SDD again

Tim Bray tbray at textuality.com
Sat May 9 06:36:37 BST 1998


At 06:52 PM 5/8/98 -0700, alex at veosystems.com wrote:
>No, it is not false.  I hightlighted the word 'safe'.  If you absolutely
>*must* know that everything was read and interpreted correctly, you *must*
>use a validating parser.  There are many applications where this is not
>an absolute requirement and, thus, you may use a well-formed parser.

No.  This claim is without technical merit and I cannot let it pass 
unchallenged.  It is trivially possible to achieve correctness and 100% 
unquestioned reliability without the use of a validating processor.  If you 
either (a) construct your DTD so that standalone='true' or (b) don't have 
external markup declarations, then you can have NO DOUBT WHATSOEVER that the 
document is being parsed correctly, for any sane denotation or connotation of 
"correctly".  So please stop making this demonstrably false assertion.

Not only is your premise false in theory, it is vacuous in practice.  If you
think, for any real-world application, that its validation against some DTD
guarantees "correctness" in any nontrivial sense, then I don't want to 
go anywhere near your software.  Validity is a highly specific claim, one
which is of great utility in many applications, but it does not equate to
having "safety" or "correctness".  Equally, lack of validity does not 
equate to lacking "safety" or "correctness".

>In addition, there are some applications that need some level of guarantee
>about whether external declaration subsets will be read and honored.  It is
>this class of applications that we cannot address today with the current
>definition of well-formed.

This statement is correct, except for the unnecessary temporizing about 
"some level of guarantee".  "Guarantee" is a binary condition; if you need 
a guarantee that the organization to which you are sending information
will have external declarations read, then you need to specify the use of 
a validating processor at that end.  If not, then not.

But please don't equate this particular guarantee with general concepts 
of "safety" or "correctness" - doing so gives the impression that the use of 
documents which are merely well-formed is in some way sloppy or irresponsible; 
such a claim is fatuous and very, very, very unhelpful.

 -Tim Bray


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)




More information about the Xml-dev mailing list