MIME-Version: 1.0
From: Andy Bierman <andy@yumaworks.com>
Date: Tue, 6 Nov 2018 09:31:19 -0800
Message-ID: <CABCOCHSedhq1bZBn6ZCERinC8gTUb+gXuMVvNs8p_SWqPdbyRQ@mail.gmail.com>
To: NetMod WG <netmod@ietf.org>
Content-Type: multipart/alternative; boundary="000000000000ad4e3f057a025ea0"
Archived-At: <https://mailarchive.ietf.org/arch/msg/netmod/MOc-0tqdxpEU1TbT_CYFUSBlIQw>
Subject: [netmod] Issues with draft-ietf-netmod-nmda-diff-00
Precedence: list

--000000000000ad4e3f057a025ea0
Content-Type: text/plain; charset="UTF-8"

Hi,

I understand the WG is going to delay this work 1 or 2 meetings to
see if any proposals for a diff format are proposed.

Any diff command assumes some algorithm that requires:
  (1) copy of the source document
  (2) set of diffs against the source document
  (3) procedure for applying diffs

With this information the tool can produce the target document.

The WG has to decide if this <compare> operation is for
textual representation of a datastore or is it a conceptual compare
of a datastore.  They are not the same.

The first problem that comes up with a textual compare is document order.
There is no defined order for top-level data nodes within a datastore.
There is no defined order for augmenting nodes which are siblings.

The next problem is that some data types have no canonical representation.

A textual compare will be affected by both types of false positives.

So while we are waiting up to 8 months for proposals...
What problem is this new diff format trying to solve?


Andy

--000000000000ad4e3f057a025ea0
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Hi,<div><br></div><div>I understand the WG is going to del=
ay this work 1 or 2 meetings to</div><div>see if any proposals for a diff f=
ormat are proposed.</div><div><br></div><div>Any diff command assumes some =
algorithm that requires:</div><div>=C2=A0 (1) copy of the source document</=
div><div>=C2=A0 (2) set of diffs against the source document</div><div>=C2=
=A0 (3) procedure for applying diffs</div><div><br></div><div>With this inf=
ormation the tool can produce the target document.</div><div><br></div><div=
>The WG has to decide if this &lt;compare&gt; operation is for</div><div>te=
xtual representation of a datastore or is it a conceptual compare</div><div=
>of a datastore.=C2=A0 They are not the same.</div><div><br></div><div>The =
first problem that comes up with a textual compare is document order.</div>=
<div>There is no defined order for top-level data nodes within a datastore.=
</div><div>There is no defined order for augmenting nodes which are sibling=
s.</div><div><br></div><div>The next problem is that some data types have n=
o canonical representation.</div><div><br></div><div>A textual compare will=
 be affected by both types of false positives.</div><div><br></div><div>So =
while we are waiting up to 8 months for proposals...</div><div>What problem=
 is this new diff format trying to solve?</div><div><br></div><div><br></di=
v><div>Andy</div><div><br></div></div>

--000000000000ad4e3f057a025ea0--