[apps-discuss] Schema comments on draft-ietf-mile-sci (fwd)

Claudio Allocchio <Claudio.Allocchio@garr.it> Wed, 22 January 2014 09:34 UTC

Date: Wed, 22 Jan 2014 10:34:05 +0100
From: Claudio Allocchio <Claudio.Allocchio@garr.it>
To: apps-discuss@ietf.org
Message-ID: <alpine.OSX.2.02.1401221033270.38263@mac-allocchio3.garrtest.units.it>
User-Agent: Alpine 2.02 (OSX 1266 2009-07-14)
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset="US-ASCII"; format="flowed"
Subject: [apps-discuss] Schema comments on draft-ietf-mile-sci (fwd)
Precedence: list

---------- Forwarded message ----------
Date: Wed, 22 Jan 2014 07:23:17 +0100
From: Martin Thomson <martin.thomson@gmail.com>
To: draft-ietf-mile-sci.authors@tools.ietf.org,
     draft-ietf-mile-sci.chairs@tools.ietf.org,
     Claudio Allocchio <Claudio.Allocchio@garr.it>
Subject: Schema comments on draft-ietf-mile-sci

I did take a quick look at the schema.  This is a fairly big, yet simple 
schema.

At this stage, the only comments I'd be prepared to advance :

There is no need to wrap <xs:choice> with <xs:sequence> when both have
cardinality 1.  That's redundant.

schemaLocation on the import is unnecessary (and not especially useful
given that it's a URN).

use of xsd:string throughout is an anti-pattern.  xsd:token is better,
even more so if you can restrict the value-space further;
unconstrained strings can be a real burden later.  In particular,
xsd:string causes usability issues, and can lead to some major issues
with whitespace normalization.  Unless you need to carefully preserve
whitespace, I'd strongly recommend against its use.

[apps-discuss] Schema comments on draft-ietf-mile… Claudio Allocchio