2007/1/17, Jean-Francois Dockes <<a href="mailto:jean-francois.dockes@wanadoo.fr">jean-francois.dockes@wanadoo.fr</a>>:<div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"> Mikkel Kamstrup Erlandsen writes: > 2007/1/17, Jean-Francois Dockes <<a href="mailto:jean-francois.dockes@wanadoo.fr">jean-francois.dockes@wanadoo.fr</a>>: > > The user-level language needs to have a way to force stemming off. > > I still think it should be an optional thing - as it is now. The language > supports stemming control, but it is not required. As discussed a bit > earlier such extended features should probably be introspectable with a a > getExtendedFeatures() method (both in simple and live api). Not sure we understand each other here. I was referring to the simple search language from <a href="http://wiki.freedesktop.org/wiki/WasabiDraft"> http://wiki.freedesktop.org/wiki/WasabiDraft</a>, and I meant that there should be a way for the user to express her wish for an exact match, with no transformation of the search term. The backend will then do its best. I don't see how the *language* feature could be optional, not having it would unnecessarily cripple the capabilities of engines which can switch stemming on or off. Maybe we could simply say that terms enclosed in double-quotes are not to be stemmed if possible ?</blockquote> <div> Ah, sorry. I was still thinking about the xml language, I see now that you wrote *user-level*  language :-) Regarding the user-level language now. Perhaps "flying Dutchman" would mean the unstemmed phrase (if supported), and 'flying Dutchman' (single quotes) could allow for stemming. This is not completely standard, but it doesn't break user expectation in horrible ways (as far as I can see). </div> <blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"> > Do we all agree that we need to define a user level language as well? +1 > from me at least. Yes for me too. > > I would be much in favour of using the raw user entry here. > > I would be much in favor of allowing userQuery elements embedded in the xml. > It does introduce a bit of extra coding work for servers... I think that I'd prefer to keep xml out of the simple API, but I could also work with your approach.</blockquote><div> We have a tie :-) What do you other guys think? I see two arguments for the xml language in both simple and live:  - Common api between live and simple interfaces  - Both apis can have the powers of both the user-level and the full xml languages, while still having an easy way of doing direct user queries without much parsing. Cheers, Mikkel </div> </div>