Episode 138: Learning as a Part of Development with Allan Kelly

Filed in Episodes by on June 15, 2009 15 Comments

Recording Venue: Skype
Guest(s): Allan Kelly

Host(s): Arno
In this episode, Allan shares his insights about how learning is a necessary part of software development. He covers the personal as well as the team and the organizational level and offers practical advice.

Links:

Tags: , ,

Comments (15)

Trackback URL | Comments RSS Feed

  1. Jooly says:

    Interesting discussion, and excellent interview. Too few podcasts debate the issues the interviewee is raising. this was refreshing in that some discussion occurred.

    While the classical view sees software development as largely separate from the wider organizational environment the organizational learning perspective embeds software development within this environment. From this point of view the act of software development has a role to play in higher order organizational learning by the company.

  2. mwdawson says:

    Something definitely went wrong with this interview. The topic is important and I was interested. Allen Kelly did have a problem with not getting concrete enough. What was distrubing was that Arno started asking very aggressive confrontational questions. Arno had a whole series of questions kind of like. “Whats a manager to do? Develpers can’t be allowed to sit all day just learning new stuff and not getting work done”. It seemed like Arno wanted to start some kind of debate. Kelly would not rise to these provocations. You could tell that he found it difficult at times.

    These questons were strange anyway. I have never worked in an organization where the big problem was developers spending too much time goofing around with new technologies. The problem is usually exactly the opposite. Developers are swamped with project work and don’t invest enough time in keeping current.

  3. mbrodersen says:

    I can’t speak for anybody else of course but I personally enjoyed the SE radio episodes covering DSLs, model-driven development and architectures. Those podcasts were informative and with a high signal to noise ratio.

    I find it interesting that, according to Erich, “A lot of software developers would consider those (DSL) episodes as vague and too abstract”. Really? I wonder if Erich has spoken to a statistically significant percentage of the people listening to the podcasts before reaching that conclusion? If that is the case then I for one would be interested in seeing the research data presented here.

    Cheers
    Morten

  4. epawlik says:

    I apologize that I have ignored that highly significant sample of size one Morten has presented as a convincing proof that the latest podcast is fluff.

    I have spoken to developers that did listen to those DSL episodes, some of them because of my recommendation. I personally liked some of those episodes very much (in particular the discussion between Juha-Pekka and Markus), but most of them didn’t. Although they are bright people, they perceived those discussions as too abstract – they just couldn’t see a relationship to their day to day work. Most of the developers I know are just more interested in the latest features of the Spring framework or of Mule or of Weblogic or in a timesaving Eclipse plugin.

    My colleagues were too polite to say that they feel it’s a bit strange learning things that cannot be used immediately. By the way, this is something Allan has touched on during the podcast several times.

    I guess that all those episodes on SE radio have listeners that consider a topic as too abstract or as very relevant to their daily work or just as thought provoking or enlightening or interesting. To my opinion, the broad coverage of topics relevant to the software engineering community is a strength of SE radio. That also means that for any episode there will be listeners that don’t like it.

    May I again suggest that we as listeners respect the expertise, thought and effort that interviewers and their guests are putting into a podcast. I strongly believe that unless we are real experts in the topic of a podcast, we should stay away from wording such as “Fluff”. And real experts in a topic tend to use other language to express that they don’t consider a contribution as very deep.

    Last point: I don’t need a statistically significant sample to proof that the topic of knowledge acquisition and learning is relevant to the vast majority of software development projects. That is evident.

    Regards

    Erich

  5. mbrodersen says:

    The difference between what I am saying and what Erich is saying is that I make it clear that I am expressing my personal opinion while Erich seems to elevate himself to speak on behalf of a large group of SE podcast listeners he does not know.

    If instead Erich had said: “I believe that …” then Erich would have expressed a personal opinion and people can then agree or disagree. There is a big difference.

    It is of course interesting to hear that Erich have talked to friends who thinks that the SE podcasts are too abstract. But it is quite a a jump to assume that Erich’s friends represent an accurate picture of what a majority of SE podcast listeners think :-)

    It could be that a majority of SE listeners find the podcasts to be too abstract. But then again it might not.

    One way to find out would be to look at how many times listeners return to listen to more SE podcasts. If the majority of SE podcast listeners come back for more then I believe it is strong evidence that the podcasts are not too abstract/boring. However, if listeners never return after having listened to a SE podcast then there is a problem.

    I for one have listened to almost all SE podcasts and in general the quality is high and the signal to noise ratio is excellent. However that was not the case in this podcast.

    Cheers
    Morten

  6. mbrodersen says:

    I normally enjoy the SE podcasts so please take this as an honest attempt to give constructive feedback that can help improve the podcasts.

    My impression of Allan Kelly is that he is sincere and have all the right intentions but unfortunately his message is fluff. In other words, vague, non specific, too abstract to be useful and in general consists of a collection of tautologies backed by long winded examples.

    Digging through the fog, the key message seems to be “only learn something if you can use it straight away (if you can put the learning into action)”. But later: “there is value in learning things that you can’t put into action right away (learning a new language)”. So learning is good. Right…

    I stopped listening after 20+ minutes. The only reason I lasted that long was in the hope that something interesting/useful would appear out of the thick fog.

    I highly recommend that you pick somebody to interview that has something interesting to say and knows how to say it. You have done that many times before but this time was a miss.

  7. Eek says:

    I too stopped listening half way through, as it was becoming toe-curlingly embarrassing to listen to Kelly’s fumbling attempts to clarify what he was saying when prompted by Arno. :s

  8. epawlik says:

    First of all thank you to Arno for setting up an interview about a topic that is both relevant and interesting for real world software development.

    After reading some posts on Allan’s blog, I also feel he has something interesting to say. During the podcast, he made some good points. On the other hand, I felt that the podcast failed because it has tried to cover far too much ground – learning at all levels, the relationship between learning and agile, practical tops for developers, project managers and executives, … .

    The relationship between software development and learning/ education/ knowledge acquisition is a deep and very practical topic. On the other hand, software engineers typically are not experts in this domain.

    Remember some SE radio episodes on DSLs, model-driven development or architectures. A lot of software developers would consider those episodes as vague and too abstract, simply because they lack the background knowledge and experience to follow the line of argument in the podcasts.

    I admit that I didn’t understand a lot of Allan’s arguments. I am not sure whether this was because the speaker didn’t know how to say it or because of my limited background knowledge or because of poor content. Therefore, I feel it’s unfair to classify the episode as fluff.

    Without an introductory episode about this topic, a typical listener to this podcast might not even by able to get the point in an interview. May I suggest that you set up such an episode.

    In addition, I would love to listen to an episode about the topic of learning in software development projects featuring Gerald Weinberg. He will be difficult to get in front of your microphone, but you also managed to interview Grady Booch and Tom De Marco – so there is hope.

    Regards

    Erich

  9. chrisawalik says:

    Arno had a whole series of questions kind of like. “Whats a manager to do? Develpers can’t be allowed to sit all day just learning new stuff and not getting work done”. It seemed like Arno wanted to start some kind of debate. Kelly would not rise to these provocations. You could tell that he found it difficult at times.

  10. The0retico says:

    I listened to whole episode and there was a silence gap, where the guest couldn’t do much more than smile in attempt to answer a reasonable question. It would be better if the guest introduced some practical techniques, because although everything what he says seems to be quite right, there is not enough reasoning at the level he talks in this interview, maybe concrete examples could be more convincing.

  11. rowinggolfer says:

    As an amateur hacker (so not your target audience) I have to say I enjoyed this episode tremendously. Some of Allan’s idea’s seem a little like modern management BS, but the take home message that developing is a process of learning even for professionals is one of comfort to me.

    Interesting discussion, and excellent interview. Too few podcasts debate the issues the interviewee is raising. this was refreshing in that some discussion occurred.

    Keep up the good work seradio.

  12. chrisawalik says:

    liked this because I’m a big fan of Philip Armour’s work in this area in The Laws of Software Process, where Armour argues early in the book that software is not a product but rather a medium for storing executable knowledge (and therefore software development is a knowledge creation activity). I also like the early reference to Peter Drucker’s work.

  13. tlooy says:

    I feel compelled to come to the defense of the interviewee here. I think that the interviewer really misses the point of what Allan Kelly is talking about regarding the learning that takes place on a software development project.

    In software development the learning that Mr Kelly is referring to is the learning of the domain and the technology stack and the design patterns that are constantly emerging as the product is being built. Even with an inordinate amount of up front design and requirements gathering on a project there is still constant learning that takes place as the solution emerges.

    The interviewer seems to be focused only on the the learning of cool new toys and API’s to program to, those ‘fun’ learning activities that are not necessarily related to the delivery of the project that the project manager is responsible for. While staying current in the latest technology is important for developers, this is not the learning that Mr. Kelly is talking about.

    We must approach each software development project as a learning exercise. That’s why we learn to build software in a fashion that can be easily refactored. Without this learning we end up with mis-understood requirements, badly designed domain models and a lack of properly implemented design patterns.

  14. anmuelle says:

    I really like SE-Radio a lot, but honestly this was the worst episode ever!

    And in my opinion not because of the topic, content or because Allan Kelly kept on an too abstract level. In contrast it could have become much more concrete, if the interviewer would have allowed that to happen…

    To me it appeared like Arno was not prepared regarding the topic at all and he just re-iterated the only one question that came into his mind over and over again (kind of because is seemed to be the only thing that he could think of and was able to ask).

    Anyways it’s definitely no good interview style to try to convince the guest of the own opinion, and just insist on repeating a question until one gets the “correct” answer or the interviewee just does not know what to say anymore.

    Several times a was really tempted to just switch off, because Arno rephrases the exact same question once again. Unfortunately poor Allan could not escape this “pseudo-inquisition” just like that – even though he sounded like he’d loved to, which is quite understandable…

  15. Tarek says:

    The subject is indeed very interesting but I was expected concrete tips & advises for acquiring knowledge during the development process, basically pragmatic tips & advises for learning on the field …

    This is a very interesting subjects but the interview went totally wrong,
    I’m sure that Allan has got serious credentials but as a Software Engineer I was hoping for practical stuffs …

    I suggest you redo a podcast on the same subject

Leave a Reply

Your email address will not be published. Required fields are marked *


*