• Tidak ada hasil yang ditemukan

Family Engineering

Dalam dokumen Digital Information Products (Halaman 182-187)

Part III Tool Support & Case Study

10.3 Family Engineering

This Section illustrates how family engineering (see Chapter 6) was performed for this special case. The created family engineering overview document is presented and explanations are given for the choices.

Feasibility and Risk Assessment

Figure 10.1 shows the initial part of the family engineering overview document, which focused on feasibility assessment and risk assessment. The rationale behind the choices in the context of the case study is explained below.

10.3 Family Engineering 167 PLANT - Family Engineering Overview Document

Product line identifier 1

Main domain of intended use Information Systems

Description Products are variants of courses in Information Systems

Date of inspection of this document Date of next planned inspection

Feasibility

a) infeasible a) unknown or unpredictable

b) almost completely infeasible b) only partly known or unpredictable

c) unsure c) unsure

d) feasible, but with restrictions X d) most known, rest predictable

X e) completely feasible X e) all known in advanc

a) infeasible

b) many arguments against it;

resources not available c) unsure

d) feasible, some management support

X e) feasible, strong management support

Risk Assessment

a) often a) no advantages expected

b) very likely b) overall strategy is unclear

c) unsure c) unsure

X d) rarely X d) strategy is in most parts defined and

advantages are expected

e) never X e) strategy is defined and advantages

are expected a) no commonalities

b) some commonalities, but which cannot be technically exploited

a) not predictable at all c) unsure

d) some commonalities that can be technically exploited e) many commonalities that can be technically exploited

b) hardly predictable c) unsure d) predictable e) known for sure

3) Are there strategic advantages that are expected from a product line approach?

4) Is the product line organizationally feasible?

Feasibility & Risk Assessment

01.12.2006 28.02.2007

1) In the domain where PLANT is to be applied, how often are radical changes expected to occur?

2) Is the demand for information products predictable?

3) Are the variable parts of information products known?

1) Is the product line technically feasible?

2) Do information products in the product line have commonalities?

Fig. 10.1. The “feasibility and risk assessment” part of the family engineering overview document.

Feasibility

• 1) Is the product line technically feasible?

Choice: e) completely feasible

Explanation: the relevant file formats (Powerpoint PPT, PDF) could be handled; wrapper programs for Powerpoint existed which could append files or apply a given layout template automatically; conversion of PPT to PDF was possible; PPT files of current courses were available; the necessary technical infrastructure existed

• 2) Do information products in the product line have commonal-ities?

Choice: e) many commonalities that can be technically exploited

Explanation: after inspecting the available material, the existing courses were found to have in several parts slides that occur identically or almost identically in different courses; these slides could be extracted and modu-larized; therefore, there was potential for synergy effects.

• 3) Are the variable parts of information products known?

Choice: d) most known, rest predictable;

Explanation: there existed regularities between the common and differing parts in courses, and they could also be described on a conceptual level; the parts of the required variants could be deduced from the conception of the courses

• 4) Is the product line organizationally feasible?

Choice: e) feasible, strong management support

Explanation: the necessary staff resources were available, the product line strategy was accepted and supported.

Risk assessment

• 1) In the domain where PLANT is to be applied, how often are radical changes expected to occur?

d) rarely;

Explanation: the domain was typically stable and expected to remain stable, as the courses were part of a curriculum that was in line with the overall department and faculty strategy. The employed file formats were widely used and were likely to be available in the future.

• 2) Is the demand for information products predictable?

d) predictable

Explanation: the demand for certain course variants was predictable, as the courses were integrated in a curriculum and offered on a regular basis.

• 3) Are there strategic advantages that are expected from a prod-uct line approach?

e) strategy is defined and advantages are expected

Explanation: main aspects of the strategy were to ease the creation of course variants, make the creation process easier and less error prone.

According to the criterion defined in Sect. 6.1, all answers were either d) or e), so that the feasibility and risk of this product line were within acceptable ranges.

10.3 Family Engineering 169 Economic, Evolution, Lifecycle Aspects

Figure 10.2 illustrates the part of the family engineering overview document that focused on a quantitative effort estimation and comparison with/without PLANT. The total effort with/without PLANT was calculated according to the formulas discussed in Sect. 6.2.1 and 6.2.21.

2) Economic / Evolution / Lifecycle Aspects

N: Number of information products j: Number of content update-cycles

Effort

Assumptions: month has 20 work days with 8 hours

with PLANT Eorg Ecab

Eupdate (one cycle, estimated average) Eunique (estimated average) Ereuse (estimated average)

Result:

without PLANT

E_unique_without (estimated average) E_update_without (estimated average)

Result:

Ranges: ok if: effort with PLANT< effort without PLANT 5,20

1 4

Savings with PLANT: 2,00

[person months]

2

0,1 0,05 0,05

0,15 1 1

3,20

Fig. 10.2. The part on economic, evolution, lifecycle aspects of the family engineer-ing overview document.

The rationale behind the estimation was as follows:

• there were 4 courses mentioned in Sect. 10.2, which made up at the begin-ning 4 information products in the product line

• the number of update cycles was derived from the context, and it was assumed that there would be 2 content update cycles, e.g., after each semester

• Eorg, Ecab, Eupdate, Eunique, Ereuse were estimated by a domain expert (only one was available) after viewing existing material.

1Reminder:

Ewith:= Eorg+ Ecab+ N ∗ (Ereusewith+ Euniquewith+ j ∗ Eupdatewith) Ewithout:= N ∗ (Euniquewithout+ j ∗ Eupdatewithout)

• the effort estimations without PLANT were based upon experience of the domain expert and his observations during the last three years before this case study.

It was estimated that in this case the savings with PLANT were about 2 person months. As an aside, the average effort to create one additional product with PLANT was: Ereusewith + Euniquewith + j ∗ Eupdatewith = 0.3 person months; the average effort to create one additional product without PLANT was: Euniquewithout + j ∗ Eupdatewithout = 1.3 person months. This meant that even if the number of information products in the product line increased during the evolution, the effort with PLANT would be lower than that without PLANT.

Configuration Management

During the initiation of PLANT there were no configurations there yet for the market-based view and product-based view, and the evaluation of a match between the market-based and product-based view was overridden at that point.

After the first pass of application engineering (to be described in Sect.

10.5) it was evaluated that the realized product configuration indeed matched the market-based view of demanded configurations (to be described in Sect.

10.4), and “match” meant that all features from the market-based view were contained also in the product-based view.

Organization

The product line in this case study had few products, and for the presented context the development department model was chosen as the organizational form (cf. Sect. 4.5.2) in which every staff member can do any type of work.

The effort for the required tasks could be handled by one single person who was responsible for family engineering, domain engineering, and application engineering.

Evaluation and Controlling

Figure 10.3 shows the section of the family engineering overview document which summarized the previous results during the initiation of PLANT, to ease the decision making whether to create a product line or not. As feasibility and risk were within acceptable ranges and the usage of PLANT brought effort savings, the decision was in favor of PLANT.

The data was re-evaluated for controlling purposes after the first pass of domain engineering (to be described in Sect. 10.4) and the first pass of appli-cation engineering (to be described in Sect. 10.5). After domain engineering, there were no changes to the aforementioned data, and the product line proved

10.4 Domain Engineering 171

Dalam dokumen Digital Information Products (Halaman 182-187)