Scout Client/Server
by
Shawn M. Gordon
President
S.M.Gordon & Associates

Introduction

I am always look for more convenient ways to extract, manipulate and convert data on the HP 3000. I typically find that report writer tools end up being a bit to cumbersome, 3rd and 4th GL’s require too much work. I have looked at some client/server tools, and some of them have been good, some haven’t. What I recently found though was Scout Client/Server from Minisoft, herafter just Scout.

Scout has a fairly long history as a data extraction and formatting tool. A couple of years ago it was rewritten as a Client/Server tool using the MiddleMan product, also from Minisoft. I have used MiddleMan in the past myself, and it is a pretty snazzy middleware tool for doing network access.

I tried Scout under both Win 3.1 and 95, against MPE/iX 5.0 and 4.0 I used SOS from Lund performance to gauge system impact.

Features

Scout is a dictionary driven program, and requires that you set up the your dictionaries on the host, and then download it to each client. I found this piece to be a little bothersome, but it does allow each user to customize their own environment. So to follow the examples we would do the following to create a dictionary for an Image database, note how you can also add MPE and KSAM files to this;

RUN DICTBLDR

  SCOUTDBF Loading Program - v3.0b          

  Select Option:    
        1 - Add Database Entry to SCOUTDBF                      
        2 - Add MPE/KSAM File Entry to SCOUTDBF                 
        3 - Delete Base/File Entry from SCOUTDBF                
        4 - Show SCOUTDBF Info                                  
        5 - Print SCOUTDBF Info                                 
        6 - Modify BasE/File Info                               
        7 - Reorganization of SCOUTDBF                          
        8 - Download to PC (SCOUTDBF --> SCOUTDBF.INF)          
        / - Exit              
       >4

SCOUTDBF  --  SCOUT System Database/File Information File

    Creation:        Date =  10/26/94       Time = 10:06:48    v3.0b
    Last Update:     Date =  10/26/94       Time = 10:06:48    v1.0a
    Last Download    Date =  01/31/96       Time = 12:13:28
    Updated:          0 Times  ;  Downloaded:     1 Times
    File Status:     Limit =  1098    In Use =    44    Garbage =     0

SCOUTDBF has the following 4 Database/File Entries
  DEMO1(DB)       DEMO2(DB)       FKSAM(K)        FMPE(F)
More (Y/N)? Y

Enter Database/File Entry Name > DEMO1   

DEMO1        TYPE = DATABASE

SETS:

  CUSTOMER        (M)       HEADER-AUTO     (A)      ORDER-LINE-AUTO (A)
  ITEM-AUTO       (A)       ORDER-HEADER    (D)      ORDER-LINES     (D)
  COMMENT-LINES   (D)       SHIPPING-ADDR   (D)      CREDIT-REFERENCE(D)

ITEMS:

  ACCT-NO            X6                 ADDRESS-1          X20
  ADDRESS-2          X20                AMOUNT             P8 (S, 5, 2)
  BALANCE            P8  (S, 5, 2)      COMMENT            X20
  DATE-ORDERED       X6                 DATE-PROMISED      X6
  ITEM-NO            X8                 LIMIT              P12(S, 9, 2)
  LINE-AMOUNT        P8  (S, 5, 2)      LINE-NO            X2
  NAME               X20                NO-OF-ITEMS        I1 (S, 5, 0)
  ORDER-LINE-NO      X8                 ORDER-NO           X6
  ORDER-STATUS       X2                 PHONE              X10
  QTY-SOLD           J1  (S, 5, 0)      RATING             X2
  REFERENCE          X20                REGION             X4
  SH-CHARGE          P8  (S, 5, 2)      SHIP-ADDRESS-1     X20
  SHIP-ADDRESS-2     X20                SHIP-NAME          X20
  SHIP-STATE         X2                 SHIP-ZIP           X6
  STATE              X2                 SUB-TOTAL          P8 (S, 5, 2)
  TAX                P8  (S, 5, 2)      TAX-ID             X2
  TAX-RATE           P4  (S, 1, 2)      ZIP                X6

The part that is rather annoying about this part is that you have to download the dictionary now. This means for every user that is going to use Scout, they have to download the dictionary to their PC. Then if the databases update, you have to create a new dictionary file, and have everyone download it again. However, once the dictionary is loaded on your PC, things are a piece of cake.

See Figure 1 for the main screen. I have turned on every window, most of them are configurable, so you can have a less cluttered interface if you like. Basically you select a data source from the combo box on the right side of the screen. If this is a database then you will get a list of datasets underneath. What I have done here is drag USER-M into the Query data source window. Since this is my master set that will drive everything, I want everything to link off of it. Next I drag the MAIL-D set into the Query data source window, and drop it on top of the USER-M data source, this is so Scout knows they are linked. If you look at Figure 2, you will see the window that pop’s up whenever you drop in a new data source. This allows you to select fields for the report, and the linkages between the sets, as well as the access method along the key, which is quit nice.

So here you are visually designing your query, and the fields you are going to want to report. Scout let’s you also go through and create calculated fields as well as subscripting fields. This gives you a huge ability to customize your output. You can read key items from a file to drive your query, enter key values yourself, or perform serial reads. So you can specify which item to compare, how to compare it (relational operator), and what to compare. This pretty much covers the range of how you would be able to specify key information.

Take a look at Figure 3. Here is the result of our query operation. On the bottom half of the screen you can kind of see how Scout is internally performing it’s operations. On the top half, you can see the results. I have configured to have a simulated green bar output, so that is why there is the shading. This report is obviously in a simple grid format, which is what I wanted. If you look at Figure 4 you will see a number of the ways that Scout can download and export it’s information. The ability to save in a variety of native PC formats is very nice, you can quickly stuff something into dBase, or a spreadsheet with these options. I think the export options need a little updating though, some of these are some very old file formats, like Symphony, and I didn’t see anything for Microsoft Access.

Needless to say, or maybe not, you can specify very sohpisticated sort criteria, as well as be able to save your constructed queries. This allows you to build whole libraries of reports that you can execute on demand. Scout even supports multiple libraries, so you can easily group and maintain your saved queries.

Usability (also installation)

My biggest criteria for ease of use is if I can use it without looking at the manual. While you need to look at the manual to get things set up, the software itself is very easy to use. Whenever you are using network based client/server tools you tend to run into situations that can make installation, shall we say, interesting. If you already have a WINSOCK in place, and are using a network connection, then you shouldn’t have any trouble. The MiddleMan middleware piece of Scout uses OLE
Automation to make use of the communicaiton layer, so it’s vital that the objects get registered in the Windows registry. The Scout installation does all this for you, but sometimes you have conflicts. I didn’t have any trouble, but I do want you to be aware of some of the possible challenges that might crop up

Reliability

No GPF’s from Windows, and no program aborts on the host. Everything seemed to work as advertised, and for the queries that I tested, all the data appeared to be correct.

Performance

The client side performance is really very fast. Depending on your network connection, you can count on around 200 kps data transfer rates, so unless your are retrieving really huge amounts of data, you should have no problem. The impact is mostly going to show on the HP. Here you are going to see the same overhead that any conventinal 3rd GL is going to give you. Depending on the effeciency of your query, like keyed or serial reads, you will see different performance hits, but overall I was very encouraged by what I saw.

Supportability (including Doc)

Support from Minisoft is really excellent. You always get a human being on the phone, and almost always get to talk to a tech immediatly. In general the software is very easy to use, just some installation questions.

The manual for Scout is pretty good. I think it could use a more detailed discussion on some of the actual configuration options and settings, but overall the examples and the tutorial are well written and easy to follow.

Summary

As usual let’s start with the things I would like to see improved. My biggest complaint is with building the Scout Dictionary. This program is obviously a hold over from the early days of Scout. I really think that this should be turned into an “Explorer” type program that can be run from the client. I suppose it’s not that important other than cosmetics because you won’t be using it that often, but I would like to see this part updated.

The fact that almost every prompt is a point of no return has it’s frustrations as well, but you really don’t spend a lot of time in this program after you set up.

My other complaint is with the tool bar in Scout. It has become common for little baloon help to come up when the mouse pointer passes over it, or at least put a description on the status bar, if there was a status bar.

My only other comment is an enhancement that I think would be really cool. As I have mentioned, Scout uses another MiniSoft product called MiddleMan to do it’s network communication between the client and the host. I have used this product myself to build little client/server utilities at the office. What would be neat is to be able to build a Scout series of actions that could be executed as a MiddleMan method from something like Visual Basic, or Delphi. This would give you the ability to build high level client/server scripts essentially, and totally customize and control their execution through your own client program. It would also keep you from having to build your own server to do it. I mentioned this idea to MiniSoft, and they seemed quit excited about the idea, so I guess we will see.

With all that said, I like Scout quit a bit. It’s one of those utilities that I could definitly get some use out of at the office. The interface metaphor is unique, and fascinating, I found myself drawn into it quit easily, and liked the way it flowed. I know that sounds funny, but it’s not often that you find a non-standard interface style that flows so easily. This is really a great tool for those of you who have a need to do data extraction and PC integration, you should check it out.

At-a-Glance box

Scout Client/Server version 3
Minisoft Inc.
312 Maple Ave., Suite A
Snohomish, WA 98290
800-682-0200
FAX: 360-568-2923
Web: www.minisoft.com
e-mail: doug@minisoft.com

Call, write, fax, or e-mail to receive more information or a demo. Software includes 1 tape and 2 diskettes, and one 68 page indexed manual. The price breakdown is as follows

5 users – $995.00
10 users – 1,500.00
25 users – 2,500.00
50 users – 3,995.00

All Scout licenses are concurrent use, this means that everyone in your company could use Scout for $995 as long as only 5 people were in it at a time. Support is 15% of the purchase price.