• Aucun résultat trouvé

Introducing agile Methodology into advanced Systems Engineering Training

N/A
N/A
Protected

Academic year: 2021

Partager "Introducing agile Methodology into advanced Systems Engineering Training"

Copied!
13
0
0

Texte intégral

(1)

HAL Id: hal-02441641

https://hal.archives-ouvertes.fr/hal-02441641

Submitted on 16 Jan 2020

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Introducing agile Methodology into advanced Systems Engineering Training

Uwe Kuehne, Colin Hamilton, Stefan Brueggemann

To cite this version:

Uwe Kuehne, Colin Hamilton, Stefan Brueggemann. Introducing agile Methodology into advanced Systems Engineering Training. 10th European Congress on Embedded Real Time Systems (ERTS 2020), Jan 2020, Toulouse, France. �hal-02441641�

(2)

1

1) Airbus Defence and Space (retired)

I NTRODUCING A GILE M ETHODOLOGY

INTO ADVANCED S YSTEMS E NGINEERING T RAINING

Regular paper for ERTS 2020 in Toulouse

Dr. Uwe KUEHNE, Dr. Colin HAMILTON1), Dr. Stefan BRUEGGEMANN Airbus Defence and Space GmbH, Taufkirchen, Germany [AD]

KEYWORDS:

Agile Methods, Systems Engineering, Systems Engineering Training Program, Highly complex Systems, Backlog Management, Story Map- ping, Minimum viable Product, Scrum Process Execution

0 SUMMARY

Agile methods are today very well established in S/W engineering (e.g. [ESA15], [ESA17]).

Because they have proven to be so effective, many companies are now working to extend the agile approach to systems engineering (e.g. [LM]). In Airbus Defence and Space, this is being done with specific training courses supported by coaches in those projects using the agile methodology. In addition to these general measures, Airbus Defence and Space has recently introduced agile methods into its flagship Systems Engineering Qualification (SEQ) course for advanced systems engi- neers. This program consists of training mod- ules and accompanying (real) projects. In this paper, we will present how we integrated agile methods:

 in the training modules and

 in the projects.

Furthermore we will explain:

 Which agile elements we used and why

 What preparations were needed

 What experience we gained

 Our conclusions and Way Ahead for future SEQ programs

1 AIRBUS DEFENCE AND SPACE PROJECT / PRODUCT CHARACTERISTICS

The portfolio of Airbus Defence and Space consists of manned Military Aircraft, Un- manned Aerial Systems, Communication and Information Systems, Border Security Sys- tems, Satellites and Orbital Systems. All of

these Systems have common characteristics that make a transition from classical SE meth- ods to agile SE methods particularly challeng- ing:

 High system complexity

 Stringent certification requirements

 Long system lifetime

 Development spread between sites and countries

 Development currently based on well- established and highly formalistic process- es

Despite these challenges, Airbus Defence and Space believes that the benefits promised by agile methods outweigh the difficulties of im- plementing them. Teaching and preparing our systems engineers how to solve these difficul- ties is one of the most important tasks of our SEQ program.

2 AIRBUS DEFENCE AND SPACE SYSTEMS

ENGINEERING QUALIFICATION PROGRAM

(SEQ)

The Airbus Defence and Space Systems En- gineering Qualification Program is a training program for mid-career engineers with around 7 to 10 years of professional experience. Par- ticipants are drawn from all of the Airbus De- fence and Space sites located throughout Europe.

The main purpose of this program is to pre- pare our most talented engineers to take the responsibility for components or disciplines within major projects.

The program is embedded in a larger Learning Path for Systems Engineers (a sequence of trainings and programs) that covers a career time frame from about 3 years to 20 years of professional experience.

(3)

2

Fig. 1: SEQ embedded in the Learning Path The SEQ program was originally created in 1999 by the Space division of the former DaimlerChrysler Aerospace AG. The program was so successful that the Defence and Secu- rity division of the company - in the meantime renamed EADS - started a parallel SEQ course in 2008 with a content specifically tai- lored to the needs of the defence business.

Following the formation of Airbus Defence and Space in 2014, the two SEQ programs were merged again under common governance. We did however retain the concept of running two groups in parallel with participants in the first group coming from the Space division (typical- ly 18 students) and those in the second com- ing from the Defence division (typically also 18 students).

Since 1999, a total of 579 participants have successfully attended the program.

The program runs for over one year with a kick-off event, 6 technical training modules of one week each, 3 social skill modules of 2 days each, and a final presentation to an audi- ence of senior managers and executives.

Around 90 lectures are given in the technical modules by around 85 lecturers. Most of the lecturers (around 80) are Airbus Defence and Space employees coming from all parts of the operational business.

Beside the training modules, the participants also execute a "real-life" project provided and supported by sponsors drawn from the Airbus Defence and Space Program organizations.

Typically, three projects are sponsored by the Defence organization and three projects are sponsored by the Space organization. This means that the average SEQ project team is

made up of six students.

Project progress is assessed three times dur- ing the one-year SEQ course by two Review Boards (one from Defence and one from Space), usually after the first, third and fifth technical modules.

Fig. 2: SEQ Program Sequence In Airbus Defence and Space, our most capa- ble systems engineers are responsible for technical project management of complex projects and the major work packages within them. For this reason, the SEQ program main- ly focusses on:

 Technical program management

 Commercial and legal issues

 Company strategy and customer relations

 Social skills

 Advanced Systems Engineering Methods

 Systems Engineering Specialities

This strong focus on technical program man- agement and advanced systems engineering techniques generates a common basis and understanding among the systems engineers passing through the SEQ course. Our expec- tation is that they will then act as multipliers and spread this understanding into the remain- ing SE community within Airbus Defence and Space.

We believe that it is essential for our systems engineers to have a good understanding of all technical aspects of their projects. Only then will they be able to make correct decisions when discussing problems and solutions with

(4)

3

their specialist engineers. To ensure that this is the case, we train our SEQ participants in what we call “Systems Engineering Speciali- ties”.

Accordingly, the Defence Group is briefed on defence-related Specialities:

 Overall Aircraft Design & Flight Physics

 Aircraft Certification & Qualification

 Image Processing & Multi Source Integra- tion

 etc.

while the Space Group is briefed on space- related Specialities:

 Spacecraft System Design

 Satellite & Spacecraft Propulsion

 Extra-terrestrial Science & Earth Observa- tion

 etc.

In total, the students receive around 185 hours of lectures in addition to Project Work and Events, Reviews and Administrative Meetings.

Fig. 3 shows how this time is distributed be- tween the various topics in the SEQ.

Fig. 3: Distribution of Topics in SEQ (hours) The program curriculum is constantly updated:

 To reflect the state-of-the-art in Systems Engineering

 To address new technical topics and emerging application areas

 To train new external or internal processes In 2018 and 2019, agile methods were intro- duced into the program in line with the evolv- ing Airbus Defence and Space systems engi- neering strategy.

3 EXPECTATIONS AND GOALS

3.1 IN AIRBUS DEFENCE AND SPACE

By introducing agile SE methods into the com- pany, Airbus Defence and Space expects the following benefits:

 Faster time-to-market

 Lower development costs

 Easier entry into new markets

 Improved customer relations by demon- strating continuous, visible and testable development progress

 Concentration on finding and improving what works

 Rapid reaction to changing customer re- quirements

 Early detection and correction of design errors

 Closer and more frequent communications within distributed development teams

 Improved team motivation by reducing unproductive documentation

3.2 IN THE SEQPROGRAM

By introducing agile methods into the SEQ, we expect to achieve the following:

 Align our Systems Engineering training with changes to company engineering practice

 Maintain our competitive edge by educating our best system engineers in state-of-the- art ways of working

 Turning these engineers into influencers that can bring the learned agile methods to their sites, domains, and projects to spread the agile mind-set and experience

 Ensure that our Systems Engineering edu- cation stays ahead or at least keeps pace with the SE transformation in Airbus De- fence and Space

3.3 STUDENT LEARNING GOALS AND SEQ COURSE GOALS

When we first introduced agile into the SEQ in 2018, we identified the following Learning Goals for the students in the agile projects [Bloom]:

Learn the basic principles of Agile Methods and the Agile Manifesto

0 10 20 30 40 50

Techn. Project Mgmt.

Commercial & Legal Company  Strategy…

Social Skills Advanced Sys. Eng.

Sys. Eng. Specialties

(5)

4

Understand what Agile is trying to achieve and why it is (or can be) more efficient than Waterfall Methods

Apply (tailor) the agile approach to their project task

Analyse their project and break it down to a series of demonstrable parts

Evaluate (check, challenge) plausibility of each part before completing it

Create a solution for the overall project task within the time frame and constraints of the SEQ program

In addition to the goals we set the students, the SEQ Design Board, which is responsible for the evolution of the SEQ curriculum, also had SEQ course goals of its own. These were:

 Test the effectiveness of our new curricu- lum

 Provide the students with practical agile experience in a controlled and ‘safe’ envi- ronment (if the experiment failed, we could still revert to classical methodology)

 Allow the Organizers, Review Board mem- bers and Sponsors to develop techniques for reviewing and giving useful feedback in agile projects

 Compare the execution of SEQ projects using classical SE methods with those us- ing agile methods

 Gather feedback from the students on how they experienced the agile training

3.4 METHODS OF EVALUATION

For reasons that we will explain in Para. 7.5, our assessment of how well the students had achieved the goals set out in Para. 3.3 was largely subjective. It was based on three major reviews with defined deliverables held at inter- vals of about four months. The approach we took was as follows:

Evaluate Goal 1 - Learn Agile/Agile Mani- festo and Goal 2 - Understand the method:

At the start of each review, the Review Boards were briefed on team behaviour and progress by the project sponsors and the resident coaches. Here, we were par- ticularly interested in how thoroughly the agile teams had learned and understood the method.

Evaluate Goal 3 - Apply the agile method:

The Review Boards evaluated documents prepared by the project teams in advance of the project reviews. These documents cover all major aspects of the SE process and allowed us to see how the students had applied agile to their tasks. Different documents and tasks are covered in the three reviews, which also allowed us to see how the students had broken their project tasks down and how they had built on pre- vious work to revise their approach.

Evaluate Goal 4 - Analyse the Project and Goal 5 - Evaluate sub-task approach: In each review, the project teams made de- tailed presentations of their project status.

Together with the documents mentioned above, this allowed us to see how the stu- dents had structured their project, how they had planned their time and resources, where they had placed their priorities and what progress they had made. Following the presentations, the Review Boards asked each team clarification questions to gauge how well the team had understood their task and the SE methods they were applying. The Review Boards gave the teams specific feedback on their progress together with recommendations for correc- tions and further work. Following the feed- back, the teams were encouraged to ask questions of their own, which gave us addi- tional insights into the training effective- ness.

Goal 6 - Create the Solution: During the three design reviews, the RB monitored the emerging solution to ensure that the teams were on-track. In the “Final Event”, the teams present their completed project to Airbus Defence and Space senior man- agement representatives. This is the last piece of information used by the SEQ de- sign board to evaluate the achievement of the learning goals.

This approach as described above was also used to evaluate the goals of the SEQ Design Board.

(6)

5

4 IMPLEMENTATION STRATEGY

Airbus Defence and Space is itself in the early stages of transitioning between document- centric and agile systems engineering. Be- cause of this, the SEQ Design Board decided to implement agile methods following an itera- tive approach consisting of:

 Lectures on agile methods for all SEQ stu- dents

 Providing agile training for all SEQ stake- holders (Sponsors, Review Board mem- bers)

 Providing guidelines to show how agile methods (like time boxing and backlog management) can be applied in traditional projects

 Performing two (of typically six) SEQ pro- jects using the agile approach, with the re- mainder using traditional SE methods. This approach was chosen to start small and to gather experience with the agile content

 Creating an agile process tailored to the SEQ program for the two agile teams 5 AGILE ELEMENTS IN THE TRAINING MOD-

ULES

There are six training modules in the SEQ during which the participants come together for one week at a single site. The modules consist of lectures on various systems engi- neering topics combined with free slots for project work.

In order to introduce Agile Methodology basics to the students, we reviewed the existing SEQ curriculum to:

 Identify available free slots that could be used to present agile methods

 Find a balance between the classical con- tent in the SEQ and the new material cov- ering agile methods

 Define a prioritized list of agile elements that could be presented under the given re- strictions

With these boundary conditions, we decided to introduce agile elements into the training mod- ules as follows:

Firstly, we explained the rationale for the agile approach (cf. Para.3.1), so that the partici-

pants could understand why agile methods are important and why they were introduced into the SEQ.

Secondly, we introduced definitions used in agile methods to the participants together with first concrete examples of how they are being implemented in the Company. This included a presentation of the Agile Manifesto [AM] cov- ering the principles:

 Individuals and interactions over processes and tools

 Working products over comprehensive documentation

 Customer collaboration over contract nego- tiation

 Responding to change over following a plan

The Agile Manifesto is the foundation for agile ways of working. For the SEQ, we felt that it was important to train the students to thor- oughly understand this manifesto and for them to discuss its benefits, implications and draw- backs in detail.

The Agile Manifesto presentation was com- bined with agile values and principles such as empowerment, transparency and visibility.

Business aspects are a key part of the SEQ training program, so agile methods like Lean Start-up and the concept of creating Minimum Viable Products (MVPs) were also introduced.

Thirdly, agile methods obviously have an im- pact on most of the classical Systems Engi- neering disciplines that are taught in the SEQ.

Examples are Contracting, Quality, Configura- tion Management, Model Based Systems En- gineering and many others. In the agile brief- ings, these impacts were highlighted and the presenters of these disciplines were also asked to address agile methods in their dedi- cated lectures.

Finally, we introduced the concrete processes that make up the agile mind-set to the SEQ participants. Here, we focussed on the most popular methods, namely Scrum [Scrum], Scrum of Scrums [SoS], Kanban [Kanban] and SAFe [Safe]. Special attention was given to roles like Scrum Master and Product Owner, artefacts like sprints and backlogs, and cere-

(7)

6

monies like planning meetings and retrospec- tives.

Time slots for project work are normally used by the teams to apply the content of the lec- tures to their projects. In addition to delivering formal briefings, the agile coaches also sup- ported the agile teams during their project work as described in the next section.

6 AGILE ELEMENTS IN THE PROJECTS

In the 2018 and 2019 SEQ courses, two out of six projects in each year were selected for implementation with agile methods. Project selection was done in close cooperation with the project sponsor since it was crucial that the sponsor supported the agile approach. This ensured that project expectations were har- monized between the sponsor, the SEQ de- sign team and the participants.

The kick-off and project initialization phase in the SEQ was used to introduce basic agile methods into the projects, to understand the task, create a vision, define a team charter and other points as follows:

1. Project goal and vision: The overall vision of the project was discussed, and different concepts of scoping, de-scoping, prioritiz- ing and gathering user needs were trained and applied

2. Scrum roles: The teams were made famil- iar with the typical Scrum roles like Scrum Master, Product Owner, and Development Team and given a dedicated coaching to apply them.

3. Scrum-like process: The main meetings, such as planning, stand-ups, backlog grooming, sprint demos and retrospectives were presented, applied and tailored to the project specific needs.

4. Team Charter: The team charter highlight- ed how the scrum roles are assigned, e.g.

rotating them in the teams so that every- body can experiment with them. Working concepts like using visual management tools, concepts of push vs. pull, creating generalists or specialists were introduced here.

5. Backlog: An initial backlog was created

using user stories and epics. These were estimated and prioritized using standard estimation techniques. It was important to train that this had to be done in cooperation with the project sponsor and groomed regularly.

6. Release planning: An initial release plan was created out of the established backlog to plan the number, order and content of it- erations.

7. Minimum viable product (MVP): The MVP concept was explained to illustrate that the agile approach can produce not only doc- umentation but also usable results in an early project stage. As a first step, the teams had focus on the highest priority items to discuss what "something most via- ble" would be for the project sponsor.

8. Review tailoring: The SEQ project work normally includes three classical reviews.

During these reviews, the students present the status of their projects together with supporting documentation. This is of limited value when using agile methods, so for the agile projects it was important to tailor the review approach and the associated review items.

As mentioned in Para.5, the Agile Manifesto was introduced to all of the students in the training modules. In the case of the agile teams, it was covered again in much greater detail since these teams had to learn how to apply it in their projects.

The SEQ contains lectures on many aspects of systems engineering such as risk manage- ment, quality management, configuration management, contracting, safety and so on as described in Para. 5. For many of these do- mains, we explained to all of the SEQ students how they can change when applying agile methods. Those students in the agile projects then had to create solutions for these topics in their projects.

Finally, we also introduced agile techniques like design thinking, open space formats, world cafes, self-organization and others into the agile projects.

(8)

7

7 EXPERIENCE TO DATE

7.1 EXPERIENCES WITH AGILE ELEMENTS IN THE

TRAINING MODULES

Fitting agile methods into the SEQ was quite challenging since the course was already densely packed and we were not able to ex- tend the duration. To make room for Agile, we therefore had to shorten other topics or re- move them entirely.

In 2018, we found that none of our students were really familiar with agile methods, so as a first step we decided to give them all just one introductory lecture. We further tested our approach by training Agile "hands-on" in two of the six SEQ projects.

At the end of the program we gave all of the students an overview on how to apply agile methods in classical waterfall projects.

As the 2018 course progressed, we quickly realised that the impact of agile methods on the other lectures was far larger than we had originally expected. Because of this, we great- ly increased the agile content in the 2019 SEQ. Instead of just treating it like an add-on during project initialisation of the agile projects, we integrated it completely. This tighter ap- proach did however, lead to the disadvantage that we had to invest additional effort to train both agile and classical SE methods in paral- lel.

Following this experience, we decided to rede- sign the next SEQ course in 2020 to be com- pletely based on Agile. Only agile methods will be taught and all of the projects will follow the agile approach. A first estimate showed that we will have to include about 15 hours on agile methods in the 2020 SEQ. In parallel we made a more detailed analysis of the remaining SEQ lectures to determine if further changes were needed to implement the all-agile approach. In total, we found that about 20 lectures would be strongly affected. We contacted the lecturers to make them aware of the SEQ redesign and to discuss with them how they should adapt their lecture content. Since more than 80 of our lecturers come from within the company, they were also able to update us on the im-

plementation of Agile within Airbus Defence and Space. As a consequence the evolution of the SEQ program concerning technologies, processes, methods and tools is always in line with the evolution in the Engineering organiza- tion. What we found was that many depart- ments such as Legal, Finance and Quality Assurance are currently working to adapt their processes and ways of working to the agile approach.

In this context, management of and relations with our principle governmental customers seem to be the biggest issues we are facing during our transition to agile methodology.

This is because these customers are more used to specification fulfilment rather than being deeply involved in the responsibility for design decisions.

Clearly, we must reflect this customer attitude in the SEQ and train our internal customers (the project sponsors), to behave accordingly.

We are still considering how to do this in the best way since it is currently far from certain how our customer community will react to ag- ile-based projects in the future.

7.2 EXPERIENCES WITH AGILE ELEMENTS IN THE

SEQPROJECTS

In 2018, the general briefings on Agile were received with much interest but we did not see that the students in the "conventional" project groups attempted to apply these methods in their projects.

On the other hand, we were impressed with how fast both agile teams were able to learn the methods and how quickly they able to ap- ply them. It was however very clear that this was only possible due to the intensive and expert coaching they received.

During the year-long SEQ, we often found that it was important to repeat the message that the agile way of working implies continuous (re-) planning to properly reflect sponsor feed- back. We stressed that this feedback, gained in sprint demos or the formal reviews, should be welcomed and can influence the backlog, planning and architecture of the overall solu- tion.

(9)

8

At the end of the SEQ 2018 we compared the outcomes of the "conventional" projects with those of the "agile" projects.

Based on our evaluation methods, what clearly emerged was the higher degree of team em- powerment and the continuous self-reviewing in the agile projects. These aspects, together with the value-based task sequencing, allowed the agile teams to produce concrete results faster once the initial learning phase had been completed.

In 2019, we observed that the two agile teams perceived the task of implementing agile methods in quite different ways. Whereas Team A embraced the concept enthusiastically from the start, Team B was more reluctant and questioned the basic viability of the approach.

This was particularly apparent in the following areas:

Agile Mindset (i.e. how the Agile Manifesto and its principles were understood and ac- cepted by the teams)

Adherence to the scrum-like process (how the teams applied the Scrum-like process described in Para. 6)

Agile project management (especially goal/vision, release planning, Scrum roles, backlog management and MVP as trained in the modules)

When we looked at the way Team A dealt with these points, we noted the following:

Agile Mindset: The team fully embraced the agile mindset. For example, they greatly appreciated the value of retrospectives as a means of continuous improvement, fre- quently requesting training in new retro- spectives and applying them regularly.

They also liked the concept of "PULL" in- stead of "PUSH". All in all, the team per- formed well, had a good team spirit, com- mon goals and shared their knowledge with each other.

Adherence to Scrum-like process: Team A tailored their approach by using Sprints, nominating a "Scrum Owner" (who acts both as Scrum Master and Product Owner), but they chose not perform Sprint planning, reviews, and other Scrum ceremonies. We

considered this to be a good compromise, since these ceremonies were not useful in the frame of this particular project.

Agile Project Management: The team was greatly impressed by the benefits of user story creation, prioritization, estimation, and release planning. They also used these tools when communicating with their spon- sor.

When we compared how Team B had dealt with these three topics, we noticed some sig- nificant differences:

Agile Mindset: When the project started, Team B was initially reluctant to accept the agile mindset, seeing it as more of a bur- den than a benefit. Their attitude changed later, with some of the team members even asking for support in their regular company projects. This suggested that our concept of creating "influencers" has been success- ful.

Compared with Team A, Team B did not really form an integrated team, preferring instead to allocate their project tasks to the individual members.

Adherence to Scrum-like process: This was not really done. For example, whereas we felt that the concept of timeboxing could have helped them, they chose not to apply it.

Agile project management: Team B made good use of agile concepts including scop- ing/de-scoping, user stories, epics and MVPs like models, simulations, and proto- types. They learned a lot from the MVPs, and their user stories helped them to modu- larize their system and selectively detail the parts they felt were most important. This team also created very good traceability from Epics to User Stories, Functional Re- quirements and System Design.

By the end of the projects, the two teams had tailored their respective approaches to agile quite differently, but in a way that worked well for both of them. We believe that the differ- ences in their approaches are due to the com- pletely different nature of the Defence and

(10)

9

Space business areas and reflect reality in the Company itself.

7.3 EXPERIENCE WITH THE AGILE MANIFESTO

In addition to the agile elements mentioned above, the teams also applied the Agile Mani- festo in their project work. The principle of

"responding to change over following a plan"

proved to be very useful since it allowed the teams to cope with the changing expectations of their sponsors as the projects progressed.

They found it was easier to respond to these changes by creating an overall plan with more detail for near-term iterations and less details for later iterations.

Providing "working products over comprehen- sive documentation" was perceived as chal- lenging because the teams first had to learn what the initial working products should look like for their specific project. As time went on however, we observed that they became profi- cient at working with user stories and creating products for them. This was a very good way of stimulating sponsor feedback during demos.

With regard to documentation, the Review Boards expected pre-defined documents spe- cifically tailored to the agile projects. What we saw was that the teams needed quite some time to understand what they should deliver and what not. Following this experience, we decided to revise our definitions for deliverable documents in the all-agile 2020 SEQ curricu- lum.

"Individuals and interactions over processes and tools" was understood and applied very well by the agile teams.

"Customer collaboration over contract negotia- tion" was understood by the teams to mean frank discussions with the sponsors regarding project objectives followed by adaptations where necessary. This was implemented by defining user stories and prioritizing them to- gether with the sponsor.

7.4 EXPERIENCES WRT. REVIEW BOARDS AND

PROJECT SPONSORS

As we were at the very beginning of the im- plementation of agile methods within the com- pany, most of the Review Board members and

the Projects Sponsors only had a very limited knowledge of agile methods.

We provided the Review Board members with agile training before executing the first project review. We also found that continuous support from the agile coaches was very helpful during the preparation and execution of the reviews.

The agile project teams requested more ex- change with the Review Board as well as more frequent but shorter reviews. This led us to question whether having just three reviews during the project execution time will be suffi- cient in an all-agile SEQ. More intensive cus- tomer interaction is, after all, a central feature of the agile approach.

Currently we see no way to fulfil this require- ment due to the limited availability of the Re- view Board members. To mitigate this issue, we have asked the Project Sponsors to meet with their project teams more often and to give them more frequent feedback.

7.5 FACTORS INFLUENCING LEARNING ASSESS- MENT

Ultimately, our assessment of the learning effects was based on how well the teams had completed their projects. In the 2018 and 2019 SEQ courses we specifically wanted to see if the agile projects were more or less successful than the waterfall projects. We believe that we were able to do this, but for completeness we would like to point out some of the difficulties we encountered:

 Projects are different in terms of scope, complexity and familiarity to the participants

 Random composition of the teams and their expertise

 The considerable freedom given to the teams to tailor their projects

When we made our assessment we were very much aware of the above problems. On bal- ance however, we reached the conclusion that our concept for incorporating Agile into the SEQ had worked well and that we should pro- ceed to the next step of running all projects in Agile.

(11)

10

7.6 FURTHER OBSERVATIONS

In addition to the experience we noted in the previous paragraphs, we would like to pass on some more anecdotal feedback from the stu- dents themselves. When reading these com- ments, we should remember that they were trying Agile for the first time. Typical remarks were:

 “Agile is good for software, but it is hard to adapt to hardware”

 “The procedure of Sprint Planning, timing etc. is too rigorous; too fast, too stressful.

The process should not be strictly applied”

 “Hard to really complete items properly; we had to return to them every time we started the next one”

 “Hard to maintain a fixed schedule because the team members come from different sites”

 “Learning the process was a burden, but only at the beginning”

 “Really liked the focus on the ‘User Point of View’”

 “Continuous adjustment in project was very good”

 “In real life, the Space business area has a much closer relationship with the Customer than the Defence business area”

 “Not sure how Agile will work in very large projects”

 “Acceptance of process varied widely among team members”

Although somewhat unstructured, we found this feedback very valuable for our further planning.

8 FULFILLMENT OF EXPECTATIONS

8.1 IN AIRBUS DEFENCE AND SPACE SEQPRO- JECTS

In Para. 3.1 we listed the expectations of Air- bus Defence and Space with respect to the introduction of agile methods. Of course we are not able to do an evaluation for the overall company, but we will try to make some re- marks with respect to our experience in the SEQ training program.

We would like however to stress that the SEQ project environment is a special one as de- scribed in chapter 7.5.

Faster time-to-market: In 2018, for the first time in the SEQ, one of our agile projects produced a component that could be inte- grated and work in a real system. Based on this experience we are convinced that the agile approach will help us to shorten time- to-market even if the capabilities generated will not be the final ones.

Lower development costs: As the SEQ projects do not cover the complete devel- opment cycle it is not possible to do an evaluation of this expectation.

Easier entry into new markets: This is out of Scope for SEQ.

Improved customer relations by demon- strating continuous, visible and testable development progress: This is something we observed in our projects. The agile pro- jects were able to show to the sponsor a continuous output as opposed to a com- plete solution at the end of the project.

Concentration on finding and improving what works: We observed a greater focus on finding what works through the MVP ap- proach. Several different types of MVP were defined in the projects, such as Matlab models, 3D printed prototypes, GUI mock-ups, and Hardware breadboards.

Rapid reaction to changing customer re- quirements: This is something we actually observed in one project. The Project Spon- sor (Customer) changed the focus of the project halfway through the course. Despite this, the project team was able to follow the change and still come up with a viable re- sult.

Early detection and correction of design errors: We did not observe this situation in the SEQ projects.

Closer and more frequent communications within distributed development teams:

While we saw this happening, we also saw that the SEQ structure prevents the team members from interacting as frequently as they wished. To improve this situation, we

(12)

11

will adjust the SEQ curriculum in 2020 to al- low for more discussion time.

Improved team motivation by reducing unproductive documentation: All SEQ teams were highly motivated. We did how- ever observe that the agile teams were more focused from the start on tangible outputs rather than documentation.

8.2 IN THE SEQPROGRAM

In Chapter 3.2 we listed our expectations with respect to the SEQ Program.

Align our Systems Engineering training with changes to company engineering practice:

Currently, the introduction of agile methods in the SEQ is mainly based on the Airbus Defence and Space engineering strategy.

In order to ensure that we are fully syn- chronised with this strategy, we are coordi- nating directly with the organisational unit within the Company that is responsible for implementing Agile. Examples of this coor- dination are that the curriculum for SEQ 2020 was defined together with this unit and that they also supply most of our agile coaches.

Maintain our competitive edge by educating our best system engineers in state-of-the- art ways of working: Systems Engineering is considered to be a key competence with- in Airbus Defence and Space. In 2019, Systems Engineering had one of the high- est priorities in the Competence Develop- ment Plan. The planned SEQ re-design is embedded in the overall activities of the Company related to Systems Engineering.

Turning these engineers into influencers who can bring the learned agile methods to their sites, domains, and projects to spread the agile mind-set and experience: We are confident that this will happen and have al- ready seen small examples of it working (see also Para. 7.2)

Ensure that our Systems Engineering edu- cation stays ahead or at least keeps pace with the SE transformation in Airbus De- fence and Space: This goal will be met by the curriculum for the SEQ 2020 and the subsequent updates.

8.3 FULFILLMENT OF STUDENT LEARNING GOALS AND SEQCOURSE GOALS

Based on the student Learning Objectives defined in Para. 3.3, we can report that the agile project teams in SEQ 2018 successfully passed all reviews and the Final Presentation.

To date, the agile project teams in SEQ 2019 have successfully passed all three reviews.

The Final Presentation is scheduled for De- cember 2019 and we expect that this too will be successful.

Our overall conclusion is that the students have achieved all the learning objectives we set them.

Concerning the SEQ course goals of the De- sign Board (see also Para. 3.3), our conclu- sions are as follows:

Effectiveness of new curriculum: In princi- ple our 2018/19 experiment was success- ful, but we realized that running classical and agile methods in parallel was inefficient and somewhat confusing for the partici- pants. This is why decided to run future SEQ Programs completely in Agile.

Provide practical agile experience: This was successfully done for the agile pro- jects.

Educate Review Board members and Sponsors: The briefing for Review Board members was sufficient for the SEQ. We observed however, that agile projects with a sponsor experienced in Agile run better than projects where the sponsor is inexpe- rienced. The short training that we offered was not sufficient for these inexperienced sponsors. Our mitigation from 2020 on will be that sponsors not versed in Agile will re- ceive continuous support and on-the-job training from the agile coaches.

Comparison of agile vs. classical project execution: As described in Para 7.5, it is difficult to extract the explicit effect of agile on the project outcome. What we can say is that the agile projects were completed as successfully as the classical projects. How- ever there was a stronger focus in the agile projects to produce tangible outputs (MVP) at an earlier stage.

(13)

12

Gather feedback from the students: Gather- ing feedback from the students in a Retro- spective format was successfully done and incorporated in the redesign for SEQ 2020.

9 CONCLUSIONS AND WAY AHEAD

As a teaching experiment, we felt that our pilot program to introduce Agile has been very suc- cessful, but it also exposed some limitations as indicated above.

We have taken the lessons learned from the pilot programs and used them to design an all- Agile SEQ course starting from 2020. In paral- lel with the new content, we will closely moni- tor the real-life experience of Airbus Defence and Space as it continues to roll out Agile Sys- tems Engineering, especially in the area of customer relations. Where appropriate, we will adjust the content of the SEQ to reflect this experience.

We expect that in three to four years the agile approach will be fully implemented in the En- gineering organization of Airbus Defence and Space. At that point of time, we believe that every new SEQ participant will have a good knowledge in agile methods before they enter the SEQ program. This means that we can then remove the lectures on basic agile meth- ods to make more room for more technical content or other new advances in Systems Engineering.

10 REFERENCES

[AD] Airbus Defence and Space www.airbus.com

 Defence &  Space [AM] Agile Manifesto

http://agilemanifesto.org

[Bloom] https://tips.uark.edu/using-blooms- taxonomy/

[ESA15] Stefan Brüggemann (Airbus), David Escorial (ESA): Challenges and op- portunities for Software Product As- surance with the introduction of Ag- ile methods, ESA SWPA Workshop 2015

[ESA17] Stefan Brüggemann (Airbus) Manuel Fernandez (ESA): Status of Agile

ECSS Handbook, ESA SWPA Workshop, 2017

[ESA19] Stefan Brüggemann (Airbus) Manuel Fernandez (ESA): Status of Agile ECSS Handbook, ESA SWPA Workshop, 2019

[Kanban] Kanban (Development)

https://en.wikipedia.org/wiki/Kanban _(development)

[LM] Lockheed Martin Case Study, 2018 http://www.omgwiki.org/MBSE/lib/ex e/fetch.php?media=mbse:patterns:is 2018_-_case_study.pdf

[SAFe] Scaled Agile Framework

https://www.scaledagileframework.c om/

[Scrum] Scrum.org https://www.scrum.org/

[SoS] Scrum of Scrums

https://www.agilealliance.org/glossar y/scrum-of-scrums/

Références

Documents relatifs

The application of tools for development of the open online courses and promising ways of use of free access resources to improve the technologies and electronic means of

thesis contributes to the knowledge base of software development by providing a process model that offers an iterative and incremental improve- ment of product quality in

In einer sich schnell wandelnden Welt werden agile Methoden im Software Engineering allein schon aus wirtschaftlichen Gründen immer interessanter und nehmen auch in den Lehrplänen

In this lecture, students are introduced to a rationale model, to capture and exploitation methods, and tool support for rationale management.. The goal is to motivate them to

To let participants already experience these difficulties and to work on making agile hard- and software development a reality, we propose the format of Interdisciplinary System

Four critical success factors for ICT4D need to be satisfied before an Agile method can work: the projects needs to be demand-driven, skills pertaining to Agile need to be taught to

The Group Methodology of Using Cloud Technologies in the Training of Future Computer Science Teachers.. Oleg Spirin 1 [0000-0002-9594-6602] , Vasyl Oleksiuk 2 [0000-0003-2206-8447]

The basic idea of the method is to use a combination of DEMO methodology and the theory of transac- tion costs [5] to build a single quantitative method that could be used