|
General
Electric
FOR
INTERNAL USE ONLY
|
| INTRODUCTION
The following guidelines should be used in evaluating the facts shown in the Application Audit (CPB-565). The information shown below relates to each section and questions asked in the Audit Brochure. Possible meanings shown for each point do not represent the exact meaning of the answers obtained, but should serve as clear identifiers or indicators of facts or problems about a prospect's business and what it could mean to him.Note: Use of these Evaluation Guidelines should be used in conjunction with a knowledgeable pre-sales AE and/or your Regional application specialist to insure maximum impact and accuracy. A. Suggested Approach - Application Audit Presentation Use as outline for summarized report for presentation: 1. Section 1 deals with general organization of the customer's company and assignments of responsibility in the Bill of Material area. 2. Section 2 deals with the characteristics of the Bill of Material information now available in the customer's present system. 3. Section 3 deals with the characteristics of the customer's present system. 4. Section 4 deals with the Bill of Material file organization approach currently in use in the customer's business. We suggest that the findings from these four areas be used as lead topics in the summary presentation report-prior to the recommendation portion. GUIDELINES FOR EVALUATING Parts Explosion Application Audit Questions The following points are suggestions, critical points to look for, etc., tied directly back to questions in the audit and capabilities of the Parts Explosion System. The prime objective of this documentation is to assist you in relating capabilities of the system back to specific business requirements pointed out by the audit. Section 1: Business Organization and Responsibility A. It is absolutely essential to find out just where, organizationally, the responsibility lies for information content and accuracy. The questions asked in this area look simple on the audit, but getting the true picture is generally very difficult. Watch out for situations where duplicate files are kept by the various functional areas. They will each feel that they are primarily responsible, particularly if they add information to the file. These "mixed" responsibility situations should be pointed out to the customer in a very objective, factual manner. Don't get trapped on the question, "Where should the responsibility lie?" Only the customer can, and/or should, make this decision. IDS can contribute to the solution of this common business situation in this manner: 1. The IDS technique of establishing and maintaining the relationships between various pieces of information in the business is a significant step towards solving this problem. In order to keep control in a data base, where information comes from various functional sources, we must first determine the relationships and interactions of the information. Until this is provided, you cannot effectively assign responsibility. The chaining design techniques inherent in IDS make this possible. In a true common data base, the file is not organized around a prime function-it is organized by information relationships. B. Covered by points under A. c. You'll find that duplicate files are the norm, rather than exception. This occurs because people requiring information cannot get it from a master file in the time frame required. The integrated IDS file and inquiry response capability can help in this area. The problem, business wise, with multiple files is inconsistency of data. There are real advantages to everyone working from the same data base, even if some data is in error! D. This is a key area to look into. Even if it is stated that product structure information is conveyed via a specific medium, you will often find that the people using the information are really getting it in some other way or combination of ways. (Bill of Material plus drawing as example) Point this out _n_ow_, because it is disastrous to dry up these informal sources by installing a new system. E. This group producing the present documentation is a key group to point out the functional capabilities of the system and how it can help their work area. Page 2 Section 2: Characteristics of Bill of Material Information A, B,C: The information obtained under these three points is useful in several ways: 1. Use (in conjunction with Point G) for file size estimates. 2. General statistics of interest to customer management. 3. Use as some general indicator of common usage of parts-a prime area for improvement with IDS elimination of redundant information within the same file as well as across multiple files. Approach: a = 100,000 individual, unique items. b = 5 parts, average per assembly. c = 150,000 total records in present Bill of Material Files (Includes common usage parts) Step 1: Divide (a) the total number of individual unique items in system by (b) average parts per assembly. 100,000 divided by 5 = 20,000. This answer provides the total unique assemblies in system (20,000). Step 2: Subtract this value for unique assemblies from total unique items in system. 100,000 - 20,000 = 80,000. This represents the total number of unique component parts in the system. Step 3: Subtract the value from Step 1 (total unique assemblies) from (c) the total records in B/M System File. 150,000 - 20,000 = 130,000. This value is the total number of detail part records (component record callouts) in the B/M System File. Step 4: Subtract the total number of unique detail parts (answer from Step 2) from the total number of detail component records in system (answer from Step 3). 130,000 - 80,000 = 50,000. This value is the total number of multiple uses of _it_e_m_s as opposed to unique uses. (Used only once.) These records will be both detail parts and assembly records called out at higher levels as components of some higher assembly. Relating this value of multiple uses (50,000) back to the total records in the file (150,000) tells us that 30% of file falls into this category of multiple use items. This is a tremendous area of possible improvements using the IDS techniques: Reducing redundancy: a) the master-detail concept of IDS is very effective in reducing redundancy by lowering or eliminating the repetition of match keys; b) additionally, the use of cross reference pointers allows the practical storage of common data only once in the file; c) the capability to link all master records that are related to (affected by ) a detail not only helps eliminate redundancy, but is a critical requirement to achieving a truly integrated, common data base. Communication capability between data records is essential. The hierarchial file solution of_a_ll sequential oriented approaches does not effectively have this solution Note: One condition could distort this analysis of multiple use items. Typically, master part files in accounting areas (Question A) will be kept from day one forward. Little or no attempt will be made to purge unused items. Most people, on the other hand, will make some attempts to purge Bill of Material system files. The net effect is an overstating of unique items in the system, which results in understating the percentage of multiple uses. This will make your estimates of multiple use very conservative, a point you should point out to the customer. D. This statistic is important to the customer and us for estimating possible file space savings through a purging scheme. The solution for this is not easy, nor is there one pat answer for all businesses. The point is that IDS capability to establish relationships between information, makes this solution much more practical. Remember, maintaining the relationships between files (like index sequential) is only a superficial solution to this area. You have to get down to the individual data elements level to be successful! E. Any customer that has a mix of pure service parts requirements (parts not used in current production) with current production parts requirements has a control problem of Major significance. He must have the capability to identify "service part only" items, both "service and production part" items, and "production part only" items. The Parts Explosion System, as released, does not provide a canned solution to this control. The IDS data base techniques, however, do offer sound, practical solutions. (1) Carrying a code in the Master Item record (only once) is a possibility. . . with usages linked back. (2) Item type records are a possibility. (We are doing this in the Inventory Management Phase of MIACS) The point is IDS offers a variety of possible solutions. F. This area is a somewhat similar problem of control to (E). Items from different types of sources require different types of control. The possible ways to control with IDS are covered in prior paragraph. One point, beside control, is important here. In the Parts Explosion System we use a bottom justified level technique. (See System Manual for details). This effectively places items in levels page 3 and 4 |
| related to their use, as opposed to where
they appear in a structure, (low level position.) In our system, all
purchased parts and raw materials appear in level (l), the lowest level.
This can offer some interesting possibilities for user extensions in this
area of controlling purchased items. For example, in our present level by
level explosion reports, all items in level (I) represent the purchased
parts or raw material requirements for the given explosion. From this
point, a user could introduce
logic for a variety of control approaches. G. Required for sizing files (present) and estimating the possible savings using IDS in the redundancy area. Characters required in IDS to establish a component record (product structure record) will typically be much less than any sequential scheme. H. The majority of Bill of Material Systems in use today, regardless of physical system used, will carry information other than pure product structure data. By going through the analysis provided in the audit, you can open up many significant areas. 1. By just going through this area with your customer, you will convey an awareness of the manufacturing problem and a desire on our part to truly try to understand his business requirements. 2. In sequentially organized systems, knowing what this information is, and where it is maintained may open up additional redundancy elimination. 3. Maintenance of codes and control data in a multi-file environment is a fantastic problem. . . and this condition is present in many, many customer systems today! (As a matter of fact, in almost all systems except those utilizing an IDS data base. We in GE are naturally prejudiced, but look around. . . it's a fact!) We have a proven answer in the Parts Explosion System. . . the Master-detail IDS concept, as applied in our Material Item Master and Component record structure. We carry the common data in the Master regardless of where it is used. The component record is exactly that. . . a relation record that records usage information. Your customer can add his control information in this Item master and conveniently tailor the system to his specific control requirements. We have this at every Parts Explosion test site and are in a position to prove it through results.Another significant dimension of control has come out of our test site experiences. We have demonstrated the ability to provide different types of control, for the same item, as affected by it's specific use in different assemblies or use at a given point in time. Test site users have added certain control data to the component record. For example: a. It is possible to carry effectivity data in the component record and select the proper record based on defined change points. b. The above can also be accomplished by an effective date record with items linked to it. Page 5 |
| J &K
c. Make or buy decisions may be affected by where an item is used. Carrying this code in the component record will allow for varied decisions, based on actual use. d. The key point is capability in IDS to control at the level required by your customer. . . families of like parts, end items, item level, or usage in specific assemblies. I. The standard Parts Explosion system does not provide control for such items. This can be handled in IDS file approach, however, by including a code in the material item record (like any other control code) and add decision logic to the explosion or inquiry programs. You will find entry points in all Parts Explosion System programs that allow for such additions. . . at the COBOL source level not machine code level. These two areas are complicated and controversial. . . but we should be aware of them if the conditions exist. The IDS data base will handle these situations because it is based on relationships of data, not any significance of item numbers.(1) (1) usually exists because obtaining where used information either does not exist or is difficult to obtain. (2) (K) usually exists because control capabilities are impossible or difficult to build into the existing system. (Exception is use of this technique to eliminate drawings in Engineering for similar items.)In any case, we can accommodate such item numbering systems. Be very careful about suggestions to alter the present numbering scheme. . . just because we can provide the same capabilities in an IDS file structure. Changing item identification policy is a traumatic experience in a manufacturing organization that will affect areas never dreamed of. Approach this a "possibility that must be carefully planned and thought through."L. This is basic information required to develop some estimates of processing requirements. This is also the area where "random" vs "sequential" approaches meet head on! The Parts Explosion System is purely transaction-oriented, utilizing a random data base. We are dedicated to this approach in GE for many valid reasons. (Refer to the first section of Application Review for Parts Explosion System). Additionally, you will find that many customers are batch and sequential file oriented because nothing else has been available to them. We cannot afford to overlook or be hypercritical of this fact! Sequential schemes have been effective (in simple problem areas) but now you can offer a broader possible solution to the total information requirement and integrated systems environment.If the customer has full random capabilities in his objectives, (transactions, inquiry, and usage of 2 common data base) he will appreciate our IDS and Parts Explosion System Products. page 6 |
| Incidently, in even "high volume B/M
transaction situations" . . . break these volumes down to a daily
basis and compare hit ratio to total file size. Batch processing looks
less and less attractive.
M. Point I is needed to identify that portion of time "outside the system." It may prove to be a significant portion of the overall time required to process an engineering change, and the Parts Explosion System is not going to solve this area (except for responsive inquiry for engineering research tasks.) Identify the magnitude so that improvement in the processing area can be identified. There are many possible areas of improvement in Engineering work, through the use of computers, but they are outside the scope of this application.Point 2 is required to measure possible improvements in our system vs present or proposed. Remember one point, your competition is sequential approaches. Be sure to point out to your customer that not only actual file maintenance runs must be considered, but also pre-sorts, and restructuring processing, inherent with sequential file overflow area approaches.Section 3: Characteristics of Present System A. This area of the audit is designed to clearly identify the customer's present "system vehicle." We could probably spend the rest of our lives debating the respective problems of moving from one level of equipment to another. Let's look at some general points. Point A-(l) - Going from a manual system to any mechanized system is pure brute force work! We do believe in GE, however, that the most effective step, long term, is to go after a basic area like Parts Explosion in the random device environment. Do it right and establish a sound base for future system growth in the manufacturing area. So called short-range, interim steps, only postpone the final decision. Other work such as Financial can be picked up parallel. . . but aim for the manufacturing control area!Point A-(2, 3a, 3b, 3c . . .) - All card-oriented systems have certain common characteristics and problems. I. Sequential orientation (Refer to Parts Explosion Application Review. . . Evolution of Techniques . . Section I.) 2. Quality of Data is typically poor. Card files are wide open to manual interruption and multiple location maintenance problems for codes and other common data. Expect and p_l_a_n for unexpected data quality and conversion requirements when going from a card system to any magnetic recording mode. Point A-3-f - If your customer now has a disc B/M system, _y_ou_ have a prime prospect! He has experienced, personally, the problems of sequential file organization on disc, or the magnitude of a self-designed random file approach. page 7 |
| Provide him with documentation available on
IDS and the Parts Explosion System. Ask for a sample of his actual Bill of
Material data and process it through the Parts Explosion System. Provide
him with available standard outputs from the system and an IDS formatted
file dump. Have qualified Technical Support people work with the customer
in these areas. You will generally find this prospect extremely receptive
to the IDS/Parts Explosion System approach. He has been there!
Point B - This area in the audit is designed to identify the customer's present system capabilities. It can also serve to identify unsatisfied requirements he has. We have intentionally structured the points around capabilities now included in the GE Parts Explosion System. To truly serve your customer's needs match his business requirements with available capabilities. Every type of function except points 4, 10, and 12 are available in the Standard Parts Explosion System release. Examples are displayed in Section 3, Application Reviews. Point 4 - Selective Bill of Material reports, is possible with minor modifications. Add control codes to Material Item records and selective logic to existing Inquiry and Explosion modules. Point 10 - IDS Fild structure used in the standard release has the capability to build on implosion capabilities with minimum effort. One test site, Abbott Laboratories, is implementing this capability as an extension of the standard release. Point D - Effectivity control is being implemented as an extension to the standard release by another test site, Aeroneutronics Division, Philco-Ford. Effectivity control will be a standard function included in MIACS, phase 2, Inventory Management. Many of the capabilities provided by the Parts Explosion System are standard expected functions. (Bill of Material reports, where-use, etc.) Two areas are particularly outstanding and unique. This system is capable of providing both gross and n_e_t, level by level explosion, with lead time set back, multi periods. We also offer some fundamental input data editing capabilities which is also unique as compared to competition. All of these capabilities are covered in detail in the Parts Explosion System Application Review. Point F - This point was included to get some indication of the customers' explosion requirements. You will often find that his frequency and approach are now what he would truly like to do, but is limited because of system and/or processing capability. Be sure to dig out "what he would like to do."Point I - The system as released is not capable of tracking and controlling discrete orders. Orders are pulled together in summation during the explosion, for a total net explosion of all items entered. Inventory Management, phase 2, will have this capability. page 8 Section 4: Bill of Material File Organization Approach This section is designed to identify the type of file organization now in use. It is structured into: (1) (2) (3) (4) Card Systems Tape Systems Tape/Disc Systems Disc Systems The strong and weak points of each type is covered in Section I of the Parts Explosion System Application Review. Once the file organization type is identified, use the Application Review material and other IDS documentation (I) to point out the respective improvements possible with the IDS data base approach used in the system. The IDS managed data base approach utilized in this system (and all current developments) is as significant, or greater perhaps, than the actual application software. We have a unique product available to us in IDS and it warrants a major role in the overall approach. In addition to identifying file organization, sequence schemes, and storage devices utilized, there are several additional key points asked for in each section:1. Where-Used Approach. . . both file capabilities and organization. . . and present system's functional capabilities. This is one critical area that the IDS file structure, used in this system, is far superior than any other technique known today. The Master Detail relationship used in IDS provides efficient where-used capabilities with proven file integrity. 2. Absolute Addressing vs. Relative Addressing. IDS utilizes a relative addressing approach as opposed to IBM's Index Sequential which is absolute. Experience (GE and others) has proven that absolute addressing (physical hardware addresses) is much more vulnerable to changing file devices and particularly to overflow problems.3. How is overflow handled, particularly in IBM's Index Sequential? Degrading of performance is characteristic of predefined overflow area techniques as the overflow area fills up. Variations have been implemented to try to solve this problem such as "overflow to another overflow area" . . .In the BOS/MT version of IDS, we also fall off in performance when the file utilization reaches 65 to 70%. This is vastly improved in IDS releases utilizing available page inventory capabilities. This can be pushed up to the area of 90% utilization. . . excellent by anyone's standards. (1) IDS - A New Concept in Data Management, CPB483A IDS - Data Base Study - CPB48 I A IDS User's Guide (first release of a series of documentation of User's Experiences) page 9 |
| The prime difference in approaches between
IDS and other techniques is that IDS "spreads" the file across
the total defined available space. Overflow space is not restricted to
predefined physical locations . . . tied to device characteristics such as
a track, etc. This, in effect, allows for available space for additions to
the file to also be "spread" over the file. The logic internal
to IDS looks for the nearest physical space, as required, and stores the
new record in the relationship predefined by file structure selected. It's
important not to misinterpret the term
"spread". The page concept and place near capabilities of IDS will physically store related information in the same page, or nearest page with available space, to reduce accessing of the file. The design logic for overflow in IDS is geared to the concept that all remaining space is available for overflow. . . as opposed to specific logic allocations for overflow. Our performance is tied to total file utilization. . . not the filling up of individual defined overflow areas.Rebuilding of files to unload overflow areas in Index Sequential is very time consuming. Throw the multiple file approach in on top of this, and it becomes a significant task. 4. Linking techniques are also covered in this section. Bi-directional linking is truly possible in IDS with all of its inherent capabilities in system design. Without effective bi-directional linking, problems such as scheduling, re-scheduling, and where-used become unwieldly. page 10 |
| General Electric Parts Explosion System Application Audit |
