Web Enabling the Enterprise WITHOUT Reprogramming

Oracle Representatives Visit the Treehouse

Editor's Sproutings

Bully for the Mainframe

Manual Labor

tRelational Goes GUI

SECURITRE Verification Programs

New NATURAL Features in TRIM

tRelational and DPS Sales on the Rise!

If It Were a Perfect World...

T1 Connection Increases Speed and Expands Services at treehouse.com

Operations News

TSI Education and Consulting Update

Do you want to add dynamic content to your organization’s Web site?

Do you want to provide a common, browser-based user interface to a client/server, midrange, and/or mainframe system?

You CAN, because Treehouse Software, Inc. (TSI) is now marketing an exciting product called ForeSite™ which allows organizations to Web-enable mainframe, midrange, and client/server systems without re-programming. This advanced integration server and deployment environment for the Web has been developed by InfoSpinner, Inc. of Richardson, Texas. And through a partnership agreement with InfoSpinner, TSI will now provide sales, support, development, design consulting, and training for ForeSite.

What is ForeSite, and how does it work?

ForeSite is the most advanced Web integration server and deployment environment available on the market today. It has been designed to address many of the usability, extensibility, and scalability problems associated with creating sophisticated, high-volume data access and Web commerce applications. By utilizing Web technology, ForeSite breaks new ground, providing an open architecture that assures simple, rapid, and reliable delivery of Web-enabled applications.

One of the key features of ForeSite is its patent-pending “clustering” technology which is implemented in two separate deployment modules: Dispatcher and PageServer™. These modules are based on a distributed, multithreaded architecture that has been designed to maximize performance, throughput, and availability of applications. ForeSite also includes Integrator, a powerful and easy to use integration module that simplifies the integration of applications with the Web. The dynamic content stored in applications and databases can be easily Web-enabled when using this integration module, and since it is compatible with most popular HTML editors and authoring tools, HTML pages can simply be dragged-n-dropped into Integrator to integrate applications and databases.

In addition, ForeSite supports a broad range of data sources and application types through its unique System Integration Module (SIM) data sourcing architecture. This architecture includes several pre-built System Integration Modules (SIMs) for accessing common data sources, such as 3270 applications, ActiveX servers and controls, Java classes, and ODBC databases. Any additional data sources may be accessed through the (Component Object Module) COM-based SIM Developer’s Kit (SDK).

ForeSite can integrate with any Web server that supports either the Microsoft Internet Server Application Programming Interface (ISAPI), the Netscape Server Application Programming Interface (NSAPI), the Common Gateway Interface (CGI), or the Novell Web Server.

Call us at (412) 741-1677 to set up a demonstration of ForeSite or to learn more about how ForeSite can Web-enable your enterprise systems.

TSI and Oracle Alliance
A consulting and distribution business alliance has been established between TSI and Oracle Corporation.

Under this alliance, a license and distribution agreement enables Oracle consultants to use TSI’s ORACLE Performance Monitor, DELPHI, during site visits. In addition, Oracle Corporation has been granted permission to distribute trial copies of DELPHI to current and prospective ORACLE for OS/390 customers. Currently, a trial of DELPHI ships with every ORACLE for OS/390 order.

DELPHI is a Real-Time Performance Monitor which provides comprehensive statistical and accounting data about ORACLE performance, supplements the ORACLE SQL*DBA Utility, and gives database administrators and system programmers concise information about the performance of their database system.

Refer to the article entitled Oracle Representatives Visit the Treehouse.

TSI Expands Management Team With Aquisition of InfoCatalyst, Inc.

TSI is pleased to announce that InfoCatalyst, Inc. of Houston, Texas has been acquired and that its founder, Tom Berrisford, is the newest member of the TSI Management Team. When asked why he joined TSI, Tom replied, “I have been working to build InfoCatalyst into the kind of company that George has already built with Treehouse. Everyone I talked with was very impressed with the TSI people and the support TSI provides to its customers. I decided to join the existing TSI team rather than build my own over time."

Tom’s initial duties will include serving as the ForeSite™ project manager, and adding ForeSite to TSI’s portfolio of products by assisting with training, documentation, sales, marketing, and product support (see cover story to learn more about ForeSite).

George Szakach, President of TSI stated “It is really great to have Tom on board. I have known him for over 23 years. He brings 30 years worth of technical and marketing savvy to the Treehouse. As Director of Business Development he will be exploring relationships with partnering companies, researching the acquisition of new technologies, and assisting in the controlled growth of TSI staffing. In addition, Tom will help ensure that our burgeoning consulting opportunities are properly handled.”

TSI Is Proud to Add Steve Robinson to Our Team . . .

We are delighted to welcome Steve Robinson of S.L.R. and Associates as a part-time consultant. With more than 25 years of experience in technical writing and SAG software education, Steve is an ideal addition to our consulting team.

Steve will contribute to TSI’s training and documentation efforts. He will compose product training materials, upgrade our marketing and technical documentation, and serve as a guest columnist for our company Web site and newsletter.

We also plan to utilize Steve’s background in software education. In this capacity, he will teach TSI product related courses and instruct other staff members in software education procedures. In addition, he will serve as a representative of TSI. He will be speaking at selected customer sites and attend SAGGROUP meetings or SAG conferences on our behalf.

Steve will get involved with our newest offering GENART for NATURAL, an automatic remediation tool for Y2K compliance, by providing interesting articles, insightful documentation, and detailed training materials.

Other New Faces Added to TSI’s Growing Family

Dan Sycalik was hired as a senior consultant. Dan’s initial focus is in the areas of data migration, propagation, and warehousing. He has already become a tRelational and DPS expert, and is handling customer demos, presentations, installs, and training. If you would like to discuss these products from a technical standpoint, call and ask for Dan.

Mike Kuechenberg joined TSI as a sales-technical representative. Mike has considerable experience with the Software AG product line, and is familiar with TSI's SECURITRE, N2O, tRelational, DPS, and more. He will be assisting the Sales Department with installs, demos, presentations, training, and consulting.

Marisa Dohanich recently joined TSI, working with Katya Davis who is heading up our Technical Support Department. Kara Phillips recently joined TSI as a technical writer, working with Joe Brady, Manager of Technical and Marketing Documentation (see Kara’s article on page 4). Michael Szakach has been “hanging around the Treehouse” for years, and has assisted with documentation, operations, and sales. Recently Mike became a full-time employee.

The TSI Y2K Statement Still Stands

TSI is purposely keeping the original date of September, 1997 on our Y2K statement which is posted on our Web site. Periodically, we will update the introductory message.

As of November 1998, we note that any updates for “Y2K compliance” for operating systems, TP systems, database systems, and languages have not affected our software adversely. Therefore, our 1997 statement stands unaltered.

As a minor update, we acknowledge that many businesses, including government agencies and universities, have created Y2K compliance rules of sort. This is excellent. However, these rules are sometimes hazy or they conflict from one customer to another. Compliance for one customer may cause noncompliance with another. Compliance has a vague meaning when it comes to a problem with multiple, sometimes complex, “solutions.” We have seen many statements, rules, and demands. Still, we stand by our robust, meaningful statement of being “as compliant as possible at this time.”

Product Releases/Enhancements

• ADAREORG V2.3.0 and ADASTRIP V3.0.0 have been released. These versions provide Y2K functionality and support for ADABAS V6.2.

• DELPHI V2.4.0 has been released. If you are a current DELPHI customer, contact TSI to upgrade.

• DPS V3.0 is scheduled for release in the first quarter of 1999. This new version will offer many enhancements, including support for the Microsoft SQL Server and DB2 as target RDBMSs, VSE support, new built in transformations, automatic retrieval of the ADABAS FDT from ADABAS, compare of ADABAS FDT produced by tRelational with FDT stored in ADABAS, expanded statistics by ADABAS File and Database-ID, sample external transformation routines for common transformations, and a utility for analysis of transaction arrival rates.

• N2O V3.4.1, is now available. This new version of N2O incorporates the N2OSCAN feature and the powerful Documentation Tools. This version offers PREDICT V3.3 support.

• PEEK V2.2.2 has been released, and is fully compatible with ADABAS V6.2 and NATURAL V2.3 and above. PEEK V2.2.2 documentation is now available in electronic HTML format as a series of MS-DOS files.

• PROFILER for NATURAL V4.0.0 has been released.

• RACE V1.4 includes a greatly improved percentage of eliminated RCs, optional use of Multifetch, and handling of files with up to the maximum number of ADABAS fields.

• SECURITRE V3.1.0 has been released with full compatibility with the latest versions of ADABAS and NATURAL.

• tRelational V3.0.0 has been released.

• TRIM V6.1.0 now supports NATURAL V2.3. And TRIM V6.2.0 is soon to be released.

Hard Copy, CD, or Both?

Want a free TSI Documentation CD? TSI has begun sending out all of our product manuals, product overviews, and fact sheets on a single CD to anyone who requests one. TSI customers and potential customers have the option of receiving their documentation as hard copy, CD, or both. Any individual receiving our documentation is permitted to make as many copies of it as they wish -- Free!

Many companies charge $50 to $1000 for their documentation CDs. We at TSI believe that product documentation is a great marketing tool that should be used as such. What better way to see what a product is all about than reading the manual? So here they are, all on one CD, in portable document format (pdf). Contact TSI to get yours today. If you already have a TSI Documentation CD, you may not have the most current. Give us a call and tell us the date on the CD, and we'll let you know if a newer one has been released.

User Group Meetings

Steve Robinson was invited to speak at the SAGA NE Regional User Group meeting in October. He spoke about Y2K manual assessment and remediation versus automated assessment, and remediation using GENART examples.

TSI has extended SAG User Group Meeting support through 1999! Call us to find out about TSI’s offer to financially sponsor your 1999 User Group meeting.

Fun at the 2nd Annual TSI Golf Outing

It was a “scramble”, and who better to scramble a scramble than TSI's staff.

This past summer, TSI held its second annual Golf Outing. Among some of the sights were people teeing up on the fairway, carts on the green, and the ball mysteriously reappearing on the tee after a drive.

When the dust (and sand) cleared, the team captained by Tom Ray walked away with the honors, and the rest of us are still wondering exactly what a scramble is.

16th Annual Ho Ho Holiday Party

On December 13th, TSI had its annual holiday party. Among the good food and festivities was the TSI Who’s Who Contest, where baby photos of TSI staff members were mounted on a board and everyone tried to guess which gray-haired employees matched the precious little baby faces.

Letter to the Editor from the President

As a Software AG (SAG) customer and partner, TSI is very interested in any discourse we see about SAG and their products. This includes the Software AG discussion list (SAG-L) hosted by the University of Arkansas. We think the University does an excellent job of coordinating, and sometimes policing, the SAG-L correspondence.

All SAG customer sites should be “listening in” to SAG-L. At TSI, we print the SAG-L discussions daily (usually about 50-100 pages) and route it to all staff.

SAG-L discussions are generally informative and interesting, but at times they can be annoying or disheartening. The interaction is typically good, but we note that only a small percentage of SAG sites correspond on SAG-L. In some cases, there is conversation of a self-serving nature along with blatant marketing, job searching and recruiting, and other no-no’s.

TSI doesn't respond as often or as robustly as we could, instead keeping our responses short and to the point, and as non-marketing oriented as possible. As a customer we try to make our responses “fit” with correspondence from others, and we are careful not to cross the line into “vendor talk.”

Regarding the numerous Y2K related questions and comments we see on SAG-L, for example, we certainly could say: “We have GENART, no more discussion necessary.” On the correspondence related to data movement from ADABAS to anywhere, we could say: “We have been promoting tRelational and DPS as your data migration, propagation, and warehousing solution for several years. There simply is no other solution.” However, we limit our vendor talk to TREETIPS and our Web site.

The comments blasting SAG, their support, and their direction, along with questioning the future of ADABAS, NATURAL, and SAG themselves have been somewhat annoying. Maybe the rule should be that you can knock SAG if, and only if, you own a piece of them. How many SAG-L respondents or their employing companies support SAG by owning SAG stock?

George Szakach

For decades the mainframe computer has symbolized the power and speed behind major corporations by providing centralized processing, presentation, and data storage. However, as computers continued to evolve, a dark cloud hovered over the mainframe computer industry. As you know, the microcomputer or PC and client/server platforms were major threats to the survival of the mainframe, and the image of the mainframe grew faint during the 1980s and '90s.

In our September 1997 issue of TREETIPS, an article entitled “Vitality in use of Mainframes” reported findings from an IBM S/390 Market Analysis Department survey report. Despite current trends, the article had accurately predicted the continued vitality in the use of the mainframe computer platform, and anticipated the continued presence of the mainframe during the new millennium.

Compared to other platforms, mainframes have been perceived as costly, slow, unreliable, and bureaucratic. Yet, the centralized computing power behind mainframes continues to survive the test of time and technology. One might wonder how these “technological dinosaurs,” which were once thought to be on the verge of extinction are now thriving in companies throughout the world.

Actually, several factors have contributed to the revitalization of mainframes. These include new technologies, such as the Complementary Metallic Oxide Semiconductor (CMOS), which has enabled mainframe vendors to create smaller and cheaper machines. CMOS-powered machines typically cost 10 times less to run than traditional mainframes, and air-cooled units require 90% less power than their water-cooled predecessors do.

In addition, the struggle to contend with high-end midrange servers, and the price wars between mainframe manufacturers have resulted in competitive pricing, significantly reducing annual revenues for mainframe manufacturers. In fact, recent market research studies have concluded that running a large microcomputer-based network can cost up to three times more than supporting the same number of users on a mainframe.

Although microcomputers, mainframes, and client/server platforms continue to evolve and share similar characteristics, it seems that each will thrive as it serves its own purpose in the industry, now and beyond the year 2000.

Recently, while visiting TSI, I had an opportunity to sit in on a GENART presentation by Alison Ogier and Brian James from Generation Systems Limited (GSL), the developers of GENART, and an excellent learning experience was had by all (especially me). I will share some of my observations with you in a moment.

Quite apart from my association with TSI, I received a call recently from a company where I have done some NATURAL education. They were interested in my views concerning the Y2K problem. I first pointed out my bias to them. Having just finished a GENART tutorial, and close on the heels of my association with TSI, I had to indicate that I was not an impartial source of information.

Nonetheless, the discussion went quite well. One of their concerns was budget. Not surprising. Many shops have met with quite a bit of resistance from the “budget people” when seeking to acquire Y2K remediation software.

The other alternative this shop was asked to investigate was manual remediation. Those of you who read Inside Natural are familiar with my aversion to doing things the hard way. If I can save a few keystrokes for common tasks, I do. Manual remediation is definitely doing something the hard way, and it is also the expensive way.

I have asked around a bit, and played around a bit myself, regarding the time required for manual remediation. The general consensus seems to be about 400-800 lines of code per day. My own personal estimate is around 500 lines of code. Now let me clarify my estimate a bit. This obviously does not include comments. It also does not include DEFINE DATA clauses. Why?

A tangent

Suppose I see a variable in a DEFINE DATA clause named #DATE-POSTED. Does this variable require remediation? Suppose it is formatted as A8. Does it require remediation?

Before you say no, because it is A8 and not A6, realize it might contain MM/DD/YY, and the two slashes bring the character total to eight, not the two characters for the century. So does it require remediation? Where will I find the information I need? IN THE CODE. So my 500 per day might be someone else’s 600 or even 1000 lines per day.

Don’t believe the last number? I have seen higher ratios of comments and definitions to code. Quite a few years ago I coined the term "NATBOL" to refer to COBOL code written in NATURAL. This is typically the result of giving a COBOL programmer a NATURAL manual and telling them to “go forth and write NATURAL code.” One characteristic of such code is huge LDAs. A record will be read into a view, say with 40 fields. Then there is either a MOVE BY NAME or individual MOVEs to put the data into a different set of 40 fields that will be displayed on a Map. After the INPUT statement, the data is moved into another 40 fields for processing. Then into another 40 fields for updating. One hundred and sixty fields for what could probably have been done with a single view with 40 fields. No wonder I have seen LDAs go on for 20 screens or more.

Back to the main thread

Okay, now you know what I mean when I say 500 lines of code per day for code remediation. If they have about a million lines of code, which approach is cheaper? I am not the definitive source for pricing, but I believe GENART would therefore be priced at about $100,000 (ed. Pricing per line of code is reduced as the volume increases).

One million lines of code, remediated at 500 lines per day, yields 2000 person days. Lets assume that your applications (and available staff) are such that you can remediate code at twice the rate I assumed. You are still looking at 1000 days of effort, which is about four years of billable effort. Take whatever figure you use for the “cost” of an employee for a year and multiply by four (most shops take an average salary and multiply by 2.25 to derive such a “cost”). You are not done yet. GENART assisted remediation requires more than the cost of the software. For a million lines of code you will probably require two people for two to three months to accomplish your remediation. I would be willing to bet that the manual approach “prices out” at least two to three times as expensive as GENART assisted remediation.

Not figured into this analysis, but certainly something to consider, is accuracy. GENART is quite accurate, and can be enhanced by appropriate use of what are called “search criteria” to be very accurate. What is very accurate? Far more accurate than what would be achieved manually.

How can I make such a blatant statement? Basically, GENART duplicates manual procedures. How do you remediate manually? You start by SCANing for “obvious” strings like *DAT, DATE, and (D). This gives you a first cut at your problem. Then you add to the list new items you found (e.g., using a MOVE EDITED *DATX (EM=MMDDYY) to #HOLDER). Now you SCAN for the new items, ad nauseum. Besides its many other remediation facilities, GENART does this. Guess which approach is more appropriate? Letting software do the work, or a programmer, who is bored by the process and anxious to get home? With GENART, programmers (or analysts) get to do more creative work, they are far less likely to get bored, and the end result is far more accurate.

So, when compared to the manual process, GENART produces a more accurate analysis for less money. In the vernacular of today’s youth, “Duh, it's a no brainer dude!” (Note my fluency with the lingo. I find it useful when trying to converse with younger members of the family who do not speak English.)

Okay, enough for now. Next time I will endeavor to demonstrate how GENART basically duplicates what you would do manually.

The recently released tRelational V3.0.0 offers new features that will not only make the task of data warehouse design possible, but (dare we say?) fun.

One of the new features added to this data modeling and mapping tool is tRelationalPC, a Graphical User Interface (GUI) that allows users to see their entire Relational Database Management System (RDBMS) data model in an instantly familiar, ERD-like layout on their desktop Windows 95/NT PC.

Drag-n-drop Functionality

tRelationalPC offers drag-n-drop functionality which permits easy manipulation and editing of all components of the data model, including tables, columns, indexes, and foreign keys.

Automatic ADABAS-to-RDBMS Mapping

ADABAS data elements may be dragged-n-dropped from the left pane into the data model, automatically creating RDBMS structures which are “mapped” from the originating ADABAS data elements. This ADABAS-to-RDBMS mapping metadata is what tRelational passes to DPS, to instruct it to extract/transform ADABAS data for migration to the receiving RDBMS data store or warehouse.
Interface to the Mainframe

tRelationalPC downloads implemented ADABAS file structures and RDBMS data models from the mainframe through Treehouse Remote Access (TRA) middleware that is provided free of charge to all tRelational sites (TRA requires standard TCP/IP connectivity between the mainframe and PC).

For more detailed information on tRelational and DPS, visit our Web site or contact us at (412)741-1677.

What are tRelational and DPS?

tRelational and its sister product DPS are the two components of TSI’s solution to a problem that more and more sites are facing, extracting data out of legacy ADABAS files and into RDBMS-based data stores, data warehouses, or packaged applications, and (perhaps more importantly) keeping the data in RDBMS tables synchronized with the data in ADABAS files.

The Traditional Approach

Many institutions have tried the traditional “homegrown” approach, which typically consists of extracting their data directly out of ADABAS. However, this process is lengthy, and it requires teams of programmers to write NATURAL-based extraction and transformation programs, complete with extensive rules for extracting specific logical files out of larger physical ADABAS files. These extraction programs must then be run as part of their nightly batch processing. A subsequent synchronization of RDBMS data with altered ADABAS data requires a complete rerun of the extraction process.

These institutions are consequently faced with a two-pronged problem. They have incurred a costly development, debugging, maintenance, and enhancement cycle built into their NATURAL-based data extraction and transformation system; and as the needs of their data warehouse unavoidably grow and change, they are faced with an ever-shrinking “window” of nightly batch processing time to complete this cycle. One site told us that a proposed homegrown data extraction solution would have required 14 days of processing time to complete a single cycle.

The tRelational/DPS Approach Saves Time

tRelational and DPS allow a site to quickly and easily map ADABAS data structures to RDBMS data structures, and then materialize (initially load) and propagate (subsequently keep synchronized) the ADABAS data into the RDBMS without requiring direct access to ADABAS.

At a recent tRelational/DPS trial, TSI consultants were told that these tools helped the site complete a process in two days that had originally taken two years of development to accomplish manually using a similar homegrown process (Figure three full-time consultants over a two-year period at their hourly rates. . . you do the math!).

We are always looking for ways to provide better customer support in less time. However, it is often difficult to determine exactly what is causing a product to function improperly in a user’s environment, since we are unable to see the particular components, including parameters, datasets, load modules, JCL, and CLISTs at the customer site.

We have found that most problems are due to either the parameters being loaded incorrectly or the components not being installed correctly. When the customer calls our support center to ask for our help, it is sometimes difficult to determine the source of the problem. Consequently, we are in the process of developing a solution to this support problem by developing verification programs for SECURITRE, our security interface package for ADABAS, ADABAS Utilities, and NATURAL.

These programs will offer our customer sites assistance in the diagnosis of installation problems, and in the event that assistance from TSI support is required, the information provided will help our staff resolve the problem in a more timely manner.

Once the installation of SECURITRE is complete, the site may execute the utility security and NATURAL and ADABAS file security verification programs which will verify the proper installation of SECURITRE User-Exits 1, 4, and B; SECURITRE for NATURAL; and SECURITRE Utility Security.

In addition, the programs will report the datasets from which the individual modules/parameters were loaded, since these datasets are often concatenated by log-on procedures and substitution JCL which can cause the SECURITRE parameters or user-exits to be loaded from an unexpected location. This causes errors due to the use of old modules, “vanilla” modules (modules without zaps applied), and improper parameters. With this tool, a user will be able to verify that the module that they expected to use is actually in use.

To verify the installation of SECURITRE’s ADABAS file and NATURAL security, a NATURAL program (STRVRFY) is executed. This program generates a single call that all components of SECURITRE will recognize. Then, each component will populate the return buffer with the appropriate information, and STRVRFY will check the return buffer for valid information. If any piece of SECURITRE is not installed or installed incorrectly, the user will be informed with detailed messages.

Since all of the components are verified at the same time, the user is able to identify and correct all of the problems without re-executing STRVRFY. This program also reports the installation status of the SECURITRE RTM, log-on security, and STEPLIB support. However, the verification program is only able to determine the information in the NATURAL nucleus that it is executing. Therefore, the program should be run on all NATURAL nuclei where SECURITRE for NATURAL is installed.

The second verification program is executed in batch to verify proper installation of SECURITRE for ADABAS utilities, since all ADABAS utilities are executed in batch. This program verifies that the current version is installed in ADARUN, and that the parameters can be loaded properly by loading the first ADARUN module found in the dataset concatenation. The information reported includes the sold/trial status of SECURITRE and the dataset where ADARUN was found.

Both SECURITRE verification programs display dataset information about the various components that are loaded when NATURAL, ADABAS, or SECURITRE are started. This is determined in the same fashion used by OS/390 itself when loading a module. First, the Job Pack Area (JPA) is searched using the standard search path: STEPLIB, JOBLIB, and LINKLIST. If the module is not found in the JPA, the following areas are searched in this order: Fixed Link Pack Area (FLPA), Pageable Link Pack Area (PLPA), Modified Link Pack Area (MPLA), and Common System Area (CSA). If the module was loaded from STEPLIB or JOBLIB, the dataset name is given. Otherwise, the area and dataset number is given.

This type of verification will help the customer and TSI to resolve problems quickly and more efficiently. These verification programs will be included in future versions of SECURITRE, and we plan to incorporate similar capabilities into other TSI products.

TRIM V6.2.0 provides access to information from the NATURAL nucleus in use during a current TRIM Real-time Monitor (RTM) session. This function makes another level of environment information available to the Database Administrator (DBA) using the TRIM RTM.

This information is useful when attempting to identify possible product conflict conditions, maintenance issues, and other interaction considerations. In addition, information about ADABAS and the TRIM product environment is available using the TRIM RTM.

The TRIM RTM can now be used to perform some of the NATURAL Buffer Pool management functions that are useful during peak performance periods.

For example, the RTM can be used to display up to 200 objects from the NATURAL Buffer Pool.

If an ‘I’ is entered next to an object on the previous screen, the screen below is displayed.

Additional information may be displayed on this object by pressing the PF9 key.

Although this NATURAL Buffer Pool Information has been available using other sources, it is now conveniently accessible through the RTM by selecting the new screens to view the data.

Note that LumeNAT has been removed from the TSI list of products. The LumeNAT functionality (presented initially in TREETIPS Issue #21) has not disappeared. We placed the total LumeNAT functionality into TRIM, renaming it the TRIM NATURAL Monitor (TNM).

With TNM, NATURAL activity is monitored and stored as summary data with certain optional detail data also being captured. Summary reports offer comprehensive views of system activity (e.g., ADABAS calls, program executions, response times) broken down by program and/or by user. Detail reports show program relationships in a nested format, along with statistics about the most heavily used programs, programs making the most ADABAS calls, and programs with the best/worst response times. The User Trace Report (shown below) displays events resulting from user activity.

The TNM User Trace Report offers extensive user/program information: ADABAS call start, ADABAS call end, program start, program end, and terminal I/O. Program execution flow, including elapsed time, is clearly displayed.

TSI has recorded record sales this past year! Our ADABAS and NATURAL customers have played a significant part realizing that our 20+ ADABAS and NATURAL products are second to none in the industry. With the TSI product suite being continually enhanced by our development staff, the addition of a few new products from around the globe, and our focus on consulting services I predict another record year for 1999.

Particularly on the rise are tRelational and DPS, TSI’s answer to your data warehousing and data migration needs. With the increased use of ERP packages from Oracle, SAP, Peoplesoft, etc., and the increased use of RDBMS data warehouses, customers are looking for an efficient way to migrate their existing ADABAS data to these platforms. The benefit of not having to write NATURAL extract programs to get to existing ADABAS data and the time saving aspect of being able to propagate ADABAS data instead of doing total refreshing has sparked the interest of many sites. Many sites today don’t have the resources to accomplish a migration or a data warehouse project and are looking for an out of the box solution. Call TSI for the box. Inside it is tRelational and DPS – the solution!

Many sites also have special needs, such as near real-time propagation, unique ADABAS file structure situations (e.g., hundreds of files with identical structures), and data warehousing to several RDBMS platforms (even in a VSE environment) that we are able to address with help from our migration and data warehousing consulting division.

SAGA and SAGD have both visited TSI and have seen tRelational and DPS in action. They were extremely impressed with these products. Oracle has also recently visited TSI to discuss incorporating the products into their ADABAS to Oracle migration strategy (see article on page 1). Oracle has visited several government ADABAS sites recently and proposed the use of our products to assist in Oracle data warehouse projects or migration to the Oracle platform. Also, several sites are trialing ERWIN (from PLATINUM Technologies) and the ADABAS tRelational interface.

Since the start of tRelational development in 1994, we have had tremendous success with the sites that have trialed the products or have done a pilot project with us. We have yet to have a site reject tRelational or DPS due to them not doing what we claim they do. Many sites have tried various products from vendors that claimed that they handled ADABAS data propagation, but they have eventually come to TSI for the solution that works!

N2O, PROFILER for NATURAL and GENART also played key roles in our success for 1998. The Y2K effort resulted in a significant increase in sales of these products. Customers realized that they needed N2O to manage their change control, GENART for the remediation of their NATURAL code and PROFILER to ensure that they have thoroughly tested their applications before putting them back into production.

TSI met with many different companies that wanted us to market their so-called Y2K NATURAL solution. After rejecting these “solutions” we came upon GENART from GSL. TSI goes through considerable effort evaluating products that we are considering marketing. The fact that SAGD was selling GENART throughout Europe made are decision for GENART a lot easier. Recently GSL has added GENADA and GENTEST to the “GENSET” suite of products. Now sites wishing to have a total Y2K solution from remediation to data aging and test scripting can contact TSI for the solution.

We owe a lot to our customers and we will continue to take care of you in the future. You've seen very little rise in product prices, while our costs have risen dramatically. You have a vendor in TSI with 15 percent of its staff involved in technical support. The industry standard is 2 percent.

Our documentation is available to you for free, and now on CDs. We continue to support your regional or local User Group meetings, CAUCUS, etc. If you would like to have a meeting of SAG users in your area, we can help you set this up with an appropriate agenda, speakers, materials, equipment, meeting room, announcements, and the all-important meals. We take care of all expenses. Contact us at tsi@treehouse.com or (412)741-1677.

Have you ever spent a frustrating day and a half tracking a production-halting problem only to discover that someone made an unannounced change to a loaded table that caused an application to run incorrectly?

Maybe you’ve experienced a mind jarring call at 3:00 a.m. from your DBA telling you that an application that you are responsible for is exceeding its share of ADABAS resources, and that you better do something about it!

Or perhaps you missed a strategic project deadline (and a healthy bonus), because your programmers neglected to properly test all aspects of an application and quality assurance sent it back with a dreaded stamp of disapproval.

You ponder . . .

“If it were a perfect world, I would easily track troublesome bugs and inefficiencies in my NATURAL code by utilizing a slick interface that could pinpoint where the potential problems lie, and I would possess powerful tools that could instantly signal whether all phases of an application have been properly tested or not. Even finding and correcting Y2K problems would be a snap, because I would know exactly how my applications work. And best of all, I’d never receive another mind jarring phone call from my DBA at 3:00 a.m.!”

Well folks, your perfect world has arrived.

TSI is proud to announce the release of PROFILER for NATURAL V4.0.0, a newly improved application profiler that is suitable for the new millennium.

The original menu-driven system of PROFILER has been completely revamped into a user-friendly tool with easy to navigate pop-ups (see the sample screen below) that instantly provide the user with the information needed to track bugs and problem programs, and pinpoint inefficient code.

PROFILER illustrates exactly which instructions in which objects have been executed; and the order in which they have been executed when using the trace feature. PROFILER graphically projects CPU usage by object and object type, and instantly signals which objects are responsible for excessive consumption of ADABAS resources.

PROFILER’s Enhanced Reporting feature allows the user to either run enhanced reports, or off-load collected data into a separate reporting file to write unlimited customized reports. In addition, PROFILER can be run in background mode, providing project leaders with the flexibility to monitor application development and testing from their own desktop, whether their programmers are using PROFILER or not.

PROFILER for NATURAL V4.0.0 is a Y2K compliant, easy to install, user-friendly profiling solution that runs under NATURAL V2.2 or V2.3 and is compatible with ADABAS V5 and V6. So stop pulling your hair out, and call TSI today to ask for a demo of PROFILER. If you are currently using an older version of PROFILER, call and ask for an upgrade. We're sure you'll love living in a "perfect" world.

A New User Interface

The PROFILER user interface has been revised with V4.0.0. A “session list” and “action list” style of interface has been developed.

With Web site statistics climbing to more than 1000 hits per week, we decided to host our site in-house and upgrade our internet connection to a T1.

Our new server has enabled us to substantially increase the speed of our Web site and expand our storage capacity to more than 27 Gb.
This expansion will allow us to add many new features, including an on-line TSI support page which will provide customers with the following services:

• A form (right) for special customer information which can help Support representatives to better diagnose problems, including a list of questions (e.g., What platform are you on?, What is the name and version of the product?, and What is the level of priority for the problem?). There will also be a space to include screenshots or other information that is relevant to the problem

• A knowledge base of “How-to” documents and FAQs

• Notices:
- An option that will allow customers to be automatically notified of product upgrades, early warnings, and zaps by e-mail
- Early warnings
- All available Zaps with descriptions of the problems resolved by each
- Any product information that could be helpful to the customer, including compatibility, Y2K, and OS/390 issues, as well as NATxx, ADABASxx, and PREDICTxx news

Visit our Web site for updates on the status of our support page and other enhancements to our site.

We are moving full speed ahead into the 21st Century at TSI by utilizing many new advances in technology. In fact, we have recently completed migration from an IBM 9370 running VM/IS to a Multiprise 2003 (Model 203).

The features of our Multiprise 2003 include LPAR capability, a CMOS processor that runs at 6 million instructions per second (MIPS), 256 Mb of main memory, and 72 Gb in our internal direct access storage device (DASD) which is set up with Raid 1, yielding 36Gb of usable mirrored disk storage.

IBM's OS/390 V2.4 and VM/ESA V2.2 are currently installed on the system. However, we are already looking into upgrading to IBM's recently released OS/390 V2.6. In addition, CICS V4.1 has been installed, as well as the latest versions of ADABAS and NATURAL. We are in the process of bringing up NATURAL V2.3.1 under CICS, and we are completing the installation and configuration of DB2.

The 2003 has an onboard OSA card for connectivity to our Local Area Network (LAN), using TCP/IP or SNA, and an internal uninterruptible power supply (UPS) system that will keep the 2003 powered for approximately 1.5 hours in the event of a power failure.

VTAM and TCP/IP are used to handle the communication requirements between the mainframe and the workstations on our LAN, which has been completely rewired and organized in preparation for our new Web/e-mail T1 Server and firewall.

Recently a new set of handouts has been developed for our on-site N2O training classes. These handouts contain pictures of all of the screens that are discussed during the training classes. Attendees can follow along and take notes. Since most attendees never receive copies of the Administrator and User Manuals, these handouts can serve as a useful quick-reference guide even after the class is completed.

In addition, a training guide has been developed for GENART, our Y2K remediation tool. The guide provides a step-by-step procedure for quick proficiency, since few developers can spend their valued time learning how to use a Y2K tool. Like the N2O handouts, this training guide can also serve as a quick-reference guide.

This new material has been successfully utilized at several sites. If you are interested in on-site training, contact our sales division at tsi@treehouse.com or (412)741-1677.

If you would like to receive a printed copy of this or any other issue of Treetips, please contact us.