- From: David Dorward <david@dorward.me.uk>
- Date: Sat, 18 Jun 2005 18:17:50 +0100
- To: Felix Miata <mrmazda@ij.net>
- Cc: www-validator@w3.org
On Sat, Jun 18, 2005 at 12:54:39PM -0400, Felix Miata wrote: > You people have a bizarre definition of where an error occurs. Line 40, > standing alone, was perfectly valid. No it wasn't. Validity can only be taken in the context of the document. Line 40 could form *part* of a valid document. > Therefore by any normal logic, any error recognized at line 40 > either occurred previously, or is a consequence of an error spanning > multiple lines that include line 40. Your normal logic can only be applied with the benefit of an understanding of what the characters actually mean. The validator isn't an artificial intelligence, it can't have that level of understanding. It can't tell when you've made a mistake that doesn't break the rules (such as failing to include the end quote for an attribute value), only when a rule is actually broken (such as missing that = sign from after what (according to the rules) is an attribute name (even if you ended that string of characters to be the start of an attribute value, but your opening quote actually closed the earlier quote you missed the quote mark from). > The validator has no problem pointing back to an opening tag on a > previous line when a required end tag is missing That is a special case, it has to know about the opening tag in order to know that the closing tag is missing. > so why can't it point out or at least allude to some mistake on a > previous line? It tells you where the error actually occurs. It gives you the line number and character position so you can find it in the full document. To help out, it gives you a small amount of the code from around the error. It can't know what you intended to type, only where the error actually is. -- David Dorward http://dorward.me.uk
Received on Saturday, 18 June 2005 17:17:55 UTC