W3C home > Mailing lists > Public > www-forms@w3.org > December 2001

Need to re-work Chapter 2 substantively?

From: T. V. Raman <tvraman@almaden.ibm.com>
Date: Mon, 17 Dec 2001 11:23:34 -0800
Message-ID: <15390.17974.131471.201352@bubbles.almaden.ibm.com>
To: AndrewWatt2001@aol.com
Cc: www-forms@w3.org, xforms@yahoogroups.com
First off --this is an 
excellent observation --and I'll first fix the entries in
the Purpose column 
to actually match "purpose".

The reason you see the three-way separation --Purpose,
Presentation and Data--
is that we conceptually think of XForms having three pieces,
components, stages ...

XForms --the model and ui together --encapsulate the
"purpose " of the application--

the model encapsulating the data structures and constraints,
the UI encapsulating a high-level representation of the
interaction.

Presentation --the XForms markup embedded in a host document
e.g. XHTML

When done interacting, the collected "data" --the XML
instance we populated --is shipped back --submitted--

Dropping "Data" in the three-way separation loses the final
stage which is just as vital.

Incidentally, this is also a response to an earlier comment
from  Chris  pointing out that the "XML makes things
internationalized"  claim was sweeping.

What we meant is that because XForms submits the data as an
XML instance, the submitted data  is internationalized
without us having to do anything special in the specification.
>>>>> "AndrewWatt2001" == AndrewWatt2001  <AndrewWatt2001@aol.com> writes:

    AndrewWatt2001> I guess since the XForms WD is almost at
    AndrewWatt2001> Last Call WD it is probably time to
    AndrewWatt2001> raise an issue which has been niggling
    AndrewWatt2001> me for several versions of the spec.

    AndrewWatt2001> I think the division of Purpose /
    AndrewWatt2001> Presentation / Data is a little wooly.

    AndrewWatt2001> There are two issues, from my
    AndrewWatt2001> perspective: 1. The examples in Chapter
    AndrewWatt2001> 2 are not ideally structured and need to
    AndrewWatt2001> be re-drafted 2. I suspect the Purpose /
    AndrewWatt2001> Presentation / Data structure itself may
    AndrewWatt2001> be fundamentally flawed or redundant

    AndrewWatt2001> I feel a little like the wee boy daring
    AndrewWatt2001> to suggest the Emperor has no clothes,
    AndrewWatt2001> so I hope you will be patient as I try
    AndrewWatt2001> to explain my concerns.

    AndrewWatt2001> Let me deal with the more superificial
    AndrewWatt2001> concern first. In the table in Chapter 2
    AndrewWatt2001> of the WD under the "Purpose" header we
    AndrewWatt2001> see terms like "Time Card" and "Order
    AndrewWatt2001> Form". It seems to me that those **in
    AndrewWatt2001> terms of purpose** (assuming it has its
    AndrewWatt2001> natural meaning) would be more
    AndrewWatt2001> appropriately expressed as "Collection
    AndrewWatt2001> of worker time data" and "Collection of
    AndrewWatt2001> order data". That sort of term/phrase is
    AndrewWatt2001> a "purpose", as I understand the
    AndrewWatt2001> term. Terms like "time card" and "order
    AndrewWatt2001> form" are actually presentations (or
    AndrewWatt2001> include an element of presentation) in
    AndrewWatt2001> my view.

    AndrewWatt2001> When you come to use phrases like those
    AndrewWatt2001> I have suggested under the Purpose
    AndrewWatt2001> heading you tend to find that it is a
    AndrewWatt2001> litany of "[whatever type of] data
    AndrewWatt2001> collection". And the data is listed
    AndrewWatt2001> under the Data heading.

    AndrewWatt2001> Which brings me to my deeper concern. Is
    AndrewWatt2001> the three way division of Purpose /
    AndrewWatt2001> Presentation / Data needed at all?

    AndrewWatt2001> For the scenarios which readily come to
    AndrewWatt2001> mind I find the three way division
    AndrewWatt2001> working out as: 1. Purpose - collecting
    AndrewWatt2001> X data 2. Presentation - [potentially
    AndrewWatt2001> multiple] - order form on desktop PC -
    AndrewWatt2001> order form on palm computer etc etc
    AndrewWatt2001> 3. Data - X data (maps usually -
    AndrewWatt2001> always??? - one to one with item 1,
    AndrewWatt2001> Purpose).

    AndrewWatt2001> The obvious scenarios, at least to me,
    AndrewWatt2001> work out as 1 purpose : multiple
    AndrewWatt2001> presentations : 1 set of data So if
    AndrewWatt2001> purpose and data map one to one do we
    AndrewWatt2001> really need a three way structure?

    AndrewWatt2001> The other possible scenario is: 1
    AndrewWatt2001> purpose: multiple presentations :
    AndrewWatt2001> multiple sets of data (attenuated for
    AndrewWatt2001> wee devices) But is that "one" XForms
    AndrewWatt2001> model at all? Or, when we have multiple
    AndrewWatt2001> sets of data don't we, when we think
    AndrewWatt2001> about the situation more precisely, have
    AndrewWatt2001> multiple purposes?

    AndrewWatt2001> As an example we might have: full
    AndrewWatt2001> personal info collection : desktop PC
    AndrewWatt2001> presentation : full personal info abbrev
    AndrewWatt2001> personal info collection : mobile phone
    AndrewWatt2001> form : abbrev personal info

    AndrewWatt2001> Again, in this type of scenario where
    AndrewWatt2001> the purpose (carefully spelled out)
    AndrewWatt2001> differs the data set seems to move in
    AndrewWatt2001> parallel - again with a one to one
    AndrewWatt2001> relationship.

    AndrewWatt2001> I hope I have been able to express why I
    AndrewWatt2001> have doubts about the current
    AndrewWatt2001> presentation of Chapter 2 and the
    AndrewWatt2001> underlying structure fairly clearly. If
    AndrewWatt2001> WG members have scenarios where we have
    AndrewWatt2001> 1 or many purposes : multiple
    AndrewWatt2001> presentations : 1 set of data then I
    AndrewWatt2001> guess we may actually need the three way
    AndrewWatt2001> Purpose / Presentation / Data division.

    AndrewWatt2001> If not, is simplicity and clarity not
    AndrewWatt2001> better served by a two way Presentation
    AndrewWatt2001> / Data Model view?

    AndrewWatt2001> I would like to suggest that the WG
    AndrewWatt2001> considers that a two dimensional
    AndrewWatt2001> Presentation / Data Model
    AndrewWatt2001> explanation/structure is more concise
    AndrewWatt2001> and perhaps better represents the
    AndrewWatt2001> potential advantages which XForms brings
    AndrewWatt2001> to the Web.

    AndrewWatt2001> Andrew Watt <HTML><FONT
    AndrewWatt2001> FACE=arial,helvetica><FONT SIZE=2>I
    AndrewWatt2001> guess since the XForms WD is almost at
    AndrewWatt2001> Last Call WD it is probably time to
    AndrewWatt2001> raise an issue which has been niggling
    AndrewWatt2001> me for several versions of the spec.
    AndrewWatt2001> <BR> <BR>I think the division of Purpose
    AndrewWatt2001> / Presentation / Data is a little wooly.
    AndrewWatt2001> <BR> <BR>There are two issues, from my
    AndrewWatt2001> perspective: <BR>1. The examples in
    AndrewWatt2001> Chapter 2 are not ideally structured and
    AndrewWatt2001> need to be re-drafted <BR>2. I suspect
    AndrewWatt2001> the Purpose / Presentation / Data
    AndrewWatt2001> structure itself may be fundamentally
    AndrewWatt2001> flawed or redundant <BR> <BR>I feel a
    AndrewWatt2001> little like the wee boy daring to
    AndrewWatt2001> suggest the Emperor has no clothes, so I
    AndrewWatt2001> hope you will be patient as I try to
    AndrewWatt2001> explain my concerns.  <BR> <BR>Let me
    AndrewWatt2001> deal with the more superificial concern
    AndrewWatt2001> first. In the table in Chapter 2 of the
    AndrewWatt2001> WD under the "Purpose" header we see
    AndrewWatt2001> terms like "Time Card" and "Order
    AndrewWatt2001> Form". It seems to me that those **in
    AndrewWatt2001> terms of purpose** (assuming it has its
    AndrewWatt2001> natural meaning) would be more
    AndrewWatt2001> appropriately expressed as "Collection
    AndrewWatt2001> of worker time data" and "Collection of
    AndrewWatt2001> order data". That sort of term/phrase is
    AndrewWatt2001> a "purpose", as I understand the
    AndrewWatt2001> term. Terms like "time card" and "order
    AndrewWatt2001> form" are actually presentations (or
    AndrewWatt2001> include an element of presentation) in
    AndrewWatt2001> my view.  <BR> <BR>When you come to use
    AndrewWatt2001> phrases like those I have suggested
    AndrewWatt2001> under the Purpose heading you tend to
    AndrewWatt2001> find that it is a litany of "[whatever
    AndrewWatt2001> type of] data collection". And the data
    AndrewWatt2001> is listed under the Data heading.  <BR>
    AndrewWatt2001> <BR>Which brings me to my deeper
    AndrewWatt2001> concern. Is the three way division of
    AndrewWatt2001> Purpose / Presentation / Data needed at
    AndrewWatt2001> all?  <BR> <BR>For the scenarios which
    AndrewWatt2001> readily come to mind I find the three
    AndrewWatt2001> way division working out as:
    AndrewWatt2001> <BR>1. Purpose - collecting X data
    AndrewWatt2001> <BR>2. Presentation - [potentially
    AndrewWatt2001> multiple] <BR> - order form on desktop
    AndrewWatt2001> PC <BR> - order form on palm computer
    AndrewWatt2001> etc etc <BR>3. Data - X data (maps
    AndrewWatt2001> usually - always??? - one to one with
    AndrewWatt2001> item 1, Purpose).  <BR> <BR>The obvious
    AndrewWatt2001> scenarios, at least to me, work out as
    AndrewWatt2001> <BR>1 purpose : multiple presentations :
    AndrewWatt2001> 1 set of data <BR>So if purpose and data
    AndrewWatt2001> map one to one do we really need a three
    AndrewWatt2001> way structure?  <BR> <BR>The other
    AndrewWatt2001> possible scenario is: <BR>1 purpose:
    AndrewWatt2001> multiple presentations : multiple sets
    AndrewWatt2001> of data (attenuated for wee devices)
    AndrewWatt2001> <BR>But is that "one" XForms model at
    AndrewWatt2001> all? Or, when we have multiple sets of
    AndrewWatt2001> data don't we, when we think about the
    AndrewWatt2001> situation more precisely, have multiple
    AndrewWatt2001> purposes?  <BR> <BR>As an example we
    AndrewWatt2001> might have: <BR>full personal info
    AndrewWatt2001> collection : desktop PC presentation :
    AndrewWatt2001> full personal info <BR>abbrev personal
    AndrewWatt2001> info collection : mobile phone form :
    AndrewWatt2001> abbrev personal info <BR> <BR>Again, in
    AndrewWatt2001> this type of scenario where the purpose
    AndrewWatt2001> (carefully spelled out) differs the data
    AndrewWatt2001> set seems to move in parallel - again
    AndrewWatt2001> with a one to one relationship.  <BR>
    AndrewWatt2001> <BR>I hope I have been able to express
    AndrewWatt2001> why I have doubts about the current
    AndrewWatt2001> presentation of Chapter 2 and the
    AndrewWatt2001> underlying structure fairly clearly. If
    AndrewWatt2001> WG members have scenarios where we have
    AndrewWatt2001> <BR>1 or many purposes : multiple
    AndrewWatt2001> presentations : 1 set of data <BR>then I
    AndrewWatt2001> guess we may actually need the three way
    AndrewWatt2001> Purpose / Presentation / Data division.
    AndrewWatt2001> <BR> <BR>If not, is simplicity and
    AndrewWatt2001> clarity not better served by a two way
    AndrewWatt2001> Presentation / Data Model view?  <BR>
    AndrewWatt2001> <BR>I would like to suggest that the WG
    AndrewWatt2001> considers that a two dimensional
    AndrewWatt2001> Presentation / Data Model
    AndrewWatt2001> explanation/structure is more concise
    AndrewWatt2001> and perhaps better represents the
    AndrewWatt2001> potential advantages which XForms brings
    AndrewWatt2001> to the Web.  <BR> <BR>Andrew Watt
    AndrewWatt2001> &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</FONT></HTML>

-- 
Best Regards,
--raman
------------------------------------------------------------

IBM Research: Human Language Technologies
Phone:        1 (408) 927 2608
Fax:        1 (408) 927 3012
Email:        tvraman@us.ibm.com
WWW:      http://www.cs.cornell.edu/home/raman
PGP:          http://emacspeak.sf.net/raman.asc
Snail:        IBM Almaden Research Center,
              650 Harry Road
              San Jose 95120
Received on Monday, 17 December 2001 14:24:07 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 10 March 2012 06:21:50 GMT