W3C home > Mailing lists > Public > www-voice@w3.org > April to June 2004

RE: Comment on SRGS PR 18 December 2003 w.r.t ambiguity

From: Serge LE HUITOUZE <slehuitouze@telisma.com>
Date: Thu, 1 Apr 2004 10:52:28 +0200
Message-ID: <3AB3C7F30279254494C258DE38FE7CB969F5DD@cedre>
To: <www-voice@w3.org>

Guillaume Berche wrote:
> In the following example with an ambiguous grammar, [...]
>
> Expansion:  t0 (t1 {tag1}) <0-2> t1 t2
> Input: t0,t1,t1,t2
> Output 1: ["t0",{!{tag1}!},"t1","t1","t2" ]
> Output 2: [REJECT after evaluating: "t0",{!{tag1}!},"t1",{!{tag1}!},"t1"]

First (cosmetic) remark: Output 1 is not the expected output, according
to Appendix H of SRGS Proposed Recommendation from 18 December 2003.
It should be: ["t0","t1",{!{tag1}!},"t1","t2" ]

Second, more important, remark: This grammar is *not* ambiguous.
The only three inputs that are recognized by this grammar are:
1. t0, t1, t2
2. t0, t1, t1, t2
3. t0, t1, t1, t1, t2

Each of them are *unambiguously* parsed, giving respective unique outputs:
1. ["t0","t1","t2" ]
2. ["t0","t1",{!{tag1}!},"t1","t2" ]
3. ["t0","t1",{!{tag1}!},"t1",{!{tag1}!},"t1","t2" ]


Best regards.

--Serge Le Huitouze
Received on Thursday, 1 April 2004 03:55:53 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 30 October 2006 12:48:59 GMT