Over Atos Origin
Contact
Diensten
Nieuws
Home
Feedback  |  Atos Origin.com  |  Syndicatie
Syndicatie
 
Links
Archief
Meer over de bloggers
Onderwerpen
Alle onderwerpen
Agile
Algemeen
Architectuur
Besturingssystemen
Bpm
Business intelligence
Business proces design
Eai
Governance
Ibm
Integratie
It consultancy
Java
Microsoft
Nljug
Open source
Oracle
Process standaarden
Project management
Requirements engineering
Rich internet applications
Saas
Security
Sharepoint
Soa
Testing
Virtualisatie
Xml

OutSystems Business Case: OutSystems is Highly Productive
Geplaatst op 15 July 2009, 22:04 door Hans ten Berge in java, microsoft, oracle, open source, integratie, project management, architectuur, it consultancy, agile, algemeen, governance, process standaarden, business intelligence

A study conducted by Atos Origin measuring the productivity of OutSystems' Agile Platform shows that OutSystems is 22% cheaper for Agile software development. With OutSystems, the study delivered equal levels of quality in a shorter time period even with an increased workload.

The original blog post was in dutch but lots of request came from other languages to translate this blog, that is why the original blog by Hans ten Berge is translated here.

This study measured  the productivity of a project executed by Atos Origin for one of its customers using OutSystems' Agile Platform against the productivity of other Agile projects at Atos Origin; non-agile projects implemented in 4GL languages; as well as projects in a global Quantitative Software Management (QSM) database.
We measured the following three criteria: "Development Scale of effort" + "Software" + "Development time."
Note: In this blog I will not describe OutSystems'Agile Platform, for this, I refer to a previous blog: OutSystems , Agile Application Lifecycle Management.
Within this blog I discuss the productivity characteristics of OutSystems and the project which has been successfully implemented.


Starting points:

This was the team's first project with OutSystems. The people involved in this project have participated in the online certification and a 2 week bootcamp. They are experienced .NET software engineers but have never implemented a project with OutSystems before. The OutSystems project was a new experience for the project members, except for the engagement manager, committed to the project. The Engagement Manager was the liaison between the business and the IT team and is a crucial role in translating business requirements into development requirements.


Project Characteristics:

  • Project size: 444 function points
  • Development methodology: Agile, OutSystems Agile methodology  (Time Boxed by 2 week Sprint)
  • New application project
  • Contract type: Fixed Price / Fixed Date
  • Team Size: 1 Engagement Manager, Delivery Manager 1, 3 Developers
  • First OutSystems project for the developers
  • Team skill Background: Microsoft .NET developers


Comparison criteria:

The study compared the new OutSystems project with:

  1. An agile project in Microsoft. NET in the Atos Origin Software Factory (the SDMC)
  2. A Unified Process Microsoft 4GL - all other 4GL languages under Microsoft. NET, in the Atos Origin Software Factory (SDMC Software Development & Maintenance Center)
  3. A comparison with international QSM traditionally conducted business in 2005 (This is the global benchmark in the QSM Database).

For the record, there were no comparison made with 3rd generation languages (3GL's like Java).  Note: I do not include figures in this Blog from other projects due to the sensitivity of the information, however, that type of information can be shared with customers by special request for customers that want to start a project with Atos Origin and Outsystems.


Productivity comparison plan:

Atos Origin registers project metrics from the various projects in a specific database using the SLIM Data Manager, the QSM device for setting up project metrics in a company specific database.  Larry Putnam, founder of QSM defined the QSM Software Process Comparison and implemented in the SLIM (Software Lifecycle Management) toolset.
A comparison plan was devised to compare Agile projects implemented in .NET; traditional, non-Agile 4-GL projects; and all projects from the QSM database (non-Atos Origin projects). The last of which contains the comparison of all measured data and projects that provided a type of international benchmark.

Next, project metrics were defined for both planned and realized projects:

  • Turnaround (Start and end of project)
  • Effort by business group (in hours)
  • Size (in Function Points and/or Lines of Code)
  • Quality during the project (number of defects)
  • Global Sourcing Ratio (to run successfully in e.g. in India)
  • Influence factors on productivity (this includes a number of parameters under which can influence productivity, including: clean order, complexity, development process, project conditions)
  • Other specific KPIs
  • Specific project conditions


Conclusions:

Development with OutSystems has led to a very positive picture regarding productivity. OutSystems themselves claim two-to-four times higher development productivity and ten times higher application management productivity levels. In this study, Atos Origin concluded that an inexperienced team (with only initial training in OutSystems but who were experienced. NET developers) were 22.9% more productive with OutSystems Agile Platform than developing a similar project with a 4GL, and 29% more productive with OutSystems when compared to a comparable Microsoft .NET project done with Unified Process. When compared with the QSM global benchmark, OutSystems is 260% more productive.



Team Efficiency

The Team Efficiency (QSM PI - Productivity Index) is 20.8. The higher the better.
The Team Efficiency has to do with the processes and tooling. The result of these two parameters leads to productivity and efficiency. Overall, the team efficiency is good.

01-PIvsEffectiveSLOG


Time

The Time Pressure (QSM MBI - Manpower Buildup Index) is 6.8. This index measures the time of software development on a 14 point scale. If this index exeeds above 8 you have a problem with the timepressure within the team. Then results are no longer good, because the team is under "above normal" stress levels. The pressure with OutSystems was high but still good for this project. Generally, it can be said that the (positive) timepressures in Agile projects is quite high. You can consider this to be "healthy" time pressure.

02-MBIvsEffectiveSLOG 


The Quality

The Quality of the program is determined by the number of defects in the system test filtering. Of course, the lower the better. The quality for the OutSystems project was 78. That is not distinctive from other 4GL projects. This is just good. The international average is somewhat lower, but of course biased because this OutSystems project has been developed using Agile methodology, and an Agile project by definition is based on changes and refinements during the process.


Productivity and Turnaround

If productivity in function points is compared to an overall conversion factor e.g. SLOC (Service Lines of Code) for the construction phase, you can see OutSystems delivers a great deal of code in a short period of time. Looking at the x-axis that shows the lines of code versus time on the y-axis, you see that you have more code in less time than average. It is even beyond the normal scale of the number of lines of code. OutSystems delivers code rapidly.

03-TOREIvsEffectiveSLOG


Productivity and Effort

Finally, when we compared the cost per function point unit converted to Service Lines of Code (SLOC) against the effort (effort,) instead of time, we get the following picture.
Here we can conclude that OutSystems' Agile Platform is highly productive.

QSM Productivity and Effort
Productivity-influencing Factor

This number is determined by a set of 18 questions which are given a particular rating.  The higher the factor number, the more complex it is, resulting in negative impact on productivity.  I will not name all the factors but you can think of the programming language, presence of a standard project management as Prince2, clean order (to what extent the specification is ready and approved), technical complexity and other project conditions.  The productivity impact factor for this project was 51 compared to the standard average of 41. This was therefore a more complex matter than average.


Summary

Developing with OutSystems is very interesting. It is a highly productive development environment.  We only studied a new project implemented by a team that was new to OutSystems (although experienced in.NET). Given this, we can assume even higher productivity levels for a project being implemented by people experienced in OutSystems. In addition, OutSystems is a lifecycle management platform and claims to be 10x faster when it comes to managing the application but we have not yet investigated this.


For the development of new software projects, OutSystems' Agile Platform is 22.9% more productive than a standard Agile .NET web development project. Since it is a lifecycle management environment which allows you to manage the production application with one-click-deployment that can be monitored, maintained and managed, OutSystems will also probably be more productive when it comes to application management and change.


The duration of the project was more than 40% lower than average and the relative speed was more than 25% higher than an average project. The quality is comparable to other Atos Origin projects.


In times of shrinking budgets and rising ambitions for web and mobile applications, doing Agile with OutSystems represents a very good alternative. Where time-to-market is also crucial - OutSystems represents a good alternative when compared to other mainstream development paths.



Hans ten Berge
Principal consultant at Atos Origin 
Hans ten Berge BLOG: 


For more information: 
OutSystems , Agile Application Lifecycle Management
http://en.wikipedia.org/wiki/OutSystems
http://www.outsystems.com/

 




Share this | 2668 keer bekeken | 4 reacties
Reacties Syndicatie en RSS

Rachel reageert, op July 17, 2009 om 18:55 (GMT +01:00):

woohoo!! an English entry :)


Hans ten Berge reageert, op July 18, 2009 om 08:57 (GMT +01:00):

@Rachel

At your service :-)


Tiago Castanheira reageert, op October 19, 2009 om 17:14 (GMT +01:00):

1st of all, it's not Naam!!! I'm not some arabian piece of shit!!!
I don't doubt that the ousystem plataform is good... but 22.9% more effective than the .net platform?!?!? I disagree completely because your explanation isn't clear enough! Have you put your eyes on your graphs? Because they look like they where drawn by some 3 month old kid in microsoft windows paint!
Be more careful in what you say and please put correct english in your pages, if the english in the outsystems platform os as good as this you can put your platform in a bucket full of shit that no one will notice the difference!!!


Hans ten Berge reageert, op October 19, 2009 om 20:08 (GMT +01:00):

@Tiago

This study is conducted with great care. And if you would be a little bit more interested in the content than just juging the graphics you would know that these graphs are rendered from the QSM system that is world wide used.

And yes they look a bit old fashioned but it is the data that counts.

The tekst is translated from dutch, and yes that is not our native language. 

The study does not only look at the .NET platform but also at the approach the process the system that helps with deploying, the testing and the fact that you generate code instead of type it all togethter.

Next of all, if you read the whole blog you get your explanation of functionpoints and productivity. If i have for example a productivity of 10 hours per functionpoints in .NET and get a 6 in outsystems (the real figure is competitive information and I will not publish this) how much % is that?

Do you really think we would put these figures out when we did not check it in depth?

If you are considering Outsystems I can give the details of the study. And you have to contact me.

Hans






Reageer
 
 
Top artikelen
  • J-Spring 2009:Hop on board the Java Troubleshooting Platform
  • Softwaredocumentatie met Sandcastle
  • Developers zijn kikkers
  • Blik op kwaliteit door middel van Six Sigma
  • Visual Studio controlled light bulb
Recente reacties
  • Onndertussen is er ook een LDAP (en ActiveDirectory) ...
  • Livelink heeft ook (simpel) versiebeheer en WebDAV toeganghttps://km.atosorigin.com/livelinkdav/~nlXXXXXVoeg ...
  • Binnen AtosOrigin is deze link helaas afgeschermd: You ...
  • @Roel: bedankt voor die extra 250 MB kerel ...
  • Inderdaad erg handig!
  • Terms of use
  • Legal