27 NESMA 通用格式化系统 案例分析

27 GENERIC FORMATTING SYSTEM(通用格式化系统)

通用技术

问题描述

公司使用多个后台信息系统（A、B等）来履行不同的法定职责。在每个信息系统中，定期生成大量信件。这些信件在后台信息系统中进行设计。IT部门决定解决格式化功能和信件的发送，并将其放在单独的格式化模块中，如下图所示。

对于每个信件，固定的信件文本存储在格式化模块X中。格式定义了多个设计样式。其中还包括称呼和签名等部分。后台信息系统（A和B）在传输文件中提供所有可变数据。这包括地址信息。

问题是，信件是如何计数的？我们是每个信件识别多个用户功能，还是每个信件仅识别一个外部输出，该功能属于哪个信息系统？

讨论

每个信件都是由后台信息系统创建的计为一个外部输出。当信息系统A中的信件发生变化时，可能会出现多种情况：

1. 信件中使用的一个或多个变量的内容计算不同。这意味着信息系统A中的外部输出发生变化。格式化模块X不需要修改，因为变量已经存在，只是内容的确定方式不同。
2. 必须在信件中添加新数据（变量）。在这种情况下，必须在信息系统A中修改外部输出，但在格式化模块X中也必须进行修改，以便读取新数据并将其放置在信件上。
3. 这封信的固定文本有一处修改。对于此更改，只需要修改格式化模块X。变量或内容未被引用，因此信息系统A中的外部输出不会更改。

此外，还可以对设计样式（布局、企业标识、徽标、字体等）进行其他修改，这些修改适用于所有字母。在这些情况下，仅格式化模块X受到影响。

因此，在后台信息系统中存在的每种类型的信件在格式化模块X中都有对应的。这是否意味着后台信息系统的每一个外部输出，后面都有格式化模块X的外部输入和外部输出？模块X中的文本和设计样式由专业程序员维护。因此，需要付出额外的努力。

从最终用户的角度来看，创建信件是一个输出过程。最终用户不知道有一个单独的格式化模块，其中收集所有信件进行格式化和发送。从这个角度来看，每个信件有一个外部输出。信件的任何修改都意味着本外部输出的修改，无论修改是否必须应用于信息系统A、格式化模块X或两者。在项目规划中，必须估计两种类型的工作。

答案

最终用户未明确要求格式化模块X，因此这被视为技术解决方案。每种信件计算1个外部输出。进行一般性修改时，计算外部输出的数量与更改的信件数量相同。

27.2 功能通用

问题描述

与上述情况相同，但现在终端用户明确要求一种解决方案，其中可以以简单和通用的方式维护信件的格式。为识别的信息系统X起草了功能设计，以便于维护信件格式。

讨论

对于后台信息系统，每种信件计算1个外部输出。

对信息系统X的功能设计进行功能点分析。假设记录了信件的历史，计算至少1个 外部输入 和1个外部输出，以及所有要实现的维护功能。

在修改的情况下，计算所有更改的功能。

答案

为后台信息系统计算每种信件为一个外部输出。

对信息系统X的功能设计进行功能点分析。假设记录了信件的历史，统计至少一个外部输入和一个外部输出，以及所有要实现的维护功能。

在修改的情况下，计算所有更改的功能。

27.3 对这两种变体的评论

IT部门选择使用格式化模块来维护变体1中的 信件 ，以便更便于维护。最初必须购买单独的格式化系统，但对于信件的修改可以更简单的进行，这就是投资回收的方式。

在变体2中，最终用户选择使用格式化模块来维护 信件。这里的原因也在于，投资将在维护 信件 时收回。

在功能点方面：在变体1中，建设项目的功能点数量少于变体2。当 信件 发生修改时，变体1中的功能点计数多于变体2。

就生产率而言：在变体1中，每个功能点的施工时间相对较长，而维护时间相对较少。在变型2中，每个功能点的小时数相对正常，在维护期间也是如此。

上面显示了可以在信息系统边界上做出的选择的效果。相同的（技术）解决方案会产生不同的功能点分析结果。

---

27.1 Technically generic

Problem description

A company uses multiple back-office information systems (A, B, etc.) for performing different statutory duties. Out of each information system, regularly large quantities of letters are created. These letters are described in the designs of the back-office information systems. The IT department has decided to resolve the formatting function and the sending of letters generically and to put it in a separate formatting module, see the figure below.

For each letter, the fixed letter texts are stored in the formatting module X. The format is defined in number of design patterns. Also parts like the salutation and the signature are included therein. The back-office information systems (A and B) supply all variable data in a transfer file. This includes the address information.

The question is, how are the letters counted? Do we identify multiple user functions per letter or do we identify only one EO per letter, and to which information system does this function belong?

Discussion

Each letter that is created from a back-office information system counts as an EO. Upon a change of a letter in information system A, a number of situations may occur:

1. The content of one or more variables used in the letter is calculated differently. This means a change in the EO in information system A. The formatting module X does not need to be modified, because the variables already exist, only the content is determined differently.

2. New data (variable) must be added to the letter. In this case the EO must be modified in information system A, but also in the formatting module X a modification has to be made in order to read the new data and place it on the letter.

3. There is a modification to the fixed text of the letter. For this change, only formatting module X needs to be modified. The variables, or the content are not referenced, so the EO in information system A does not change.

In addition, there can also be additional modifications introduced to the design patterns (layout, corporate identity, logo, font, etc.), which apply to all letters. In these cases, only formatting module X is affected.

Every type of letter that exists in the back-office information systems thus has a counterpart in the formatting module X. Does this mean that every EO in the back office information system is followed by an EI and EO in formatting module X? The texts and design patterns in module X are maintained by specialized programmers. Therefore, extra effort is to be delivered.

From the end-user perspective, creating a letter is one output process. The end user does not know that there is a separate formatting module where all letters are collected to be formatted and sent. From this perspective, there is one EO per letter. Any modification of the letter means a modification of this EO, regardless whether the modifications must be applied to information system A, formatting module X, or both. In the project planning both types of effort must be estimated.

Solution

The end-user has not explicitly asked for formatting module X, therefore this is regarded as a technical solution. Count 1 EO per letter. When a generic modification is made, count as many EO as the number of changed letters.

27.2 Functionally generic

Problem description

The same situation as above, but now the end-user explicitly asks for a solution where the formatting of letters can be maintained in a simple and generic way. A functional design is drafted for the recognized information system X with the purpose of easy maintenance of letter formatting.

Discussion

Count 1 EO per letter for the back-office information systems.

Perform a function point analysis on the functional design of information system X. Assuming that the history of letters is recorded, count at least 1 EI and 1 EO, plus all maintenance functions to be realized.

In case of modifications, count all changed functions.

Solution

Count one external output per letter for the back-office information systems.

Perform a function point analysis on the functional design of information system X. Assuming that the history of letters is recorded, count at least one external input and one external output, plus all maintenance functions to be realized.

In case of modifications, count all changed functions.

27.3 Comments on the two variants

The choice to use a formatting module to maintain letters in variant 1 is made by the IT department in order to carry out the maintenance more profitable. Initially a separate formatting system must be purchased, but changes that affect all letters concern can be applied simpler and that is how the investment is recouped.

In variant 2, the choice to use a formatting module to maintain letters, is made by the end-user. Here the reason was also here that the investment would be recouped in the maintenance of the letters.

In terms of function points: in variant 1 less function points are counted for the construction project than in variant 2. When modifications to the letters occur, in variant 1 more function points are counted than in variant 2.

In terms of productivity: in variant 1, relatively more hours are required per function point for the construction, and less in maintenance. In variant 2, the number of hours per function point is relatively normal, and also during maintenance.

The above shows the effect of the choice that can be made on the boundaries of an information system. The same (technical) solutions produce a different result for function point analyses.