The aura of eXtensible Markup Language (XML) has lured members of SIIA, in conjunction with SIIA's Financial Information Services Division (FISD), to foster an organization dedicated to drafting a new data interchange specification specifically for financial market data. Will this volunteer group successfully decipher the complex composition of our global markets?

Article by MDDL Technical Committee Co-Chair, James E. Hartley, Reuters, for publication in the October issue of Upgrade

While, perhaps, not as monumental as the Human Genome Project (HGP), mapping all market data is a ponderous enough task - entailing the appropriate sequencing of data, vendors, and users - that many would not consider it possible. Such an attempt is underway. The goal of Market Data Definition Language (MDDL), is to map all market data into a common language and syntax to facilitate the interchange and processing of this complex and voluminous data set. Unlike the fifteen year planned lifetime of the HGP, however, fewer than fifteen months have been allotted to define and implement MDDL within the financial industry.

And so much progress has been made that the MDDL team is poised to present its "MDDL 1.0" specification at the fifth World Financial Information Conference (WFIC) in London, England on 02 November 2001. The architects of the standard believe that MDDL will be sufficiently mature by the WFIC that vendors and users will be able to begin to implement this data specification within their products and processing systems.

How has MDDL happened?

Jeremy Sanders, formerly of Merrill Lynch, spearheaded the concept of a common eXtensible Markup Language (XML) specification for market data in a convincing presentation during the FISD Meeting of members on 06 December 2000 in New York City. On 17 January 2001, FISD held an organizational meeting and accepted the requests of its members to guide the creation of a new XML specification. Under the wise yet capable direction of Michael Atkin, Vice President of SIIA and Director of FISD, the disparate members of the financial markets community have banded to together as a volunteer team to decode market data into a common interchange format.

FISD defined an organizational structure for the development of the specification involving a Technical Committee and Vocabulary Committee reporting to a Steering Committee so individuals and responsibilities could be matched consistently with skills and interests. The Steering Committee, the overall guiding force and authority for MDDL, is comprised of the co-chairs of the other committees and several key members active in SIIA and FISD to balance the technical and business related firms as well as the market data vendor and user communities.

As the specification became more palpable other XML initiatives in the financial industry began to take notice. Accordingly, the Steering Committee recognized the need for formal representation of the specification and for interaction with other XML specifications and created the Liaison Committee. This group is building relationships among both "standards" bodies and the other XML groups.

The specification has received much attention and support from many members and is a wonderful example of cooperation, even between competitors, of forging an alliance to reach a common goal. This effort will climax on 02 November 2001 with the release of "MDDL 1.0" at the World Financial Information Conference in London. This milestone will mark several industry "firsts" eliciting a resounding affirmation that the financial market industry has entered the 21st century.

What is XML anyway?

The eXtensible Markup Language (XML) is a generic mechanism for representing structured data in textual format. Programs that produce and use data often store this information in disk files with a proprietary or binary data format but only programs that specifically know how to read and process that data format can make use of the data. XML offers a way to "tag" the data in a tree structure, using a clearly defined set of rules, allowing any XML aware computer program to unambiguously process the data and maintain any relationship between the data. Further, documents generated using the XML standard can be extended to include new groups of data without affecting existing applications.

XML documents are in textual format and humans can view them without any special computer programs. However, the data is often very complex and the additional XML markup is generally useless to a human. As such, several supporting technologies have been developed to assist computers in processing and to aid humans in viewing XML structured data. For example, XML Stylesheet Language (XSL) allows a user to define a stylesheet that, when properly processed by a general XSL transform engine, turns the XML into something more readily viewable (like HTML).

XML is already accepted as a web standard for information exchange and, unlike former such efforts, individual organizations have not made their own derivations. Thus XML is evolving as a true standard. XML is not new but has been active recently. XML is derived from Standard Generalized Markup Language (SGML) - an International Standards Organization (ISO) standard since 1986. XML has been a World Wide Web Consortium (W3C) standard since February 1998 and, in May 2001, the W3C unleashed the true flexibility of the standard by approving its newest form of presenting XML data specifications called "schemas." The introduction of schemas into XML, as well as a variety of related XML processing, has enabled XML to gain even wider acceptance.

What will MDDL provide?

The principal tenet of MDDL is to exchange financial information on financial instruments. Effectively passing this information between providers and users implies a common way of defining and representing the data without ambiguity. As such MDDL provides the following basic benefits:

A common understanding of market data through the generation of standard definitions of terms as well as normalized relationships of various data elements to one another. This "glossary," (managed by the MDDL Vocabulary Committee) provides a common data taxonomy for all market data.

A common format for data interchange so that it may be used by any processing system from the genesis of the market data through to the end user thus facilitating the exchange of the data amongst all interested parties. The MDDL Technical Committee is managing the specification.

What did Vocabulary do?

The Vocabulary Committee was chartered with the Herculean task of creating a data dictionary - a glossary of terms - to precisely define the variety of market data concepts. The intent of the glossary is to clarify market data for the providers and the users so interchange of specific data elements could be unambiguously communicated with precision, transparency and consistency.

The Vocabulary Committee is accomplishing this by first defining market data products by the instruments they represent. For example, bonds and common equities are different products. The categories of products defined by the committee encompass a wide range of exchange and non-exchange traded instruments. Each category is broken into groups called domain further differentiated by class and subclass. Each product has different terms associated with it as reflected in the specification.

The various terms associated with market data are called descriptors with each descriptor given a unique definition. In most cases, the descriptors require other modifiers to clarify the content. For example, the price related descriptors (like "high") require a modifier "currency" to clearly denote the currency units of the price. In some cases, a single descriptor may have multiple definitions and the meaning can be clarified with a controlled vocabulary in much the same way a dictionary provides enumerated meanings for ambiguous words.

What did Technical do?

The Technical Committee independently set about determining how XML would be used to represent the concepts of the Vocabulary Committee and the basic groups of products (called constructs) - snapshot and end-of-day - initially targeted. The specification has the flexibility to add other constructs, products, and properties in future revisions as well as vendor specific additions to products and properties.

The descriptors and modifiers defined by the Vocabulary Committee are collectively called properties and each can hold a single data type (number, date, string, etc.). The properties are arranged in a simple hierarchy so it is clear which properties belong to which product and which properties are used to modify other properties.

The Technical Committee chose to support both the newer XML schema and the older Document Type Definition (DTD) formats for defining the XML language used within MDDL. The specification is actually written in XML schema and then XSL scripts are used to convert the schema to a DTD. In this way, those organizations familiar with validating XML with a DTD can adopt MDDL readily and MDDL can still use the advanced features defined in XML schema in the future.

What makes MDDL viable within data processing systems?

Once an organization understands how to process XML documents, MDDL is a straightforward extension of that same processing. The specification includes mechanisms to streamline handling and support proprietary extensions not covered in the initial specification. In addition, once MDDL documents are processed from one source this same processing should apply to MDDL documents from all sources.

Common material need not be included in each instance document: In large documents, or to make smaller documents, it is possible to identify groups of common data in a special section of the MDDL document called "reference". These groups (such as the formal name and address of a corporation) can be referenced within the same MDDL document or by another MDDL document externally.

Compression techniques are available to minimize bandwidth and storage requirements: A variety of general text-based compression techniques are widely available that can compress XML documents efficiently. If desired, firms can implement MDDL-specific methods for greater compression.

Proprietary products and properties can be added: MDDL is extensible due to the very nature of XML but specific elements have been added to facilitate the identification of proprietary products and properties thus facilitating generalized processing. The addition of proprietary constructs is expected in a later release.

Properties applied to entire data tree: The properties associated with MDDL items are inherited throughout the hierarchy of products and properties defined in the specification. It is not necessary to explicitly define all common properties with every term they modify. For example, the currency can be defined at the highest level in the document and the implication is that all other prices within that hierarchy are quoted in that currency unless specifically stated otherwise.

What other specifications relate to MDDL?

Other existing XML specifications are related to market data although none specifically addresses all the data represented in MDDL. In fact, most other XML activities in the financial space will use a portion of MDDL data or would like to have a markup for that data. Of these, the following are most prevalent:

XBRL - Extensible Business Reporting Language - for transporting corporate statements and other financial reports.

NewsML & NITF - defining news industry standards for relaying stories and supporting elements.

RIXML - Research Information Exchange Markup Language - defining investment and financial research.

FpML - Financial products Markup Language - focusing on over-the-counter derivatives like interest rate swaps and forward rate agreements

SWIFT / ISO15022 - assisting in definition of electronic messages pertaining to securities and related financial instruments.

It must be noted that a particular area of interest to each of these various specifications, as well as MDDL, is the unique identification of all financial instruments. This is an open industry wide concern that affects MDDL directly though is not within the purview of MDDL to solve (however there is another activity within FISD working on symbology and unique security identification). MDDL is written with the anticipation that such a unique identifier mechanism will be made available.

What does all this mean?

MDDL, Market Data Definition Language, has been created to describe financial market data in a way that facilitates the understanding and the interchange of this very industry specific data. Financial industry and technical experts are working together in an open forum to ensure functionality and content consistent with established and evolving practices. Developed in a short time frame, the initial specification focuses on a limited number of instruments but is poised to encompass the full range of market data products.

From the user firm perspective, the primary goal is to make it simpler and more cost-effective to integrate and normalize data delivered from multiple sources -- vendor, third parties and direct from exchanges -- and to make it easier for the development of web-based applications. From the vendor and data provider perspectives, MDDL promotes data linkages by allowing the creation of relationships between internal and external data. MDDL also allows for data transparency at a very granular level. As such, it is a powerful technology for product enhancement and a useful tool for mass customization of data.

Note that MDDL will likely spawn new processing strategies and business opportunities for many corporations. As existing vendors and users incorporate MDDL into their existing systems, many groups may find it possible to transfer assets from traditional collection and operations information technology roles to data quality and assurance positions. The easier transference of data will hopefully lead to a concentration on quality and thus an overall more satisfying user experience.