TREETIPS Issue 24
Worldwide Edition
On-line!

January 1998


Table of Contents

TSI Celebrates Its Crystal Anniversary

The Evolution of N2O Continues

Editor's Sproutings

What Users Are Saying About CHART

Off to California to Meet with ORACLE

Converting from NATURAL to COBOL

TRIM ADASMP Procedure

PREDICT 3.3 Support for N2O

Got Support?

What's New in tRelational Version 2.0.0

New Versions of SEDIT and S/REXX: Available for UNIX, and Now Windows

Disclaimer


TSI Celebrates Its Crystal Anniversary

Treehouse Software (TSI) is celebrating its 15th year anniversary. Since its founding in 1982, TSI has gained an excellent reputation as a leader in its field. With a substantial number of customers in the U.S. and having affiliates and customers in many foreign countries, this reputation has spread throughout the world.

TSI's focus has been on generalized security, control, conversion, and performance enhancement software complementary to the Software AG mainframe product line. The focus has been expanded with offerings for UNIX and ORACLE for OS/390.

The fifteen year anniversary is known as the "crystal" or "watch" anniversary, and a watch naturally brings thoughts of Y2K issues to anyone involved in our business. Since we are only two years away from 2000, you may need a reminder on your desk that TSI has tools to assist you in your Y2K effort. So, as a special gift to the first 100 people who call and mention this article, we will send them a free TSI Y2K clock. Also, check out our Y2K Tools special offer on our web site at http://www.treehouse.com.

And finally, TSI would like to thank all of our customers who have continued to use our products, and those individuals who have recommended them to others over the past 15 years.

Return to the Table of Contents


The Evolution of N2O Continues

by Lynn McIntyre

The N2O Development Team has been busy developing not only N2OSCAN, which was featured in the last issue of TREETIPS, but also another significant enhancement requested by our customers. The N2O Toolbox Subsystem now contains a third function for aiding in application development, testing, and documentation. This new function is the Documentation Tools Option. It provides utilities to display or print NATURAL Objects, File Layouts, Descriptor Cross-Reference (XREF) Information, Object Flow Analysis, and Object XREF reports from any local N2O Environment. These features allow the user to specify a range of NATURAL Objects or one specific Object.

Following is a brief description of the seven functions available in the N2O Documentation Tools option.

The NATURAL Object Listing provides a simple way for the user to generate a listing of NATURAL Object(s). The user may choose to embed the source of any Data Areas or Copycode referenced within the displayed object(s). NATURAL Maps and Data Areas may be displayed in raw source format or as the associated editor presents them. Variables and processing rules contained within a map are also displayed.

The NATURAL Map Listing function creates listings of Natural Maps presented in the same format as when using the NATURAL Map Editor. Variables and processing rules contained within the map are also displayed.

The NATURAL Data Areas Listing is presented in the same format as when using the NATURAL Data Area Editor.

The File Layout Listing may be generated using PREDICT or a NATURAL DDM as the source repository. The extended description and field level comments are optionally displayed. A specific file, a range of files, or a range of files with a specific PREDICT Keyword may be displayed using the File Layout Listing.

The Descriptor XREF Information report shows descriptor usage using PREDICT XREF Data. A Summary or Detail report may be produced. The Summary report includes the file name, descriptor type, field name, field format, and field length. The number of times the descriptor is used in a search, update, read, delete, or view is also displayed. The Detail report also shows the names of the objects in which the descriptors are used.

The Object Flow Analysis report shows a breakdown of a NATURAL Object's flow, showing processing loops and statement blocks. This report provides a visual aid in the identification of IF/END-IF, FOR/END-FOR, DECIDE/END-DECIDE, READ/END-READ logic. Debugging and analysis of NATURAL Objects is enhanced with the use of this report.

The Object XREF report utilizes PREDICT XREF Data to identify all Data Areas, Copycode, CALLNATs, and FETCHes within each object. How variables are used (update, read, find) is also included.

These functions enable a developer to access source code from any local N2O Environment, and are invaluable aids to designing, debugging, and documenting any application. The Documentation Tools option will be released as part of N2O Version 3.4.0.

Return to the Table of Contents


Editor's Sproutings

by Joseph Brady

TSI Sponsors User Group Meeting
SAG Americas recently announced its "postponement" of support for User Group meetings during the fourth-quarter of 1997. Many User Groups rely on this support to ensure their user communities meet and exchange technical information on a regular basis. Without this support, some of the meetings could not happen.

TSI announced an offer to financially support any local or regional User Group that wishes to have a fourth quarter meeting. One such meeting sponsored by TSI was the Rocky Mountain East User Group Meeting in Denver, Colorado on December 12, 1997. TSI presented "Y2K and Treehouse Software" showcasing N2OSCAN and PROFILER. Software AG gave presentations on ENTIRE X and PREDICT 3.4, and MCI gave a presentation entitled "Large Database Tuning, Tips, and Techniques". TSI also provided lunch and refreshments for the attendees.

The response from the region was tremendous. E. B. Ellis, Rocky Mountain East Regional President, said, "We usually get 20 or so people to attend. This meeting attracted 46 attendees!" We would like to think the great turnout was, in part, due to the hard work and planning provided by TSI.

If you are interested in having TSI support your User Group meeting in 1998, please contact us.

New Faces at TSI
TSI is pleased to welcome Tim Baker-Finch to our staff. Tim joined the TSI development team in November, 1997. Many of you are familiar with Tim's name because he is the principal of Harvest Moon Computing Pty Ltd. and the developer of CHART. Tim has also worked in various large ADABAS/NATURAL sites in Australia since 1983, such as the Australian Taxation Office. Tim began developing CHART in 1992 and marketing it within Australia in 1995. About 20% of the ADABAS/NATURAL sites in Australia are now using CHART.

When asked about the trek from Australia to Sewickley, PA, Tim said, "From August, 1997 we spent three months sorting and packing our possessions and we finally squeezed our family of five into nine suitcases and a small portion of a shipping container. From our home in Canberra, we literally took trains, planes, and automobiles to travel half-way around the world (via Disneyland) to live and work in a lovely village called Sewickley, just outside of Pittsburgh."

TSI welcomes Tim and his family to the States after their long journey from Australia.

John Hoefflin is now working for TSI as a full-time consultant with our development staff on expanding the capabilities of DELPHI. John has worked as an ORACLE DBA in the UNIX environment, and has developed applications using ORACLE development tools. He also has experience with Software AG products, and teaches UNIX and C programming.

TSI Product Updates
- DELPHI V2.3, an ORACLE Performance Monitor for OS/390, has been released. TSI and Oracle worked together to solve a problem with DELPHI and ORACLE Version 7.3.2. Many other enhancements include compatibility with higher versions of the ORACLE CMDLOAD library, various bug fixes, and several display enhancements. For more information, contact TSI.

- We recommend that all N2O customers upgrade to V3.3.3, so they are ready to take advantage of N2O V3.4.0. Also, please note that N2O Versions 3.2.x and earlier will no longer be supported after June 30, 1998.

- RACE V1.4 is close to completion. Documentation is being finalized at TREETIPS press time. The main features in V1.4 include: greatly improved percentage of eliminated RCs, optional use of Multifetch, handling of files with up to the ADABAS maximum number of fields, and four digit years in all displays.

- TSI is currently planning to release SECURITRE V3.1.0 in March of 1998. SECURITRE has been compatible with ADABAS 6 for over a year, but is currently limited to file numbers 1 - 255 (V3.1.0 will not have this limitation). Beta testing will begin in February of 1998. Any sites interested in becoming a Beta test site for this release of SECURITRE should contact TSI as soon as possible.

TSI Meets with SAG-Americas
As a follow-up to visits by several SAG representatives to TSI, we recently visited SAG-Americas in Reston, VA. While a variety of issues were discussed, central focus was on ways in which TSI and SAG can expand upon their current relationship.

And Other Travels...
TSI representatives visited sites in the Virginia and the D.C. area in late July. In September we visited several sites in California, Ohio, Florida, New York, Texas, Illinois, and Toronto, Canada. October brought TSI to sites in Philadelphia, Delaware and New Jersey, and in November, we visited Minneapolis and San Francisco. If you would like us to come to your site and discuss your needs, please contact us.

ADABAS/NATURAL Open House Hosted by TSI
In September, TSI held an Open House for all ADABAS/NATURAL professionals in Rosemont, Illinois where guests met our developers, sales staff, members of our support team, and saw product demonstrations. Guests also enjoyed refreshments and participated in drawings for Bears and Bulls sports tickets.

If you would be interested in having a TSI Open House in your area, please contact us to make arrangements.

TSI Celebrates Turkey Day, Then Turkey Week
As announced in the last issue of TREETIPS, TSI welcomes Vizyon Bilgi Islem Ltd. Stl. to our family of international affiliates. Vizyon was established in 1993, and is based in Istanbul, Turkey. Shortly after Thanksgiving, Orhan Ergun, Software Manager for Vizyon, spent a week at TSI headquarters in Sewickley learning about TSI and our line of products, and preparing for sales, marketing and support in Turkey. Orhan informed us that the potential customer base for TSI products in Turkey is approximately six sites on MVS, ten sites on VSE and one UNIX site.

Everyone at TSI enjoyed working and socializing with Orhan, and we all look forward to working with our new partners at Vizyon. Welcome!

Bateleur Takes Marketing Initiative
Congratulations to Charlene McLaren and the entire staff at Bateleur Software Pty Ltd., TSI's South African affiliate. Charlene and the folks at Bateleur took note of TSI's special Year 2000 pricing on our web site with great interest.

After contacting TSI, Bateleur coordinated an advertising campaign that appeared in "Computer Week". Bateleur provided an advertisement and an article entitled "Year 2000 Software Rental Solution", in which Charlene is quoted.

We at TSI take our hats off to Bateleur for putting the time and creative effort into a marketing campaign that supports the Treehouse Year 2000 portfolio of tools.

Follow-up on our Icelandic Visitors
Congratulations to Eric Stannard of Blenheim Software International Ltd., TSI's affiliate in the United Kingdom, on their sale of N2O to the Icelandic Banks Data Centre in Iceland. As we reported in the last issue of TREETIPS, two representatives from the Icelandic Banks Data Centre recently visited TSI headquarters.

Return to the Table of Contents


What Users are Saying About CHART

After having an incredible response from our CHART promotion, we asked some users what they think of TSI's NATURAL analysis and documentation tool. We received some enthusiastic (to say the least) responses about CHART that we would like to share:

"I found CHART so useful, I wrote to five other people to make sure they tried it, and there was unanimous agreement."

"This is a sharp utility."

"I just tried it on-line with a HUGE document program. I think this tool can really help me."

"My first impression-in one word: WOW!"

"I haven't had a chance to experiment with the CHART utility, but I absolutely love the chart you printed out for me! I've been showing it to everyone, and they all have chart envy."

"This is GREAT!!!!! I tried this on-line with our 'problem' seed. Wow! What a great chart!"

Return to the Table of Contents


Off to California to Meet with Oracle

In November, TSI was invited to meet with various members of the Oracle staff. Five representatives from TSI attended, including George Szakach, President; Rich Jacobson, Vice President; Roy Parkinson, Director of Technology; Lori Falbo, Manager of External Relations; and Dan Acheff, DELPHI Technical Development.

The first day was spent with the OS/390 Product Group. The focus of that meeting was the DELPHI Performance Monitor for OS/390. Topics of discussion included the ongoing technical services required to support the DELPHI product as well as future marketing opportunities to the Oracle OS/390 customer base.

On the second day, TSI representatives met with the Design and Migration Technology division to discuss joint marketing opportunities for the tRelational and DPS technologies.

Everyone at TSI was extremely happy with this very productive and successful trip to the West Coast!

Return to the Table of Contents


Converting from NATURAL to COBOL?

If you are planning to convert from NATURAL to COBOL, C++, or any other language, would you be interested in a tool to assist you in the automation of this process?

If your answer is yes, please call TSI at (412) 741-1677, or send us an e-mail at tsi@treehouse.com.

Return to the Table of Contents


TRIM ADASMP Procedure

by Larry Shoup

Several ADASMP sites are using TRIM today to monitor ADABAS.

Batch CLOG (and PLOG, PRESUM, and TRELOG) reporting procedures remain relatively unchanged in an ADASMP environment. Even though the CLOGs related to a database will be generated from multiple Nuclei, the CLOGs can easily be processed in one batch TRIM run, if desired.

The TRIM Real-time Monitor (RTM) deserves some discussion. The RTM basically consists of two parts: assembler user-exits to each ADABAS Nucleus to be monitored, and a single copy (if desired) of the TRIM RTM NATURAL code. The RTM can be directed, via simple NATURAL screen input, to point to a particular database. Therefore, in most non-ADASMP TRIM sites, the RTM is used to monitor multiple databases quite easily.

For purposes of this discussion, the RTM has two distinct functions:

1. reporting on internal workings of the Nucleus of choice; and,

2. reporting on the variety of calls involving a database of choice, which could have had activity from any of multiple of Nuclei.

With ADASMP, there is normally no assurance that a user's first ADABAS call will be directed to any particular ADABAS Nucleus (except updates which always go to the single update Nucleus). A procedure is needed to assure that TRIM RTM calls are, in fact, directed to the ADABAS Nucleus the user intends to view. This would ensure function # 1 above, to be able to choose a Nucleus and ask, "What is happening with the User Queue or the Table of Sequential Commands in this Nucleus?".

To address function # 2, several problems were encountered early in the design of this function.

This function would require that the multiple Nuclei "pool their statistical data" in a centralized collection point for subsequent examination. This pooling would have to be done in dataspaces accessible to all Nuclei, or in an ADABAS file, or in a VSAM file, etc. For example, the RTM user might like to examine the number of calls in the past hour by JOBNAME, where the same job could have had access to the production database via several Nuclei. And the same or another RTM user might like to view the number of calls in the entire session against certain file numbers, where the files could have had calls coming from any of the Nuclei. It would be necessary for the RTM user-exits to update voluminous, pooled statistics on every call to each Nucleus, which might prove to be a performance problem. This concept of pooling data together is actually counter to the ADASMP philosophy of dispersing activity and doing more tasks in parallel in less time.

Another problem is that while some users might prefer to view the combined statistics, others might prefer to view ADABAS call information per Nucleus as it is done today, i. e., not combined. And, there may be a need for additional functions and monitored data about what ADASMP is doing. This would mean even more coding (and testing and documentation) effort to allow many options.

A new RTM for an ADASMP would be a major rewrite at significant cost. The resulting product would have to be overly complicated to try to alleviate any potential performance bottlenecks (which any performance monitor should be solving, not causing). Given the scope of the project, it could not be started immediately, it would not be completed for many months, and it would detract from other TRIM-related change/enhancements. The price to license this product would probably be a 6-digit amount.

The customers we talked to appreciated the design considerations and huge resource commitment involved in a new ADASMP approach, and were quite willing to retain the flavor of the current RTM approach, whereby one Nucleus is monitored with each RTM session, and the total database statistics are available by viewing the statistics for each Nucleus which has accessed the database. This would require only a procedure, rather than redesign and re-coding to make a radical extension to our TRIM product. It would cause no performance problems, and would be available NOW.

One of our customers was kind enough to share their procedure with our TREETIPS readers. The Newell Companies is actively using TRIM in their ADASMP environment. They use ADABAS in a CICS environment. For this procedure involving ADASMP and the TRIM RTM, they set up three ADABAS Nuclei, one for update, and two for read. To be able to run the RTM against each of the three Nuclei, they needed three CICS sessions to be initiated. They also needed to set up a TSO session to be able to submit operator commands to ADASMP.

Their procedure involves isolating each of the three Nuclei and then starting the RTM which would, of course, go against the isolated (active) Nuclei. From that point forward, this RTM session will always make its calls to the same Nucleus. For example, to isolate NUC01:

/F ADASMI,OP01 (ADASMI is the Start of Task Job Name)

/F ADASMI,CL02

/F ADASMI,CL03

This results in Nucleus # 1 being opened, and the other two Nuclei being closed. Nucleus # 1 is isolated. Then, the CICS session is started, and the user logs onto the RTM. The user may enter SNUC to briefly view the "Session Statistics for the Nucleus", which would help to verify that the correct Nucleus (# 1) is being monitored.

Then, the procedure should isolate NUC02:

/F ADASMI,CL01

/F ADASMI,OP02 (Nucleus # 3 is still closed)

This results in Nucleus # 2 being opened, and the other two Nuclei being closed. Nucleus # 2 is isolated. Then, the CICS session is started, and the user logs onto the RTM, again verifying that the correct Nucleus (# 2) is being monitored.

Similarly, the third Nucleus (or as many Nuclei as desired) can be isolated, and an RTM session started to ensure that the RTM will access that Nucleus for as long as desired.

We thank Newell for sharing this procedure with us. We continue to strive for the best solutions in all our products. We look forward to improving our TRIM ADASMP offering as guided by our users and their experiences.

Return to the Table of Contents


PREDICT 3.3 Support for N2O

by Ed Wolfe

In December, TSI announced the release of PREDICT 3.3 support for N2O after working directly with BETA sites, Software AG, and the TSI affiliates to ensure the quality of the release. There is a PREDICT 3.3 support update tape, which requires N2O Version 3.3.3. PREDICT 3.3 support will be integrated into the standard N2O Version 3.4.0 release. Both versions of N2O support PREDICT Versions 3.1, 3.2, and 3.3 objects. However, JCL changes require the use of a separate JCL member for PREDICT 3.3 migrations.

The PREDICT Version 3.3 UNLOAD/LOAD functions no longer work as they did in prior releases. The UNLOAD command of Version 3.2 allowed multiple unload parameters to be stacked, permitting multiple objects to be processed with a single batch job. In Version 3.3, the UNLOAD command terminates (i.e., FINs) after the first set of parameters, forcing a change in the type and format of N2O commands.

The migration process within N2O now utilizes the new EXPORT/IMPORT commands. The objects exported are written to a workfile, which is used by the import process to load the objects into the target FDIC. This implementation is preferrable to using the PREDICT coordinator, since the site may have different coordinator files for each of the FDICs involved in the migration. This method also allows for migrations in which the two FDICs are on different CPUs (i.e., remote migrations).

Changes made to the stack processing of PREDICT also required a change in the way the migration JCL is constructed. The CMSYNIN input for PREDICT 3.1 and 3.2 migrations that were controlled by the site are now controlled by N2O. However, N2O allows the user to specify the NATURAL SECURITY information that should be used for each step. This information is supplied to N2O in CMWKF01 of the N2OSEL step. N2O inserts this information into the CMSYNIN input of the N2OSEND and N2ORECV steps.

The site may wish to create a single User-ID used by all N2O PREDICT 3.3 migrations or utilize the NATURAL SECURITY auto-logon feature. This will prevent the complication of determining what logon ID, password, and application to use during the BATCH migration process. These parameters are optional. If the site is not using NATURAL SECURITY or is using the NATURAL SECURITY auto-logon feature, this information is not required.

In addition to the NATURAL SECURITY information, the site may specify the library onto which N2O should logon during the N2OSEND and N2ORECV steps. This is accomplished by utilizing the new replacement variable (&LOGONLIB). While building the CMSYNIN input for the N2OSEND and N2ORECV steps, N2O will place the value specified after &LOGONLIB= in the stream as "LOGON library". This feature is for sites that do not allow users to have logon authority to the SYSTEM or N2OLIB libraries. This parameter is optional, and if not specified, N2O will default this parameter to SYSTEM.

Return to the Table of Contents


Got Support?

by Katya Davis and Marni Harbutz

As mentioned in TREETIPS #23, we have several new faces in the Support Center at TSI. Because we support over 20 products, it is sometimes challenging to provide an immediate response to a customer problem. We have spent a lot of time learning about the many products marketed and developed at TSI and would like to take this opportunity to make some suggestions that we feel will help us to be more accurate and efficient in responding to your questions or problems.

We have compiled the following checklist to help our customers be prepared when notifying the Support Center of a problem:

Please be sure the information you provide is as detailed as possible. We need to know when the problem occurs, how the problem manifests itself, if any action seems to make the problem disappear, etc. By having this information available, we should be able to resolve your problem more quickly.

"I'm very pleased with our support effort." says Richard Jacobson, Vice President of TSI. He adds, "We have three people dedicated to support full time, and four people dedicating part of their time to supporting our products."

We are more than happy to provide any necessary zaps and fix tapes, free of charge, to our customers under maintenance. This standard service is called the TREEQUEST support program.

Support is available 24 hours a day, seven days a week.

Return to the Table of Contents


What's New in tRelational Version 2.0.0

tRelational (V2.0.0) is TSI's ADABAS C data analysis tool which analyzes ADABAS C file structures, data, and usage to aid in improving design and performance of ADABAS C databases. tRelational also assists with logical modeling and mapping of ADABAS C data into a relational format for use with a relational database management system (RDBMS). tRelational has been upgraded to be compatible with NATURAL Version 2.2, and with version 2.0.0 of the Treehouse Software Data Propagation System (DPS). tRelational will generate all necessary parameters and mappings required by DPS for initial materialization of ADABAS C data and for propagation of all ADABAS C update data.

The following is a list of some of the changes, features, and enhancements provided with tRelational V2.0.0:

Support of DPS V2, including specification of ADABAS version and DBID (for extraction of data from multiple sources).

Repeating Field and VARCHAR analysis now allows up to 926 fields to be analyzed per file.

Descriptor Analysis now analyzes up to 200 descriptors per file.

VARCHAR analysis now includes Long Alpha and 0 length alphanumeric fields. The entire length of each field is analyzed, where previously only the first 40 characters of alphanumeric fields were analyzed.

Support for DB2 "Not Null with Defaults" and "Not Null Unique" is provided.

tRelational Version 2.0.0 has been developed under NATURAL Version 2.2 and now requires NATURAL 2.2.5 or higher. Previous versions of tRelational, which required NATURAL Version 2.1.6 or higher, will no longer be supported.

Initial migration (materialization) of ADABAS C files is no longer accomplished within tRelational, but has been moved into DPS. tRelational generates the parameters required by DPS in directing the extraction and transformation of ADABAS C data for both materialization and subsequent propagation of update data.

Return to the Table of Contents


New Versions of SEDIT and S/REXX
Available for UNIX, and now Windows

Would you like to have the familiarity of the mainframe environment on your UNIX and Windows systems? SEDIT and S/REXX may be your answer. TSI is announcing the release of SEDIT V4.20 and S/REXX V2.20 for UNIX and Windows.

SEDIT is an editor which emulates the mainframe XEDIT and PDF editors, providing users with a familiar editing environment.

S/REXX is a macro language which emulates the IBM mainframe REXX language.

The S/REXX Debugger, is also available. It is a GUI-based debugging environment for the S/REXX macro language. The S/REXX Debugger allows the user to step through the execution of S/REXX programs, locating and fixing bugs.

If you would like more information on SEDIT and S/REXX, call TSI today, or visit our web site at http://www.treehouse.com.

Return to the Table of Contents


Disclaimer: All product names mentioned in TREETIPS are trademarks and/or products of their respective holders. The mention of any non-TSI product in TREETIPS should not be considered to imply support or endorsement by Treehouse Software, Inc., its employees, or affiliates.

Return to the Table of Contents



Treetips | Home