|
[XMLSCHEMA-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message] Re: A bug or incorrect logic?
From: Arshad Noor <arshad.noor@strongauth.com>
Date: Fri, 31 Oct 2008 11:10:35 -0700 Message-ID: <490B4A1B.7020003@strongauth.com> To: Michael Kay <mike@saxonica.com> CC: xmlschema-dev@w3.org Point taken, Michael. Its a reasonable assumption to make when there are many choices available to a developer of a finite state machine. Thank you (and Michael Glavassevich) for the answers. Arshad Noor StrongAuth, Inc. Michael Kay wrote: >> Given that A is optional, shouldn't the error message >> indicate that (at a minimum) "B or C" is expected? >> > > Designing good error messages tends to involve guessing what the user's > intention was. The message that this particular processor has given you is > what comes out easily from the usual strategy of compiling the grammar into > a finite state machine, where in any given state you know what symbols are > allowed to appear to move you to the next state. > > There are strategies that do better (spelling checkers use them, essentially > by trying to work out what valid construct your invalid construct is closest > to) but they tend to involve look-ahead, which means they don't scale well > when the content is very long. And as we know from spelling checkers, they > can sometimes produce hilarious results. > > Michael Kay > http://www.saxonica.com/ >Received on Friday, 31 October 2008 18:11:11 GMT |
Subscribe in XML format
|






