<div dir="ltr"><div class="gmail_quote"><div dir="ltr">On Thu, Jun 23, 2016 at 6:34 PM Stephan Oepen <<a href="mailto:oe@ifi.uio.no">oe@ifi.uio.no</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">there is an extension to UDF that allows decorating nodes (which<br>
identify grammar entitites) with their types; look or UDX here:<br>
<br>
<a href="http://moin.delph-in.net/ItsdbDerivations" rel="noreferrer" target="_blank">http://moin.delph-in.net/ItsdbDerivations</a></blockquote><div><br></div><div>Thanks, Stephan. Unless I'm mistaken, UDX only allows for types on the preterminals (i.e. lexical types), where the LOGON restful server puts types on the nonterminal nodes as well. On the other hand, the lexical types are probably more interesting than the rule types, so even getting UDX from ACE would be a nice step. Also, pyDelphin has had support for UDX for a while now, so it is ready to accommodate.</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
i expect the default derivation reader in [incr tsdb()] accepts UDX,<br>
but i would not want the redundant information to go into actual<br>
profiles (they are bulky enough already). so, if you were to go down<br>
this route, i would recommend that ACE by default output UDF when<br>
producing [incr tsdb()] format, and that the additional pieces of UDX<br>
be activated by a separate switch?<br></blockquote><div><span style="line-height:1.5"><br></span></div><div><span style="line-height:1.5">Woodley's recent changes are for an alternative stdout protocol and, I suspect, not the [incr tsdb()] profiles written by the art utility, so I don't think the profiles are being made more bulky.</span></div><div><span style="line-height:1.5"><br></span></div></div></div>