Interact: A Manager
of Free-Form Text
Interact version 2.0 (V920823.0000)
Redstone Technologies, Inc.
4786 S. Olathe Way
Aurora, CO 80015
Interact comes with all the programs necessary to batch or interactively load data, report on the data to print, or interactively query the database.
Interact runs on all HP 3000 Series, Classic or Spectrum. The software has a flat rate cost of $5,000, with discounting on multiple CPUs. Support is 20 percent of the purchase price per year and includes phone in, electronic support and new releases of the software. All prices are in US dollars.
Application uses 3000 to support files from litigation, library management or HR systems
Review by Shawn Gordon
Not to be confused with the former Interex magazine, Interact is a DBMS for the HP 3000 that has been around for at least 10 years. Interact is a very unique DBMS, and not just for the HP 3000 — it is designed to index and manage large amounts of text, especially free-form text. Some of its uses are in Litigation Support, Library Management, Records Management and Personnel Records.
For example, a personnel database would relate to an employee biography/skills database. Each employee record could contain fields for standard employee information such as name, address, hire date, department, salary, etc. You could also store the employee’s resume in a single field, with each word indexed. If you needed to know which employees had prior marketing experience, a search on the word “marketing” would produce that information in seconds. There are no limits on your ability to modify data within a key field, either.
How does it work?
Behind the scenes, Interact uses sets of KSAM files to store and index the information, but it is doing some rather fancy stuff within its API that appears to make total non-standard use of the KSAM files. I tried dumping the files to see if I could get a sense of what was happening, but no dice. Very fascinating stuff, though.
So you have three different functions: create a database, maintain a database, and query a database. Once a database is created, you can either batch-load your text files into it, or data-enter them. You can also add and delete fields and indexes at any time.
The interface is a little hard to get used to. You have to establish file equations for the database, but you can only do so through the supplied utility. Then it writes them to a control file so that they are executed when you open the database.
Each Interact database can contain over 16 million records, with up to 255 fields per record and up to 255 occurrences of values within each field. Interact allows you to define either fixed or variable length fields, which is obviously a foreign concept to any person only familiar with Image/SQL. Any field can be made a key and be indexed. The size of the database doesn’t seem to have an impact on the speed, only the number of qualifying records that might be returned from a search.
Interact allows for up to 255 users or user groups to be defined for security. Unlike IMAGE you can restrict on Inspect, Add, Modify, and Delete. You can also define acceptable values for fields at the database level.
There are some fascinating features in Interact, such as system and database specific thesaurus tables, so that synonyms for a term can be grouped with the term. A system thesaurus applies to all databases, and a database-specific thesaurus obviously just applies to the database it’s configured for. There is also a Soundex dictionary that can be maintained so you can do searches using phonetic spelling. This can be very handy for words whose spelling you are unsure of.
Another interesting feature is the code expansion table. This is also specific to a database and allows you to create a short hand alias for commonly used words or phrases. This is meant to increase your speed when entering repetitive words or phrases.
There is also something called “Stop and Go Word Lists” which operates on the field level within Interact. This tells Interact to either index only these words (Go List) or to index all words other than those specified (Stop List). This granular control lets you fine-tune your database to a very high degree, but I don’t know how intimate people would need to be with the data to maintain these lists in reality.
Installation and Documentation
Installation is the very common type of load tape, restore job, stream job, reload tape, done. Installation was quick and clean. The software is only 2 Mb, but comes with 14 Mb of example information.
The documentation is well written and clearly organized, and includes a table of contents. It is also very large. There is one big manual for using the product, and another for using the API to write your own programs to access the data. Since this is a new concept for many people, a very clear walk-through is done with the records from the Alger Hiss trial. My demo also included the Starr report and the Ollie North testimony, which were pretty fun to go through. All of this made the learning curve pretty short.
I decided to browse through the Ollie North papers for my examples. If you look at Figure 1below, you will see an example of the prompt you are given once a database has been opened. I’ve included the help text so that you can see the various options that are available. For my purposes I searched for the word “shred”, the keyword in front means that I want to search for that word in context, and the context is the /3, so I get 3 total lines displayed, with my keyword as close to the middle as possible.
Figure 1 Enter request: KDIC /3 SHRED Data may be Reported, Queried or Edited as shown below ------------------------------------------------------ REPORT functions allow currently existing data to be displayed on your terminal, printed on the system line printer or written into a sequential file that may be used for special purposes. COMMANDS DIsplay Print COpy CT (copy to tape) ID IDS Fields KD (Keyword Display) KP (Keyword Print) KDS (Keyword Display Summary) KPS (Keyword Print Summary) KDIC (Key Display in Context) KPIC (Key Print in Context) CP (Change Printer) QUERY functions allow selection of data based on specified criteria. COMMANDS SEarch SUMmary SUBset SOrt SAve RECall CLear EDIT functions allow data to be updated in the Data Base. COMMANDS Add INsert Modify REPlace DElete NOTES (1) Requests may be abbreviated with the underlined letters; (2) To request specific information for one of the above commands, enter that command and press the HELP (f1) key.
Figure 2 shows an example of the output generated. You can scroll through each record, or change your search criteria. As you can imagine, the Starr report lends itself to some interesting search criteria. The Starr report is actually a very good example of the batch load ability. Redstone downloaded the report from the Internet, and then just stuffed it into an Interact database. Now you can browse and search for anything that you like.
Figure 2 Keyword Frequency Listing Data Base: OLLIE Semiotix, Inc. ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; KDIC /3 SHRED ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; Record Identifier Field/Occ. Keyword in Context --------------------------------------------------------------------------- --- 11 TEXT/1 with the Israelis in August of 1985? I am not aware of a single shred of evidence that that was the case. LT. COL. NORTH: No, and I want to make 10 TEXT/1 REP. HYDE: Well, I would say I'm sure you didn't shred that as a document. But symbolically, I have heard many a media person, and some legislators 10 TEXT/1 and it is certainly not referring to "do I have more time to shred documents" because I I I'm not I want to answer that very clearly. I never 10 TEXT/1 General or anyone else whether or not I had enough to time to shred REP. RODINO: In other words, you did not ask for 24 or 48 hours more? LT. COL. Enter
Creating a database is a matter of using the online application to define the fields and records that you want and to assign indexes. Much easier than working with DBSCHEMA for an Image database.
There is a report generator as well. The system will do its best to intelligently design a report for various types of desired output. It’s not a particularly sophisticated layout engine, but for the most part it is functional.
Interact is a totally unique product on the HP 3000, and one you have to think about for a bit to see if it is applicable. I know that there are things out there that are designed to help keyword and index documents into an Image database. But I don’t think they are going to work as well, or be as efficient, as Interact. This is a large and rich product, so space constraints keep me from going into too much detail, but hopefully you get the idea of the possibilities for this application.
The interface is a bit on the lean side and takes a little getting used to, but once you are familiar with it, it’s not too bad. The product is very fast and versatile. I didn’t try the API, so I can’t really comment on what’s involved there — but the fact that you’re not limited to their query tool is pretty convenient.
If you are trying to manage free-form text and need to be able to search and do research, you really should look at Interact. I found it rather entertaining looking through the Alger Hiss report searching on things like “communist” and seeing what I got. With the batch input, you basically just get the text from somewhere and pump it in and start your research. You could even scan and OCR reams of legal documents so that you could do your online research. As you can see, there are a variety of interesting options.