XML and its uses

[Zamdrist]

At the expense of sounding like a complete newbie and nay-sayer, I have
not, at least in my field, yet seen a use for XML beyond for the sake
of itself.

I work as an IT professional, a database analyst/programmer in the
legal industry. No one has *ever* asked me for their data in XML and no
has ever sent me data in XML.

Is there truly a use for XML beyond using it for the sake of it? Is
there something XML provides that databases don't otherwise provide
already? I'm quite aware of the extensive use of XML, especially in
regards to the web and blogging acitvities. I've read a number of FAQs
on XML and do understand the basic principles.

I am however quite chagrined, being in the legal industry, where
documents and data are churned out like there is no tomorrow, how XML
has never been even considered.

 

[Joseph Kesselman]

Is there truly a use for XML beyond using it for the sake of it? Is
there something XML provides that databases don't otherwise provide
already?

XML is a standard for writing standards for portable data interchange
representations. It's a substiset of SGML, which had previously been
used for that purpose (though primarily in the document markup arena.)
XML has turned out to be remarkably useful as a framework around which
to build domain-independent data manipulations.

I am however quite chagrined, being in the legal industry, where
documents and data are churned out like there is no tomorrow, how XML
has never been even considered.

You forgot to say "that I'm aware of" -- and you haven't been looking
hard enough. The legal code for the state of Tasmania is now being
maintained in a database system which takes advantage of XML markup, to
take one particularly visible example.

 

[Zamdrist]

XML is a standard for writing standards for portable data interchange
representations. It's a substiset of SGML, which had previously been
used for that purpose (though primarily in the document markup arena.)
XML has turned out to be remarkably useful as a framework around which
to build domain-independent data manipulations.

Forgive me but, while that all sounds good, what practical, everyday
application does what you describe have?

We had a fairly significant, nationally based client recently ask us to
send them a report of their data to them, on a regular basis,
intellectual property law based data. They didn't want their data sent
to them in XML format...they wanted it in MS Excel.

You forgot to say "that I'm aware of" -- and you haven't been looking
hard enough. The legal code for the state of Tasmania is now being
maintained in a database system which takes advantage of XML markup, to
take one particularly visible example.

Well yes, that I'm aware of. Which is why I explained no one has asked
for their data, or sent me their data in XML format. No one has even
inquired about it. Ever. While we are not the biggest IP firm in the
nation, we are in the top 10 easily.

 

[Richard Tobin]

Forgive me but, while that all sounds good, what practical, everyday
application does what you describe have?

We had a fairly significant, nationally based client recently ask us to
send them a report of their data to them, on a regular basis,
intellectual property law based data. They didn't want their data sent
to them in XML format...they wanted it in MS Excel.

I don't see the connection between those two paragraphs. How does the
fact that your customers are using Excel have any bearing on whether
XML has practical applications? At most it tells you about the
commercial success of XML in some area.

-- Richard

 

[Soren Kuula]

At the expense of sounding like a complete newbie and nay-sayer, I have
not, at least in my field, yet seen a use for XML beyond for the sake
of itself.

That would also be a silly thing to do? One should use XML because it
can help solve a problem -- not because it's XML.

I work as an IT professional, a database analyst/programmer in the
legal industry. No one has *ever* asked me for their data in XML and no
has ever sent me data in XML.

Is there truly a use for XML beyond using it for the sake of it? Is
there something XML provides that databases don't otherwise provide
already?

XML is meta-language. You can make XML languages for a variety of
purposes. The languages are best at describing something tree-shaped, or
hierarchial.

You can define an XML language, and edit files that represent something
in the language. Can't do that with a database. You send an XML file to
someone else - can hardly do that with at database table. You can
describe almost anything in XML - Work documents, programs, mathematical
expressions, organisational hierarchies --- whatever. Databases can only
cope with some of these, and hardly in a descriptive way.

I think there is no claim that XML should replace databases. It's just a
way to define some data formats.

I am however quite chagrined, being in the legal industry, where
documents and data are churned out like there is no tomorrow, how XML
has never been even considered.

Probably just die-hard tradition. I also used to be at a place where
everything is thought of in terms of database tables. Those poor people
didn't see that a small change in the business model they are _really_
representing causes lots of changes in the database nitty-gritty. They
were constantly running back and forth, trying to keep their hundreds of
scripts working with revision #3243243. If they had looked at things
from a top-down perspective (which I think an XML description encourages
much better than a relational model), they could have gone home earlier
every day..............

Søren

 

[Soren Kuula]

We had a fairly significant, nationally based client recently ask us to
send them a report of their data to them, on a regular basis,
intellectual property law based data. They didn't want their data sent
to them in XML format...they wanted it in MS Excel.

From the next version of Office, Excel files will be (zipped) XML
too..............

Søren

 

[=?ISO-8859-1?Q?J=FCrgen_Kahrs?=]

Is there truly a use for XML beyond using it for the sake of it?

XML is often used by software running in the background.
In such cases, the XML data is not visible to the end-user.
For example, XML is used for storing the content of
object-oriented data in software applications.
Many other applications exist (like XAML and XUL), which
are mostly unknown to end-users.

You, as an end-user, might appreciate that XML has solved
the problem of representing international alphabets in an
elegant way (Unicode). I have seen Japanese documents
(being unable to actually read them), but it is commonplace
to pass Japanese files just like any English data files.

I am however quite chagrined, being in the legal industry, where
documents and data are churned out like there is no tomorrow, how XML
has never been even considered.

Considering new standards is one thing, but adopting them
is much more expensive. XML is no panacea. In some places
and industries, continuity of established standards is so
much more important than "XML-hype".

 

[Peter Flynn]

Forgive me but, while that all sounds good, what practical, everyday
application does what you describe have?

It lets me write documents which can then be published or reused in many
different formats without me having to maintain multiple parallel
versions; and to know that what I'm writing is in a file format that
will outlast both the software and myself. But that's because I write a
lot, and it saves me time, money, and effort.

In the legal field, I suspect people who actually do the writing are not
really concerned about whether their documents can be reused later or
not, nor about whether they will be long-lived. All they want is a
pretty interface, which Microsoft provides. The people who actually have
to make legal documents work in the long term are slowly discovering
that document formats *designed* to last a long time are Quite A Good Idea.

There are currently *no* XML editors suitable for direct authoring by
people who don't know XML (see my paper at Extreme Markup earlier this
month). This may well have something to do with why it's not used by
authors (except for those who learn what XML is).

We had a fairly significant, nationally based client recently ask us to
send them a report of their data to them, on a regular basis,
intellectual property law based data. They didn't want their data sent
to them in XML format...they wanted it in MS Excel.

That may well be because what you are sending them is not data: it's
results (the output of doing something to data). Or that the information
is transient or just not very important, and thus not worth preserving
(because it can always be regenerated from the original source data).
There are many good reasons for not using XML.

The recent CDs of legal code for Ireland were all produced from XML
master copies. No-one in their right mind would ever use anything else
for permanent textual information (nor has done since the late 1980s).

Well yes, that I'm aware of. Which is why I explained no one has asked
for their data, or sent me their data in XML format. No one has even
inquired about it. Ever. While we are not the biggest IP firm in the
nation, we are in the top 10 easily.

If they (and you) are happy with using Word/Excel/etc, stay with it. I
see no reason to force anyone to use a format they don't need or want.
Saving time, money, and effort isn't usually interesting when the cost
of training for XML adoption is high, the software is mediocre, and when
the nature of the information just doesn't require any of the benefits.

///Peter

 

[Joe Kesselman]

Saving time, money, and effort isn't usually interesting when the cost
of training for XML adoption is high, the software is mediocre, and when
the nature of the information just doesn't require any of the benefits.

Saying the same thing another way: XML is just a tool, not a silver
bullet. If it suits your current needs, great. If not, use something
else until your needs change.

In my experience, many legal documents often wind up going into a system
geared to handle ink-on-paper, and are generally manipulated as
unstructured plaintext before that... and are recreated de novo each
time (or from a template which is also plaintext and which must be
re-edited each time). A word processor, or even a basic text editor, is
a more than adequate tool for that environment.

If something is intended to be a _living_ document, with ongoing
generations of editing applied to it, multiple renderings, and so on,
portable semantic markup with separate styling becomes more useful, and
a standards-based text-compiler environment such as an SGML- or
XML-based system starts to gain the upper hand over the wordprocessors.
Ditto as the data becomes more structured. Of the two, XML is succeeding
better than SGML did because it hits the 90/10 point -- it delivers 90%
of the strengths of SGML for 10% of the implementation complexity and cost.

But human-readable documents are becoming the least interesting
application of XML -- even when you count the fact that the new version
of HTML is XML-based rather than SGML-based. Which is another reason
you're seeing slower uptake there than in the middleware data-exchange
layers.

Summary: If you aren't interested, and your customers aren't interested,
there's no magical reason you should care right now. At some point in
the future that equation is likely to change; you have to decide when
and how heavily you want to invest, based on whether you want to be on
the leading or trailing edge of that wave and your own best guess about
when it's going to hit your area.


Putting it yet another way: I don't think anyone's interested in getting
into a flame war about it. Use the tools that fit your way of
approaching your task to the satisfaction of your customers.

 

[Peter Flynn]

But human-readable documents are becoming the least interesting
application of XML

Actually they are the most interesting and challenging.

The use of XML for data-exchange is largely a done deal, with vast
amounts of software available, plentiful skills, relatively simple data
models, and ample support (even if people do go off at a tangent
sometimes

By contrast, modelling human-readable information is *hard*, with many
unsolved problems, poorly-developed software, and only a few experts.
It isn't "just publishing": there's a large amount of work still to be
done, and this makes it -- for me -- by far the most interesting area.

///Peter

 

[Joe Kesselman]

By contrast, modelling human-readable information is *hard*, with many
unsolved problems, poorly-developed software, and only a few experts.
It isn't "just publishing": there's a large amount of work still to be
done, and this makes it -- for me -- by far the most interesting area.

.... OK, point granted. The semantic-web stuff is one signpost on the way
to that set of problems. I still think Tim B-L has bit off much more
than he can chew in that area, but even partial success is likely to
produce some interesting results.

What XML is really buying us here is a shared platform that everyone can
agree to base their development on, with the hope of seeing it find real
applications (rather than getting stuck in academia) and the possibility
of synergy as solutions are combined and applied to new areas.

 

[David Dorward]

Forgive me but, while that all sounds good, what practical, everyday
application does what you describe have?

Err, getting the latest headlines from the BBC
<http://newsrss.bbc.co.uk/rss/newsonline_uk_edition/front_page/rss.xml>
(and thousands of other sites) springs to mind.