DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma;
 h=mime-version:date:message-id:subject:from:to:content-type;
 b=av0q1eCWLQm0YPZ0ySKy5B3H8N8kouFR8Jy9VsqpJYR4Tt+JtNrEqtN/DmfxGtJrZT
 +gidrVYOnPOCdV43uzqSzwGX7KqZAM62RqEasmAeXws5Qb/Cpnls3ctTUUTX8tDAwG0b
 B3kn7/AteZ0q1ACBkTUOm9g2/LAZi2c8GHE5c=
MIME-Version: 1.0
Date: Tue, 24 Feb 2009 19:33:39 +0100
Message-ID: <ebe4d1860902241033o7e742740wc7d438d92921fe19@mail.gmail.com>
From: Catherine Pfeffer <cathypfeffer@gmail.com>
To: mmox@ietf.org
Content-Type: multipart/alternative; boundary=0015174c436c9ebb000463ae5b74
Subject: [mmox] Top down or down top?
Precedence: list

--0015174c436c9ebb000463ae5b74
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

Christian wrote:
> Well, I know a lot of places where it's used to add additional
> attributes etc.
(it = namespaces)

Additional attributes from the same standard?

> Maybe some XML representation is really all which is needed.

I'm glad we said it in parallel without any concertation ;-).

> But actually I really would first think about a more global approach
> towards a general architecture and then defined the more details
> protocols and formats needed.

What we need is perharps both a top-down approach and a down-top approach.

For example, what happens if you write all the clothing serialization in
LLIDL and then Jon says "well, data description would be better in XML
schemas" ? ;-)

In the other direction, the high level analysis might discover that we need
to add a "time" data type to the lower level.

I feel it safer to think both at highest layers and at lowest layers at the
same time. The actual work needs not be done by the same people. If they
work in different groups, there just needs to be enough communication
between both groups so the whole picture remains consistent. The whole
approach can be an incremental process.


-- 
Cathy

--0015174c436c9ebb000463ae5b74
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Christian wrote:<br>&gt; Well, I know a lot of places where it&#39;s used t=
o add additional<br>&gt; attributes etc.<br>(it =3D namespaces)<br><br>Addi=
tional attributes from the same standard?<br><br>&gt; Maybe some XML repres=
entation is really all which is needed.<br>
<br>I&#39;m glad we said it in parallel without any concertation ;-).<br><b=
r>&gt; But actually I really would first think about a more global approach=
<br>&gt; towards a general architecture and then defined the more details<b=
r>
&gt; protocols and formats needed.<br><br>What we need is perharps both a t=
op-down approach and a down-top approach.<br><br>For example, what happens =
if you write all the clothing serialization in LLIDL and then Jon says &quo=
t;well, data description would be better in XML schemas&quot; ? ;-)<br clea=
r=3D"all">
<br>In the other direction, the high level analysis might discover that we =
need to add a &quot;time&quot; data type to the lower level.<br><br>I feel =
it safer to think both at highest layers and at lowest layers at the same t=
ime. The actual work needs not be done by the same people. If they work in =
different groups, there just needs to be enough communication between both =
groups so the whole picture remains consistent. The whole approach can be a=
n incremental process.<br>
<br><br>-- <br>Cathy<br>

--0015174c436c9ebb000463ae5b74--