UBL: Another Opportunity for FOSS in the Enterprise

E-business still lacks a universal, cheap and easy-to-implement
standard language. At least, this was the case until a few months
ago. Today,
the Universal Business Language
(UBL) is ready to fill this gap, and it looks to be solid offering
rather than yet another bunch of buzzwords. UBL comes from
OASIS), the
same folks who standardized the
OpenDocument format
for office files, and UBL is equally as open. Participation in the OASIS UBL
Technical Committee is open to all organizations and individuals.
OASIS hosts an open mailing list for public comment, as well as the ubl-dev
mailing list for exchanging information on implementing the
standard. UBL is provided on a royalty-free basis and is available to
all, without licensing or other fees.

UBL starts with three basic premises. First, XML is a wonderful
technology, but in and of itself, it is no guarantee of
standards. Second, no company works in isolation or has contacts only
inside its industry segment. Therefore, if every company had its own XML-based
system, nobody would accomplish anything. Good business requires
efficient, cross-industry communication. Third, forget
about reinventing the wheel or substituting it with some vague,
untested idea. UBL is all about doing tomorrow the same things we are
doing today but in a faster and less error prone way.

The basic goal behind UBL is simply to define a single, XML-based format
for all the usual commercial documents--invoices, purchase orders and so
on. Using this format, an American buyer could create and e-mail in
English an order for, say, one hundred carpets. The Thai employee who
receives the order in her in-box could open the file with any UBL-enabled
application and read it in her native language; the document already
would be formatted according to local policies and business regulations.
In addition to this, she simply could drag-and-drop the order to the
company inventory system without typing anything, without wasting time
and without risking errors.

UBL also can be used as the business data format in a variety of Web
services. Although solutions to this problem have existed for years,
before UBL, such solutions were cost effective for only large corporations
with millions of transactions to process each year.

To know more about this UBL standard and how FOSS programmers may add
support for it in office applications, starting with OpenOffice.org, we
talked to Jon Bosak, chairman of the UBL Technical Committee, and Lars
Oppermann, software engineer for the framework and XML projects in
StarOffice/OpenOffice.org.

Linux Journal: Let's start by looking at
how to add UBL support to some software applications. Which specifications
should a developer read? Are APIs and libraries already available; if so,
under which license? Would they be constrained, at least today, to only
a few software languages?

Jon Bosak: Everything currently available
for users and developers is included as part of the
UBL 1.0
specification. Because UBL 1.0 was released recently as a standard,
no APIs or software libraries are available for it yet. But, I expect to see those appear as soon as
software developers begin to wake up to the huge business opportunity
presented by UBL. I often compare UBL to HTML. Both are specific XML
tagging languages created to enable interoperability in specific domains,
which for HTML was hypertext publishing and for UBL is
business-to-business electronic commerce. UBL is at about the same
place on the adoption curve that HTML was in 1992. So there is not
much support available yet, but we can expect that to change very
quickly as people start to understand the advantages of a standard
approach to business data interchange.

LJ: How can UBL documents be integrated in
current office environments? Is it already possible to use this standard in
production with OpenOffice.org, Microsoft Office or any other
application, from corporate-level products to SOHO ones?

JB: The key here is the creation of plug-ins that will enable
office software packages to translate their internal data formats to
the standard UBL form for output and to perform the reverse
translation for input. Early prototypes have demonstrated the
viability of this approach. Now it's up to the producers of these
office systems to provide the necessary translation capabilities.

LJ: Let's hope that free software
products will be the among the first to provide UBL support. It might
be exactly what is needed to make more businesses switch to FOSS to reduce costs.

Is UBL only for business-to-business use, or is it also usable/needed by
private citizens?

JB: UBL defines standard forms for
basic business messages, such as purchase orders, shipping notices
and invoices. It is not intended for non-business use. However,
some pieces of the UBL standard, such as the UBL naming and design rules for XML schemas and the UBL
specification for code lists, are being adopted for other XML
standardization initiatives. So, it's quite possible that future XML
standards usable by private citizens may be based in part on these
aspects of the UBL work.

LJ: Let's discuss with Lars the
technical details of the problem. Can UBL 1.x documents already be read
or written with OpenOffice.org?

Lars Oppermann: OpenOffice.org is aiming
to support UBL through its general XML integration efforts. The software
can read or write any XML-based format, including UBL instances, by means
of its XSLT import and export filter. We have worked on a UBL import of XSLT
transformation and used it for demonstration purposes within the OASIS
UBL committee. In the future, full UBL instance editing and creation
capabilities will be provided through
OpenOffice.org's
XForms support.
This approach has shown promising results and already was demonstrated at
XML2004.

LJ: Who is working on this?

LO: The UBL Technical Committee at OASIS has formed a
human interface sub-committee (HISC) that is looking at ways to edit UBL. This
committee also looked at XForms and OpenOffice.org XForms.

JB: Lately the HISC has made considerable progress in defining
input specifications for UBL documents. These specifications are not
XForms but can be applied using XForms. Micah Dubinko, a leading XForms expert, has done a lot of work
on the spec over the last few months. To have an idea of what is
happening, check out this
wiki.

LJ: What would be the best way to start an UBL/OO.o related
project?

LO: Getting in touch with the OASIS UBL folks or with the
community at xml.openoffice.org would be the best course of
action. Because the UBL Technical Committee itself is not so much
interested in particular implementations but rather in general
approaches, OpenOffice.org might be the best place.

LJ: What about cross-language support? Imagine an English
company sending a purchase order to a French company. The promise of UBL
is it should make it possible for the French user's software to
parse UBL and show on-screen a French version of the same
document. What would OO.o need to do this? More coding--how much, at
which level and why?--or simply the right libraries, templates and so
on?

LO: Everybody would need a localized version of the XForms
document that displays the actual instance. Data in the instance
mostly isn't locale dependent. If localized data is contained in the
instance, XForms could provide means for filtering out the correct
values.

JB: The UBL 1.0 International Data
Dictionary (IDD) provides an important
resource for developers creating localized UBL software. An IDD draft
for Chinese, Japanese, Korean and Spanish is undergoing balloting as
an OASIS Committee Draft and should be available for public review
[sometime in April]. A preliminary version (Excel, sorry) of the IDD is
already available
here. In
that document, the columns labeled "business terms" are intended specifically
for use in populating localized drop-down menus and field
labels in forms. As with all the UBL 1.0 spreadsheets, the IDD will be
provided in both OpenOffice.org and Excel formats when the Committee Draft
is published.

LJ: Any other general comments or
pointers for developers?

LO: No actual coding is needed in
OpenOffice.org to support UBL. You only have to create XForms documents
that work on top of the UBL documents. Also, check out the
W3C XForms pages and
the
OpenOffice.org XForms specs.
General XML information on OpenOffice is available
here.

LJ: Thanks to Jon and Lars for their time.