XML and the New COBOL:
part 2

Articles
XML publishing
Sarbanes-Oxley

Podcast
Java and XML
SQL/XML
Portals

Webcast
Java and XML
SQL/XML
Portals

Sites
COBOL
INCITS
W3C XML

Barry Tauber of Victor Consulting is a long-time member of the committee responsible for the COBOL standard (INCITS J4). We continue his explanation of the integration of XML support into the COBOL standard.

<< Part 1

There are COBOL vendors, not one COBOL vendor; and so there is competition in the marketplace. Not only for the language, but within the language. Without meaning to slight anyone, some of the COBOL compiler vendors are: IBM, Micro Focus, Fujitsu, AcuCorp, LegacyJ. The vendors like having a standard, but they like more having customers. When their customers began requesting XML processing the vendors responded, but without a standard they had little guidance.

Since the various individual vendors had been processing XML data for quite a while, there were several implementations that were both functional and used in production in commercial applications. They differed in minor aspects, and basically what the committee is doing is picking best-of-breed of what already works and formalizing it into the standard.

What COBOL has actually done is define how the data is to be processed once it has been accessed. It is not a SAX model and it is not exactly a DOM model, it is native COBOL processing. There is a parser, but it is not seen by the COBOL code.

For those of you who have some COBOL under your belt, here is a nutshell of how XML is handled in COBOL:

External data connectors are defined in COBOL in the Environment Division.

Use a SELECT clause.

COBOL has added an organization of XML, with optional specifications of Schema or DTD validation.

Data itself is defined in the Data Division.

COBOL has added to the FD structure in the File section (which connects with the SELECT clause) to define the tagged layout of the XML. This layout uses COBOL terminology of level numbers and data names. Repeating structures use the standard COBOL occurs clause. There are new COBOL keywords IDENTIFIED and ATTRIBUTE for tags and attributes. Traditionally, COBOL handled UTF-8 and ASCII with PIC X clauses and still does. With the 2002 standard and the introduction of PIC N (National) characters, UniCode and UTF-16 are handled.

The structure can be hand coded to map to the physical XML structure but mostly is not hand coded. All the vendors provide an external utility to read the DTD, Schema, or physical XML and generate a COBOL layout.

Data is processed in the Procedure Division.

The XML FD is OPENed or CLOSEd (standard COBOL verbs). An open of an XML document sucks in the whole document, DOM style, and resolves any references, entities, etc. PIs are not passed through. The structure defined in the Data Division serves as the window to the internal structure. Standard COBOL verbs READ, WRITE, REWRITE, DELETE, and START are used to traverse the internal structure. A form of Close can write the modified internal structure back out to wherever it came from.

That is the nutshell. The TR is the detail, and is written as changes against the 2002 standard. The 2002 standard itself is available at ANSI, or you can contact me via e-mail

There were a bunch of companies and a bunch of folks involved here and I cannot name all the names here (some do not want to be named). The people responsible for checking for errors, making sure changes work and do not invalidate other working code are: