XML Signature WG

99-September-99
Chairs: Donald Eastlake and Joseph Reagle
Note Taker: Joseph Reagle [text]

Participants

• Donald Eastlake 3rd, IBM
• Joseph Reagle, W3C
• Ed Simon , Entrust Technologies Inc.
• Todd Vincent, GSU
• Peter Norman, FactPoint,
• Mark Bartel, JetForm
• John Boyer, UWI
• Richard Brown, Globeset

Minutes

Requirements

• Brown
  • 2.3.1&2 why are the sub-headings? Reagle: groups them, but not necessarily.
  • 2.3.* Formal statements are somewhat confusing. Reagle: correct, will fix.
  • 3.2.2 A capability, not a requirement over all signature. Reagle: true.
  • ACTION DON: will reword and send to list.
  • ACTION BROWN: send comments to list.
  • ACTION REAGLE: tweak final time and move forward.

Syntax Draft

Capitalization

• Capitalize all words and joined words.
• Peter Norman, does RDF reserve the first letter as capitalized for conventions.
• ACTION REAGLE: Check with everything capitalized, bounce of Ralph

Closure

• Peter wants to add something to the native document where he doesn't have control over it nor the DTD, needs to use a <?PI> to define the scope and ensure it will always be ignored by Signature applications.
  • Reagle: Let's not speak of PIs because of the property that they were ignored by a particular c14n algorithm. (That algorithm has now changed and won't ignore them.) It does make sense to speak of <PIs> if you need to arbitrarily insert some content irregardless of the content model.
• Scenario: a document with three paragraphs, assume you want to sign the second paragraph.

  <root>
    <p>this is a paragraph </p>
    <p>this is a longer paragraph</p>
    <p>this is the longest paragraph</p>
  </root>

  Can use XPath ' /child::para[position()=2] '

  Now if someone inserts text resulting in a new document

  <root>
    <p>this is a paragraph </p>
    <p>new paragraph this is </p>
    <p>this is a longer paragraph</p>
    <p>this is the longest paragraph</p>
  </root>

  The signature would break. Is this a good thing or bad thing? If you permit something like:

  <root>
    <p>this is a paragraph </p>
  <?dsig type="begin" id="1">
    <p>this is a longer paragraph</p>
  <?dsig type="end"  id="1">
    <p>this is the longest paragraph</p>
  </root>

  and use

  /descendant-or-self::node()
[
ancestor-or-self::node()/previous-sibling::processing-instruction([@type="begin"][@id="1"])
  and
ancestor-or-self::node()/following-sibling::processing-instruction(@type="end"][@id="1"]
]

  Your signature might be less likely to break.

  The breaking or keeping of the signature is a good/bad thing as defined by the application. Boyer's closure requirement means that applications have the expressitivity/power to define whether its a good or bad thing.

• Am I signing a section or a transformed document? Good to think of this as a transformed document as Tim said, "When we talk about signing parts of a document, then they only way I can see of giving meaning to this is to say that we are signing a some document which is not actually given, but is formed by making a particular transformation on the document given.... Life is then simplified. A signature is over a document. The document can be referred to, and/or enclosed, directly, or specified as a manipulation function. So long as both parties know how to do it, any function can be used. This puts the (xpath, say) function into a very similar position to the canonicalization function."
  
• Don and Joseph will try to structure discussion with a series of poll/questions if necessary to come to "closure" on this issue if necessary. Otherwise continue on list.