Lists Home |
Date Index |
From: Gavin Thomas Nicol [mailto:firstname.lastname@example.org]
On Tuesday 18 February 2003 11:40 am, Bullard, Claude L (Len) wrote:
> So, to repurpose, reuse, reapply?
>> But none of these make it much easier to work with
>> the document content as you point out. What will
>> make users happier is easier reuse? However, if this
>> is at the cost of making it harder to publish the
>> original document (say, the RFP), that robs Peter
>> to pay Paul. So the publisher isn't happy.
>Right. That's the hard part. From what I have seen, it is very unlikely that
>we will ever reach a situation where everyone is happy.
Not usually. That is "tension" I referred to as part of the
XML expert's problem to solve. They have to work out the
publishing/tool/workflow tradeoffs. David Megginson refers to the
time it takes to set this up, and he is right that it can take
a long time. In my world, that is called the implementation phase.
In this phase, after the contracts guys leave, the customization work
that adapts the system to local requirements is performed. The customer
must be an active party for lots of reasons. Then, the system must
provide for a certain amount of customizability which the owner can
do 'at their own risk'. The vendor has to be clear and the owner
has to ask about those risks. One reason for using schemas of some
kind to drive GUIs is to have a common place to put the information
that expresses the customization and bounds the risk.
>> The original document should be a data source in its
>> original format. Now XML is helping.
>Yes, the problem is the fidelity. There is a spectrum of content and the way
>people approach content, from the extremely visual, through to the highly
>structured (and possibly abstract). People who approach documents visually,
>will have a very hard time authoring using styles... they'll want to fiddle
>with fonts instead (how many documents have you seen where every style was
>Normal?). In those cases, the original document, while it *should* be the
>data source, most likely cannot be in any productive sense.
If style is the organizing principle, I agree. But why do that? This seems
to be the big gain in using Office with XML: author in the style one is
accustomed to or dealt, then export in accordance with a schema. At that
point, if reuse is a big deal, the system needs to make the document queriable
in accordance with the schema which is acting as a data dictionary to documents
of that type.
>I once dealt with a bank that produced a book of credit reports (a *huge*
>book). There were all kinds of uses the data could be put to, from setting
>interest rates, approving loans, etc. onward *if* the data could be reused.
>The problem was that in some cases, the credit reports were GIF images, in
>other cases, Excel spreadsheets, and the rest were in everything between
>them. The obvious solution was to standardize the data entry side so that
>data could be accurately captured in XML, or somesuch, but that was precisely
>the thing that could *not* be done. Prototypes based around forms, word etc.
>all failed, because the people entering data were underpaid and
>undermotivated to enter data appropriately. This is an extreme case, but I
>think there are many situations where siilar problems will raise their head.
>While I applaud XML on the desktop, I'm not seeing a silver bullet...
Yes. It is an enabling technology, not an end to itself. And what you
mention above goes directly to the "tensions in many dimensions" I am
>This is just on the data entry side too... the rest of the process of
>establishing side-by-side views, or formatting, or extracting summaries, or
>building indexes, or <whatever/> still needs to be done.
Yes. The buyer has to become acquainted with what is possible given
the tradeoffs that they make for convenience, legacy, etc. vs reuse.
XML is not a free lunch.
The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
initiative of OASIS <http://www.oasis-open.org>
The list archives are at http://lists.xml.org/archives/xml-dev/
To subscribe or unsubscribe from this list use the subscription