Skip to Content

Open Source @ SDN?

I believe that the SDN-community has an enormous creative potential. We use to discuss our ideas online and scetch programming techniques in weblogs. Unfortunately, a consequence of ABAP software logistics is that it is quite difficult to contribute the results of our work to other ABAP developers like we could do at for example.

But indeed it would be possible to start Open Source projects in ABAP. Ideally, from a technical point of view we would need

  • a central Web Application Server,
  • an administrator,
  • an own namespace (perhaps /SDN/),
  • programming rules concerning package dependencies.

Of course, we would also need rules for joint work: what has to be done before starting a new software project, for example.

A more minimalist approach would call for a repository to store transport requests. Although distributed development would be impossible, interested parties could at least import the requests into their own systems, thus obviating the need to manually recreate lots of development objects. A dedicated namespace and rules concerning the naming and allowed dependencies of packages provided via this service would be necessary nevertheless.

However, at first we should answer the question whether there is a need for Open Source software based on SAP and whether this community can manage it.

Or do think this is this too ambitious? But perhaps this suggestion could lead to platform that allows SDN members to present their prototypes/solutions to other SDN members?

What do you think?

You must be Logged on to comment or reply to a post.
  • Hi Tobias,

    I love the idea, it would be great if for all new ABAP-related weblogs the source would be available in the system and we can have a look more interactively, then a video. Really feel what is it about.
    From the otherhand it would need a lot of organizing, make sure only bloggers can maintain their code,  but everyone interested would have only display authorization.
    Considering sometimes the level of users in the forums, they cannot respect basic rules, it would be though to force them to respect also the rules in such system.
    However, it definately an excellent idea.

    Looking forward to see some progress on this topic.


    • I think we should all read the seminal article called ‘The Cathedral and the Bazzar’ which speaks about the fundementals of open source movement and then ask ourselves our question, whether Open Source is relevant for SAP.


  • Hi Trapp,
        I am completely satisfied with your thought on ‘Open Source software based on SAP’, surely it will be very helpful for all Techies across the Globe, only thing is we need a good moderator(probably from SAP Labs) who can manage us. Overall it is really a positive thought for a better tommorow in this amazing field called SAP.

    Thanks & Regards:-

  • Hello Tobias,

    that’s an excellent suggestion I’d love to see implemented. But as Peter points out, it will be very difficult to strike a balance between enabling people to contribute to a project without at the same time allowing them to disrupt it. However, I’d be very glad if something along the proposed lines could be made to work.


  • My favorite, or at least one of my most favorite topics and in fact I have it on my list to write a weblog on this very subject this week to update everyone on where we stand since we last approached the community during TechEd ’05.

    I guess I need to step that one up a notch now, huh Tobias?


    • We first need a tool that will help us share code, then we can worry about the intricacies of “managing” those open source projects. 

      If the SDN Gods smile upon me ( and to a great extent the community at large ) you may see something from the community pages jump to life at the Demo Jam this year that could be the first step we need.


  • Tobias,

    I agree with you 100% that we should be leveraging SDN more to work on more open-source, collaborative projects.

    I am not sure the solution is a central Web Application Server and an SAP administrator though.      I do agree that we need something like sourceforge, but sourceforge is far from being a central instance with a multitude of programmers trying to coordinate developments while locking each other out of objects.  The communication efforts and logistics of that could possibly turn out to be a nightmare.

    I believe the open-source/collaborative process can thrive here at SDN if we come up with a way for developers to share developments easily.

    Another key point is that most, if not all, successful open-source projects start out with a functional alpha version of a somewhat useful product.  IMHO, developers will only spend time and be truly interested in a project if they are building on something that is already useful and can be made better.


    • Hi Edward,

      you are right – development coordination will be a difficult task. On the other hand we have a powerful package concept that can help us.

      And you are right – we need a useful product, otherwise soon we will have dozens of sample applications and sophisticated generic tools that could be helpful to write better software but nothing that attracts attention.

      I’m curious about SDN day and I hope we will have a list of proposals of software projects. Perhaps we should start to collect ideas. Could this be the next SDN challenge?


      • What’s stopping all of you? SDN has provided the pages inside of the Contributor’s Corner to list any and all projects started related to SAP technologies, 5 or 6 are there.

        Also just today a blog was posted to revive the KM Weblog package project.

        You have a forum called “Technology & Tools” which you can start a thread and collect ideas – if you wait for someone else to do it do you think it will happen?

        I won’t make an extra “challenge” on SDN just for this – you have the tools and the means already the challenge is whether or not you will all step forward and do something ? 😉

        • Amen Craig…glad to see that you are earning the word “evangelist” in your title.  🙂

          Regarding the mentioned blog, Open Source Project @SDN – Weblog System using KM
          It will be extremely hard, for that project to ever come to fruition.  In the blog, you can see that Arul is already feeling the agony of trying to coordinate a collaborative, open source project with typical waterfall, corporate project management methods.  He says “This project has shown us that it is very difficult to finalize the scope and the choice of technology by discussing the feature requests with all the developers.”  This sounds like pure misery to be involved with.  Why would any developer touch this?  I get enough of these types of projects everyday at work.

          Sometimes I feel as a bit of an evangelist myself, but you cannot have a successful project of this type without an alpha version of a useful product.  The product has to be interesting with potential for improvement.  Developer’s will then want to get involved because they see the benefits, not because they just love having even more phone conferences with 100 people involved trying to define the scope of another destined-to-fail project. 

          Open source, collaborative projects should be an outlet for a developer to get away from the typical “analysis paralysis”, corporate red tape..a place where they just get to finally step away and let their creativity flow.

          Thorsten says it best in his earlier response when he says “All the successful Open Source projects I can think of are about creating software that is essentially redundant, but highly ambitious and fun.”

          So in the end, I have to give the biggest AMEN to Thorsten….Amen, Mr. Franz.


          • Hello Edward,

            I think I completely share your perception of today’s blog on the KM project. When I read it I immediately wanted to fire a cynical comment, but since I really do not want to disencourage anyone and do not bring down someone’s motivation.

            But – I immediately thought that this is an example of totally not understanding how sucessful opensource development in most cases works.

            No such project to my knowledge starts with some manager or architect and certainly not with a rigid organistorial structure. Almost any such project starts out with someone, who has already created something, sharing this something with the public or community, hoping to gain some sort of respect and finding some people who join in to further develop that something.

            I think the only opensource projects that are different are those which are transferred from a closed domain to the public domain.

            If it really worked the way the author wants to see it, it could become really absurd, e.g. in that some corporate manager starts an opensource project, finds some %&&$Ӥ programmers, whom he can milk and then run with the result and sell it. A crazy idea actually.

            Another strange idea is the one of a central development platform sponsored by -hmm – by whom?
            I have never heard of any such thing. Even big stuff like linux is in no way developped like that.

            Somehow it’S really funny how easily one can identify big-corp-thoughts.

            Anyway, i wouldn’t mind if I’ll be proven wrong in the future by witnessing some successful SAP style opensource paradigm. Probably I’d even take part one or another way but only after I’d gathered 20 personal lawyers around me to protect my IP (or the one donated to the public domain) from the big corporation’s hunger 😉

          • Hi Anton,

            “No such project to my knowledge starts with some manager or architect and certainly not with a rigid organistorial structure. Almost any such project starts out with someone, who has already created something, sharing this something with the public or community, hoping to gain some sort of respect and finding some people who join in to further develop that something.”.

            We had same opinion when we started the opensource project. However we found it very difficult to come to any kind of conclusion about the choice of technology or scope of the project. The project needs a starting point. I think that developing a protoype with some rigid rules (as you put it) would be better idea to start with before developers implement their innovative ideas.


  • Great idea, and I’ll be part of it as soon as it starts. But why do something useful? All the successful Open Source projects I can think of are about creating software that is essentially redundant, but highly ambitious and fun.
    Come to think of it, we ought to create a free Open Source alternative to something awfully expensive and proprietary… oops, I think I better stop here! 😉
  • DOXYGEN ( is a tool for creating documentation (html, pdf, etc) from sourcecode. DOXYGEN supports C, C++ and many other programming languages, but not ABAP.
    Having a tool such as DOXYGEN would be a great help for keeping my ABAP documentation up to date.
    DOXYGEN provides tools to add new languages, but it still takes time to figure out how these tools work and to apply them.
    I would vote for this (relatively easy?) project to start with as an open source project. Having DOXYGEN for ABAP could also be of help during the development of other ABAP Open Source projects.
  • Hello Tobias,

    I think it’s great.

    Notwithstanding accepted notions of open-source distributed development… I would approach it in the same manner as Thorsten “is essentially redundant, but highly ambitious and fun.”.

    If the overarching goal is to harness SDN’s “enormous creative potential”, then Project Management, usefulness of the product, even the final product becomes a technical detail.

    I don’t think it will lack the volunteers to develop – I personally know of a small group of people keen to collaborate. They just need the venue. Logistics will be an issue at first.

    Get down to : connectivity, a 80% of a central idea to develop, rules of engagement, tolerance for ambiguity  and you’re there. 

    Given a semipermanant core group leading base code and 90% of us that goes in and out can provide your contributory DNA soup for creative background.

    You bring the environment, I’ll bring some code, A mate will bring the chips (that’s fries for you folk stateside 🙂 ).

    Wilbert Sison

    By the way : In response to Thorsten’s entertaining comment : “we ought to create a free Open Source alternative to something awfully expensive and proprietary… oops ” : Just my opinion – but collaborative ERP application that can be developed in this method CURRENTLY does not provide the same value-proposition to the corporate marketplace where SAP competes at.  When running mission critical business applications, I’d stick with standard commercial ones for now. We can always hope for good things though.

  • I wonder about how or at least how well this would work.  A few questions…

    1.  Any code developed on a client’s/employer’s time and/or system belongs to them, right?  Developers would have to be very careful.

    2.  Having an interpreter is good, but is the idea to redevelop the entire SAP system in open source?  If not, you still have the problem of writing code that means anything without access to the SAP dictionary and business objects.

    3.  How would it be different from installing one of the downloadable development platforms from SDN?  You can have a fully functional SAP system on your personal computer now.

    These are just thoughts from a quick read of the blog.  I am very interested in the idea, just trying to get some clarity.


    • Replies:
      1. Yeah, developers need to be very careful.

      2. See, what I intended to say in my last post is “lets re invent the wheel” by developing ABAP interpreter in C. SAP has already ported r/3 on GNU/Linux, but am not sure about its policies and all that. I want to go on similar lines as MONO. Microsoft says that “.net platform” will be platform independent but they never relaesed a version for GNU/Linux operating systems. So Miquel De Icarza(GNOME and MONO developer) and his team thought to write mcs compiler(for C#) themselves.

      3. Nothing to say.