbus interface it could go in 1.1 if there is community momentum (more than just Tracker). However I think it is more important to get started on the Metadata Storage API, Index API (term count etc), and Shared Metadata Extraction and Harvesting Spec.<br>
<br>In the very beginning we had two interfaces, a simple (stateless) and a full live interface. Via a lot of discussions and checking of actual client needs it became clear that a stateless interface would have very few customers - hence we went for the stateful API.<br>
</div></div><br><br>Cheers,<br>Mikkel<br>