Software Research and the Industry

8 responses so far ↓

• 1 Dirk Riehle of SAP Re-Invents the Wheel « James Dixon’s Blog // May 1, 2009 at 20:24

  [...] a comment » Dirk Reihle of SAP has just released a piece called The Commercial Open Source Business Model. He does a good job of giving a broad overview of the various models and how different business [...]

• 2 Dirk Riehle // May 2, 2009 at 20:12

  James, it looks like you are upset about my coverage of related work. This is a research paper, and here is how it works:

  (In response to your review.)

  Scientific research papers, to be recognized as such, need to go through a peer-review process. In this process, scientist peers who typically remain anonymous to you, review your paper, form an opinion on its quality, and recommend to publish or reject the paper. The paper in question made it through this process. This gives it the scientific stamp of approval commensurate with the prestige of the publishing venue, the AMCIS conference in this case.

  The hallmark of research is that it is not just an opinion, but something that other reasonable people, your scientist peers, can agree upon. That’s what the review process above tries to ensure. You typically have to prove that your work is not just an opinion but is some form of truth. That’s why most research papers have sometimes laborious evaluation and validation sections. In contrast, opinion pieces try to be interesting and engaging to readers, but are not considered research as long as they haven’t passed peer review. Scientific procedure makes papers boring but is necessary.

  Scientists need to demonstrate that what they are presenting is somehow novel in comparison to what came before. Thus papers have a prior and related work section. What’s relevant for that section is work that was published after it went through peer review and thus is considered scientific research work. Opinion pieces, as well worked out as some may be, are usually not part of such related work discussion. My paper reviews some in the related work section, like Bruce Perens’ work, but that’s only because the anonymous reviewers complained that I should be covering it. This doesn’t imply non-peer-reviewed work can’t be discussed, but it is less common and not required.

  The blogosphere is full of theories on how commercial open source works, and trying to cover that would have blown the 5000 word limit on the paper without me ever getting to what I wanted to say in the first place. I had read your Beekeeper model paper a while back but it did not resurface in my mind when writing this paper. I would love to hear how I could use it as primary material to validate, for example, the value of community for product management or the percentage-wise shift in sales and marketing to research and development expenses. I’ll be happy to revisit your paper for this.

  I hope this explanation helps explain some of the mechanics and decisions around scientific writing and publishing.

  Dirk

• 3 James Dixon // May 4, 2009 at 19:20

  Hi Dirk,

  Thank you for your post.

  You include statements like ‘According to Jacob’ and ‘Augustin estimates’ and many other ‘facts’ culled from blogs and OSBC presentations. I know Jacob, Asay, and Augustin personally, and they are evolving this model as they go along with everyone else. They also freely share their opinions on how to execute and improve the model. Your paper is an aggregation of other people’s opinions. You seem to think that a scientific paper that aggregates opinions is better than the original opinions.

  Thank you also for the explaination of the process. If I understand you correctly my work – which took 2 years of observation before I even started it and includes feedback from over 30 professionals in commercial open source and acedemia – cannot be counted as prior work, because I elected to use an open feedback process instead of a closed one, and thus it is only ‘opinion’.

  It brings us to an interesting topic. You identify how important peer review is in the process, yet your paper fails to mention how important peer review is in commercial open source. The scientific peer review process is ‘closed’, you only get feedback from a selected audience. Peer review in open source is unlimited.

  Take my Beekeeper work. I got some initial feedback from people that I know and included their comments into the piece. However the best feedback I got was from people that I did not know, and I only got this feedback because of the open nature of the process. These people include:

  Roberto Galoppini, open source consultant
  Matt Aslett, 451 Group
  Tarus Balog, Open NMS
  Paul Ramsey, The Open Planning Project
  David Dennis, GroundWork Open Source
  Michael Grove, CollabWorks
  Margaret Rouse, IT Knowledge Exchange
  Martin Michlmayr, foss bazaar
  Bernard Golden, author and CEO of Navica
  Dick Selwood, Embedded Technology Journal
  Wynn et al, University of Dayton
  Mark Hinkle, Socialized Software
  Martin P, Inflection Technologies
  Chris Katz et al, Tufts University
  Dr Pretra Malik, Victoria University of Wellington

  The scientific community has had the opportunity to use the internet to improve the openness of the peer review process for a long time but has not done so. Why is that?

  James

• 4 Dirk Riehle // May 5, 2009 at 00:18

  Hi James, I agree that the academic peer review system is due for an overhaul.

  And it is not just the review, it is the publication itself. Why wait for months until something is published but becomes accessible only for a fee? Open access is trying to change this, with some success.

  As to the peer review, the current system does not allow for mass review but only for review by a selected circle of experts recognized as such by their own peer-reviewed publications. Catch 22… But I’m sure all this will change.

  At present, if you want academic recognition, you will have to make your work pass academic peer review, e.g. by submitting it to a conference or a journal. Selecting your own reviewers doesn’t count I’m not sure why you want that recognition but under the rules of today’s game, that’s what you have to do.

  Dirk

• 5 James Dixon // May 5, 2009 at 19:25

  Hi Dirk,

  I’m not all that interested in academic recognition. The Beekeeper model is used in many computer science courses around the world already.

  An issue I have with journal submissions is their requirements with copyrights. Those also seem rooted in the past.

  James

• 6 Dirk Riehle // May 6, 2009 at 04:45

  Hi James,

  feel free to tell Elsevier and Springer, some of the stalwarts of academic publishing that they should stop insisting on copyright transfer agreements. (I already did.)

  Dirk

• 7 Jabu // Jul 7, 2009 at 16:57

  Interesting reading, which is always the case whenever open source pheneomenon is discussed. Unfortunately I cant take sides

  Jabu Mtsweni
  SAP Research, South Africa

• 8 Dirk Riehle // Jul 11, 2009 at 00:04

  Hi Jabu: SAP is much more active in open source than most people think. Please contact Erwin Tenhumberg internally; maybe I can get him to post some summary here. –Dirk

Leave a Comment