QUESTION & ANSWER

Paoli: Information correlation, reasoning next step for XML

By Michael S. Mimoso, Senior News Editor 21 Dec 2004 | SearchWebServices.com

Digg This!     StumbleUpon     Del.icio.us

Jean Paoli

How did you react to winning the XML Cup and earning that kind of recognition from your peers?
Jean Paoli: It was very emotional, actually. I started in 1985 working on this subject; I know a lot of the people, and it's really a community effort. For me it's emotional because when you work on something for 20 years, it's nice to see millions of people adopting it.

Can you detail some of your early work on XML?
Paoli: I started in 1985 in Europe doing research on the migration of technology and putting those results on the market. I worked with INRIA, a French computer lab, doing work with people in the research community, and a lot of those projects started around semi-structured data.

In early 1990s, Sharon Adler Tim Bray, Jon Bosak, all had a vision about open data and opening the data format. At the time, I was also a Unix programmer using Sun before I moved to Microsoft in 1996. Bill Gates' manifesto on the Internet was in 1995, and Microsoft wanted to talk to me. I didn't think HTML was good enough, they needed something around data, not presentation.

I helped create the committee (what became the XML technical committee at the World Wide Web Consortium) with Jon Bosak, my friends from SGML and my friends at Microsoft. This was really important for me to make the bridge between classic SGML and my movement. My friends got excited because I was recruited to help build Internet Explorer 4 from the ground up. This was really important for me to make the bridge between classic SGML and my movement.

SGML was very small at the time, and Microsoft developers were giving me their feedback on the standard, which was too complicated. This was very interesting for me because a Microsoft developer helped me think about the problem in a different way to simplify the standard. That integration of ideas between SGML and Microsoft's idea of simplicity were very interesting. Early on in 1997, Microsoft announced its first support for XML. There were a bunch of submissions, and I was organizing the mix between Microsoft and the standard committee. James Clark, Sharon Adler, myself and others spent one week in Redmond to prepare the XSL submission to the W3C; we needed a transformation language to transform XML to presentation. Eventually, we put XML in IE and Windows, at the back end in SQL. I was the manager of the team that created the XML parser and jumpstarted XML at Microsoft.

For IT it's data, for users, it's documents. Then, we'll have millions of new opportunities to start correlating the information, think about reasoning on top of it and managing it. Jean Paoli Senior director XML architecture

Why start with XML at the back end?
Paoli: I did not want to do the same mistakes [as SGML], and focus too much on the user. In the 1990s, I was doing SGML editors for users, building user interfaces. But when you have a user interface and no back end, no one can go and manage the data. That's why I focused on the back end for the first two to three years, putting it in Exchange and SQL Server.

And obviously Microsoft supported your strategy?
Paoli: In 1997 when I talked with Adam Bosworth to Bill Gates, he gave us the resources we need it. Bill understood it in five minutes, but he couldn't bet the company on it. In 2000, when announced Web services, that's when he bet the company on it.

Around 1999-2000, I moved to Office. I knew there was enough on the back end, and it was time to start the front end thing: InfoPath [an collaborative application that enables the sharing and reuse of information via XML. It is integrated into Office.] and the overall Office XML support (Word, Excel etc).

Has XML exceeded your expectations?
Paoli: All of these ideas are a reality after 20 years. People don't know how much Web services and XML are part of the mainstream. It's really something that's used today. It's a huge thing -- interoperability -- I see that every day. Governments all over the world are using XML for e-government initiatives, government-to-consumer and government-to-business.

I truly believe in this open data -- I put my professional life on it. Now you're seeing people in hospitals using InfoPath in the ER. It's something where you can show the vision for documents to be processed in backend systems, to consider documents as source of information that could be treated by computers, not for humans only. It's a another way to present documents, by stressing their content, not how they look on a screen. You cannot focus on just how they look, but on what's inside.

FOR MORE INFORMATION Read about the Wisconsin DOJ's experience with XML and SOA Read about Commerce One's recent sale of XML patents

Now that XML is successfully employed on the front end as well, what's next?
Paoli: XML is at the front end. Adobe, OpenOffice, they're adding XML to the front end as well. The technology is here, that's the fun part of this for me. This is the tip of the iceberg because imagine a time when millions of documents are stored and archived in XML. Project yourself in a few years, millions of documents created by humans, stored in databases. For IT it's data, for users, it's documents. Then, we'll have millions of new opportunities to start correlating the information, think about reasoning on top of it and managing it.

That's what I call the second step. In the first step, the tools are all here.

What's your take on recent reports that XML traffic is overwhelming enterprise networks?
Paoli: I don't think personally it's an issue. It's always the same thing with a new technology, people are afraid of it and in the end, it just works. Machines are faster today than five years ago. I know about this whole discussion around binary XML, I don't know. We are trying to understand what it all means. We don't have any yes-no answer.

The problem for binary XML standard is multiple. We have to understand the problem space, and there are two problems there: The first is technical. Is it possible to have one binary XML standard that can respond to all the technical limitations to all scenarios; mobile, military, etc. Microsoft believes there are multiple ways of optimizing more than one problem. ZIP works well for specific applications. Zipping the XML just works, but it doesn't' solve other problems.

There is also a people problem -- education and adoption -- that scares me.

Tags: VIEW ALL TAGS

Digg This!     StumbleUpon     Del.icio.us