[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: more silly questions

• To: "'Amelia A Lewis'" <amyzing@t...>,<rjm@z...>
• Subject: RE: more silly questions
• From: "Michael Kay" <michael.h.kay@n...>
• Date: Mon, 9 Feb 2004 09:04:34 -0000
• Cc: <xml-dev@l...>
• Importance: Normal
• In-reply-to: <20040208202109.512d4c8a.amyzing@t...>
• Reply-to: <michael.h.kay@n...>

> ASCII or UTF-8, without the strong well-formedness 
> constraints of XML, simply *are* more robust than XML for 
> things like headers.
>

There seem to be two mutually-contradictory uses of the word "robust".
One is "process everything as best you can, even it it's garbage". The
other is "don't accept anything that's garbage".

Apart from the problem (which was deemed off-topic) that the SMTP
protocols support a flawed process model, they also have the problem
that they don't reliably transmit the data from sender to recipient. We
still see cases of XML code examples in email messages being corrupted,
for example, and non-ASCII characters being displayed incorrectly. Not
to mention the unreadability of quoted text with lines wrapped at column
72, and mail clients that remove newlines when they shouldn't. This is
essentially because there isn't enough metadata carried with the message
to format it accurately on receipt, which in turn is because a transfer
syntax is used without sufficient extensibility built in. Robustness
shouldn't mean "guessing correctly most of the time". A truly robust
protocol would be one that carried all the necessary metadata to decode
the message, and XML would be an excellent vehicle for that.

Michael Kay

• References:
  • Re: more silly questions
    • From: Amelia A Lewis <amyzing@t...>

• Prev by Date: Re: A question about REST and transaction isolation
• Next by Date: RE: more silly questions
• Previous by thread: Re: more silly questions
• Next by thread: RE: more silly questions
• Index(es):
  • Date
  • Thread