<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by John Cinnamond on Medium]]></title>
        <description><![CDATA[Stories by John Cinnamond on Medium]]></description>
        <link>https://medium.com/@jcinnamond?source=rss-5c3b5f793ccb------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*2OlofKCpfgPQQhL_0zoc8g.jpeg</url>
            <title>Stories by John Cinnamond on Medium</title>
            <link>https://medium.com/@jcinnamond?source=rss-5c3b5f793ccb------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 21:56:55 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@jcinnamond/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[How can we ensure that senior developers have the skills to mentor juniors?]]></title>
            <link>https://medium.com/@jcinnamond/how-can-we-ensure-that-senior-developers-have-the-skills-to-mentor-juniors-e42aabfd97be?source=rss-5c3b5f793ccb------2</link>
            <guid isPermaLink="false">https://medium.com/p/e42aabfd97be</guid>
            <category><![CDATA[programming]]></category>
            <dc:creator><![CDATA[John Cinnamond]]></dc:creator>
            <pubDate>Sat, 06 Jul 2019 11:55:53 GMT</pubDate>
            <atom:updated>2019-07-06T11:55:53.051Z</atom:updated>
            <content:encoded><![CDATA[<p>This is a question that came up during <a href="https://brightonruby.com/">Brighton Ruby</a> yesterday, during a chat show style presentation called “<a href="https://twitter.com/nodunayo">Nadia</a> chats with… <a href="https://twitter.com/sarahmei">Sarah Mei</a>”. It’s a question that got me thinking and I want to explore it a little more and suggest an answer to it. (Spoiler alert: it’s the culture, stupid.)</p><p>I think that the underlying premise of this question is sound. Mentoring is a good thing, and it is important to invest in the development of more junior members of the team. Being a software developer is about more than just writing code and healthy software teams are full of people who have both technical and people skills. So it seems reasonable to ask how to get better at it.</p><p>And yet… the question bothers me, especially in the way that I have formulated it*. I think it mischaracterises the problem and I want to address it in two ways: first, by making an appeal against a latent individualism in the question; and second, by rejecting an underlying imbalance of power within teams.</p><p>Let’s start by discarding the problematic terms ‘senior developer’ and ‘junior developer’. For me, these are terms without a clear definition of what constitutes being “senior” or “junior”. Instead, we can talk about people with more experience of development helping those with less experience. This is principally through sharing knowledge and experience; through providing encouragement; and through challenging people to think differently about a problem. But why would these practices be unidirectional, only ever flowing from those with more years of experience to those with less? Of course by definition, those with more experience have more experience to share, but this doesn’t grant them an exclusivity on it. And when it comes to the question of encouragement or of challenging people to think differently, the amount of prior experience becomes even less relevant. People who are new to development would be just as capable of playing this role.</p><p>This also assumes that the experience to be shared is predominantly technical. But I previously suggested that healthy technical teams are full of people with both technical and people skills. Is it reasonable to assume that people with more years of development experience are also the ones with advanced people skills? My own experience suggests otherwise, and indeed the very premise of the introductory question suggests that it is something that many “experienced” developers are lacking. So perhaps we should be asking why those with strong people skills are not mentoring those without, irrespective of the number of years they have spent being paid to write code.</p><p>Mentoring is not teaching. It doesn’t require a teacher-student relationship. And thus it shouldn’t be thought about in terms of a small group of omniscient figures handing down their wisdom to the rest of the team. It should be a constant activity that the entire team deliver to each other and, importantly, <em>receive</em> from each other. It’s a group activity, not a task for an individual to perform.</p><p>This introduces some challenges to the traditional power structures within teams, and to another problem I have with the titles “senior developer” and “junior developer”. I’m not going to deny that there will always be an imbalance of power within a team, largely expressed through the amount of influence individuals hold over the rest of the team. This distribution of power is typically based on the number of years of development experience an individual possesses, or on their acuity in a particular area. I do not wish to argue that having an imbalance of power within a team is unreasonable. I do, however, want to talk about the “expression” of power, in particular on the way in which power can be exercised to the maximum good of the whole team.</p><p>There is a tendency to assume that the primary expression of power ought to be in the transfer of skills and knowledge from those with more experience to those with less. This, I think, is a mistake. The optimal expression of power within a team — i.e., the way in which influence is distributed within a team — ought to be based on the extent to which an individual can make the whole team better. It is not a measure of how much knowledge they already hold. Often, individuals make teams better by passing on knowledge and to this extent the traditional power structures within a team make sense. But not all knowledge is technical knowledge. More importantly, there are many other ways to help make a team better. The person who notices their colleague is getting stressed, and who helps them find a way through it, is at least as important as someone with a wealth of technical skill. The power structures within a team ought to reflect this, and often don’t.</p><p>So if the question is “how can we ensure that senior developers have the skills to mentor juniors?” then I think that the answer is “you don’t.” Instead you create a culture where everybody helps everyone else in the team to develop both technical and non technical skills, and you distribute power based on the extent to which individuals make the whole team better, rather than on the number of years experience they have or on their technical prowess. If you manage this, then the question becomes meaningless, and your team becomes all the better for it. Fix the culture first. The individuals will follow.</p><p><em>(*) For the avoidance of doubt, the leading question is my interpretation of something that came up during the talk, taken out of context and rephrased. To the extent that I am critical of the way this question is asked, it is criticism of my interpretation, not of other people’s statements.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e42aabfd97be" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Some throw away thoughts on programming as mathematics]]></title>
            <link>https://medium.com/@jcinnamond/some-throw-away-thoughts-on-programming-as-mathematics-5245e49c5a48?source=rss-5c3b5f793ccb------2</link>
            <guid isPermaLink="false">https://medium.com/p/5245e49c5a48</guid>
            <category><![CDATA[programming]]></category>
            <dc:creator><![CDATA[John Cinnamond]]></dc:creator>
            <pubDate>Sun, 08 Apr 2018 10:33:42 GMT</pubDate>
            <atom:updated>2018-04-08T10:33:42.070Z</atom:updated>
            <content:encoded><![CDATA[<h3>Paul Snively on Twitter</h3><p>Real talk: &quot;Programming isn&#39;t mathematics.&quot; Yes, it really, truly, absolutely is, &amp;amp; fighting the fact will only lead to needless pain.</p><p>I have some reckons!</p><p>First up, I’m pretty sure that I agree with the sentiment of this tweet and I’m only interested in exploring how to understand it, not to argue that it is wrong. But I also think that there can be a mismatch between people’s experience of programming and the idea that programming is mathematics, and it’s important to understand this and look at how (computational) mathematics can fit into their world.</p><p>I think that there are two different interpretations of programming. The first is the act of writing code that gets translated into a series of instructions for a computer. I’m going to refer to this as coding. There is also the sum of activities that go into producing working programs, of which coding is only one activity. Programming, as defined as this sum of activities, and coding are not equivalent (although they might be isomorphic :)</p><p>When people say “programming is mathematics” I think that this often easily dismissed because it doesn’t apply to the whole act of producing working programs, and perhaps not even to the most difficult part — that of understanding what needs to be done.</p><p>There is also a problem with the definition of “mathematics”, as several replies to the tweet noted. Most people don’t recognise the nature of mathematics beyond what they have been taught in school. And this is a real shame — mathematics is not (only) a set of equations and definitions and rules; it is the best tool we have to make sense of the world we live in. It is also the best tool we have to create useful approximations of the world and this, in essence, is what we are doing when we producing working programs.</p><p>I find that idea that thinking that programming is not mathematics will ‘lead to needless pain’ a bit more problematic, even though I think that it is essentially true. There’s an implication in there that if only we understood this then coding would be pain free and we could live as happy programmers in a land of rainbows and unicorns. Now I realise that this is not what the tweet says, but I think that this is often how it is read. There is also an implication that (the majority of) programmers are the ones making programming difficult and painful through their own intellectual failings. Now again, that isn’t what the tweet says, but it could be read as implying this.</p><p>I think it’s much more useful to talk about mathematics and computation in a positive light. Rather than telling people that they’re creating pain by not being interested in mathematics, it is better to give people a glimpse of how a mathematical understanding of computation can solve a real problem that they face. Better to get people excited about the possibility of mathematics than to beat them over the head with the rod of intellectual elitism.</p><p>I’ve diverged from the original tweet now, onto the wider frustration of why programmers don’t habitually study mathematics, given that it solves a lot of the problems they face. Ultimately I think it’s a sales problem. Programming is an activity done by humans for other humans, so if you want more programmers to get enthusiastic about mathematics then you need to sell it to them. The problem with mathematics and programming is, ultimately, one of psychology.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5245e49c5a48" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[London Ruby Unconference, 2016]]></title>
            <link>https://medium.com/@jcinnamond/london-ruby-unconference-2016-2dc4b5ac8fb0?source=rss-5c3b5f793ccb------2</link>
            <guid isPermaLink="false">https://medium.com/p/2dc4b5ac8fb0</guid>
            <category><![CDATA[unconference]]></category>
            <category><![CDATA[lrun]]></category>
            <category><![CDATA[conference]]></category>
            <dc:creator><![CDATA[John Cinnamond]]></dc:creator>
            <pubDate>Tue, 25 Oct 2016 13:05:19 GMT</pubDate>
            <atom:updated>2016-10-25T13:05:19.365Z</atom:updated>
            <content:encoded><![CDATA[<p>I recently attended the inaugural <a href="https://github.com/codescrum/london_ruby_unconference_2016/wiki">London Ruby Unconference</a> and I thought that it would be good to jot down my thoughts while they’re still fresh in my mind. This was the first unconference I has attended and I wasn’t entirely sure what to expect, so I was pleased to find it both enjoyable and useful.</p><p>The day followed what I believe to be a standard unconference format. At the beginning of the day everyone gathered around a whiteboard containing a grid of 5 time slots, each slot containing up to 5 simultaneous sessions. Attendees were invited to propose a session and select a time and place to lead it. There was no discussion around the proposal, no criteria for eligibility, and no voting. If you wanted to have a session on any topic then you simply proposed it. This lack of selection felt really good — it was an opportunity to talk about whatever topic you were interested in. However, getting up in front of all of the attendees to propose something was a little daunting, particularly without any preparation. Thus session leaders were naturally self-selecting to people with the confidence to do this. I wonder if there would be a way to encourage people who are less confident to propose sessions…</p><p>(A quick diversion on diversity: I wanted to talk about the feeling of ‘inclusivity’ of the sessions, but ‘inclusion’ is a word with many connotations. I don’t know how many attendees there were, but I would guess at around 40. Of these, probably 35 of them were white men. This is pretty representative of the breakdown of programmers that I work with. There was nothing about the organisation of the conference that prevented people from other backgrounds from taking part, but inclusion is about more than just removing barriers. It is also about encouraging people from underrepresented communities to take part, and to feel as though they belong. I mention this in particular because in general the unconference was a great opportunity to learn from people I wouldn’t otherwise have spoken to, and it would have been more valuable overall had those people come from a wider variety of backgrounds and faced different challenges in the workplace compared to people like me.)</p><p>The first session I attended was a discussion on ‘Why Ruby?’, proposed by <a href="https://twitter.com/jenny_huz">Jenny Huzell</a>. It started from the point of view of both a recruiter and a prospective programmer asking why we choose to use Ruby and not some other language. This then evolved into discussions about the importance of the ruby community, comparisons with other languages, and observations about the easy of getting started with Ruby and, in particular, Rails.</p><p>Next up, I lead a session about Truth and Abstraction. This was different from most sessions I attended in that it mostly involved me talking for an hour and very little group discussion. I guess that even my laissez-faire approach to an unconf session was still dominated by my control freakery and my tendency to talk to much. What was interesting to me, though, was that this was an opportunity to explore a new and unfinished idea by discussing it, rather than having the pressure of trying to turn it into a more polished topic for a normal conference talk. Some of the questions and observations that people made during the session gave me some new ideas as to how this topic might evolve, and that probably wouldn’t have happened in a more formal setting.</p><p>After lunch I attended two group discussions: one on architecture lead by <a href="https://twitter.com/thebestie">Stephen Best</a>; and one on changes to rails 5 controller testing lead by <a href="https://twitter.com/hlame">Murray Steele</a>. This was unlike anything I’ve encountered at a traditional conference. You do get discussions around topics such as these in hallways and between sessions, but here they were thrust centre stage and given the time and space to unfold. More importantly, it wasn’t just discussions between people I already knew but it involved everyone else at the conference who was interested in that topic. Well… except that these discussions tended to be dominated by a few people. (Or, to put that another way, I apologise for talking too much in these discussions.) It would have been interesting to see how the same topics would have evolved in a more structured format, such as a <a href="https://en.wikipedia.org/wiki/Fishbowl_(conversation)">fishbowl conversation</a>. But overall these were the most interesting parts of the unconference for me, and the sessions I got most out of. Although a lot of the people at the unconference are also people who regularly attend <a href="http://lrug.org/">LRUG</a>, it was unusual to discuss, say, how to test controllers with so many other developers.</p><p>The final session was an overview of Crystal, lead by <a href="https://twitter.com/feyeleanor">Eleanor McHugh</a>. This walked through a series of increasingly esoteric ways of writing loops in Crystal and then comparing this to Ruby and, because it was Eleanor, to Go. This was kinda fun and illustrated both the similarities and the differences between Ruby and Crystal. I was a bit disappointed that the session didn’t contain more real world experience of the pros and cons of Crystal, particularly replacing existing Ruby solutions with ones written in Crystal, but that just wasn’t what the session was about. It was, however, enlightening to see just how much existing Ruby code would work without modification.</p><p>There were a few things I think could be tweaked about the unconference format to make it even better. It would have been nice to have short breaks between sessions, and to have had a longer break for lunch. Even though the unconference format is inspired by ‘hallway tracks’ — the name given to those spontaneous conversations that take place outside of the scheduled talks at traditional conferences, sometimes in lieu of attending a talk — there is still room for a hallway track at an unconf. Conversations spill out of the discussions, and smaller groups splinter off to discuss topics that began in the sessions. So a day with four sessions and more breaks would have been ideal for me. I also think that the lack of structure had both pros and cons. It gave everybody the opportunity to take part, but did nothing to ensure that everyone was given encouragement to do so. As a result I think it was slightly dominated by those with bigger personalities, or with more confidence speaking in front of groups of peers. Finally, as I alluded to earlier, more diversity always makes for a greater opportunity to learn from others.</p><p>Niggles aside, it was a good day, and I’m glad it happened. I wouldn’t like to see traditional conference or meetup formats replaced with unconferences, but I think that there is real value in having both. Many thanks to <a href="https://twitter.com/codescrum">Jairo Diaz</a> for organising the unconf, to <a href="https://www.carwow.co.uk/">carwow</a> for hosting us, and to all the volunteers for making the day run smoothly. It’s certainly left me with plenty to think about and that, at the end of the day, is what (un)conferences are all about.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2dc4b5ac8fb0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[On the fact that there are more oceans than just the Atlantic]]></title>
            <link>https://medium.com/@jcinnamond/on-the-fact-that-there-are-more-oceans-than-just-the-atlantic-726926b1cafe?source=rss-5c3b5f793ccb------2</link>
            <guid isPermaLink="false">https://medium.com/p/726926b1cafe</guid>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[formalism]]></category>
            <category><![CDATA[ruby]]></category>
            <dc:creator><![CDATA[John Cinnamond]]></dc:creator>
            <pubDate>Fri, 24 Jul 2015 16:49:28 GMT</pubDate>
            <atom:updated>2015-07-24T16:49:28.009Z</atom:updated>
            <cc:license>https://creativecommons.org/licenses/by-sa/4.0/</cc:license>
            <content:encoded><![CDATA[<p>In the 1980s there was something of a spat between two of Computing/Computer Science’s most influential figures. In one corner was the legendary Edsgar W. Dijkstra, decrying the lack of mathematical rigour of programmers in the United States. In the other was the legendary Alan Kay, rebutting Dijkstra with the observation that a lack of mathematical rigour was something of a strength when it comes to programming. The two opposing sides of this argument can be summed up in Dijkstra’s “<a href="https://www.cs.utexas.edu/users/EWD/transcriptions/EWD06xx/EWD611.html">On the fact that the Atlantic Ocean has two sides</a>” and in early remarks of Alan Kay’s OOPSA ’97 keynote: “<a href="https://www.youtube.com/watch?v=oKg1hTOQXoY">The computer revolution hasn’t happened yet.</a>”</p><p>This colourful disagreement between two of computing’s greats has largely faded from vision, but the essential disagreement has never really gone away. It surfaces from time to time and so it was that, true to form, I watched an American (Avdi Grimm) come to Blighty to tell the attendees at <a href="http://brightonruby.com/">Brighton Ruby</a> that a rejection of <em>formalism</em> is one of the strengths of the Ruby community, and one that could become its legacy. You can watch <a href="http://devblog.avdi.org/2015/05/21/the-soul-of-software/">a recording of an earlier version of this talk</a> given at <a href="http://tropicalrb.com/en/">Tropical Ruby</a>.</p><p>There were a number of themes that Avdi touched on in his talk but the one that I am most interested in is the rejection of formalism and, in its place, the promotion of a humanistic, communicative approach to software development. My interest in this theme is largely driven by the fact that I think it is not only essentially wrong, but is also a regressive attitude to hold about software development.</p><p>I struggle to fully understand what Avdi means by <em>formalism</em> in this talk, but I think he defines it as a style of software development that is driven by adherence to objectively defined rules, rather than one that reacts to the situational reality in which the software is being developed. There is more to it than this, however, because there are also implications that formalism is the world of proofs and mathematics, whereas hermeneutics — the name that Avdi uses for the alternative non-formalist approach — would sacrifice mathematical certainty for better communication, and for solutions that were better suited to the world in which they were developed.</p><p>In general I felt unsure exactly where the divide lay between formalism and hermeneutics, and I suspect that an exhaustive definition would be hard to give. There are definitely examples that I think everyone could agree were formalist — for example writing formal proofs of some property of a piece of code. Likewise there are definitely uncontroversial examples of non-formalism — for example emphasising the importance of gathering feedback from customers throughout the development process. However there are also a number of examples that are much less clear. Is the practice of Test Driven Development formalist? If I discover important properties about the problem from a customer but then capture them in <a href="http://research.microsoft.com/en-us/um/people/lamport/tla/tla-intro.html">TLA</a>+ does this make it formalist? Is formalism a question of tools? Is it a belief about the value of mathematic proofs? Is it about empathy with users? Or is it perhaps a shifting combination of some, or all, of these?</p><p>I found myself thinking about the question of what formalism <em>is</em> during the talk as some of the examples seemed to suggest that formalism is anything that is bad in software development, whereas hermeneutics is anything good. For example, the description of Inheritance being a formalist concern didn’t sit well with me. Most people I would consider formalist would disapprove of Object Oriented programming in general and I can’t see a reason why they would be in favour Inheritance in particular. I think that the reason that Inheritance is taught so early and so emphatically in, say, university courses is because Object Oriented programming is badly understood and badly taught, not because of a belief in the value of mathematical proofs, or formal logic, or axiomatic set theory, or whatever definition you want to give to formalism. Indeed the ultimate expression of formalism in Object Oriented programming is, in my opinion, something like an <a href="http://lucacardelli.name/Papers/PrimObjImpSIPL.A4.pdf">Object Calculus</a> and that doesn’t seem unduly concerned with Inheritance, as far as I can understand it.</p><p>There are good reasons to be suspicious of the formalists. If you define formalism in computing to mean a belief in the ability of mathematical logic to fully describe real world systems, then the greatest reason to be suspicious of it is that it doesn’t work. This is most evident in the analogous failure of <a href="https://en.wikipedia.org/wiki/Logical_positivism">Logical Positivism</a> in 20th Century philosophy.</p><p>There is, I think, a real danger that people continually miss the lessons in Logical Positivism’s early success and subsequent failure. For example, casual proponents of type theory often take success in solving certain kinds of problems and extrapolate it to a kind of Universal Usefulness of Types, failing to recognise that there are a number of problems for which type theory is inappropriate and we would be much better served if types would just go away and leave us all alone for a while. Conversely, there are those who recognise the failure of type theory for some kinds of problems and extrapolate it to a kind of Universal Uselessness of Types, failing to recognise that there are real benefits that types can bring us when we’re reasoning about code.</p><p>I think that both camps are equally damaging to the challenge of trying to get better at writing software. The problem here is nothing to do with the theory in question — the real problem is inappropriate reductionism to a set of rules that are mistakenly supposed to be universally appropriate.</p><p>In the context of Avdi’s talk the definition of formalism, along with an understanding of its strengths and weaknesses, matters. It matters because persuading people to reject formalism is, by extension, persuading people to reject large swathes of Computing Science. This is not a step I think we should be taking.</p><p>In general, we are not sufficiently good at developing software. We are not sufficiently good at capturing specifications, we are not sufficiently good at writing reliable code, and we are not sufficiently good at maintaining systems. We need all the help we can get to improve. Computing Science has many tools that can help us improve today, and it will continue to produce better tools in the future. If we hope to improve as software developers we need to embrace this and learn as much as we can from what Computing Science has to teach us.</p><p>But we also need to learn where the limits of mathematical logic lie. We need to understand that the world in which we operate is not cleanly delineated into propositions and proofs. We need to understand that expecting systems to be precisely <a href="https://en.wikipedia.org/wiki/Z_notation">specified in terms of first-order predicate logic</a> can overpower the essentially human concerns driving the development in the first place. We need to understand that a buggy system doing a useful thing is often better than a bug free system doing something that nobody wants.</p><p>This understanding, however, shouldn’t become another simple reductionist view of the world of software development. Software is unusual in that success in writing it requires an understand of both the formal world of mathematical logic, and the human context in which programs are being written. To present it as an either/or choice — as I felt Avdi’s talk did — is a false dichotomy: one which I think is pervasive and one which I think holds back our industry.</p><p>As a community based around a programming language developed in Japan we can well understand that the world is much bigger than two sides of the Atlantic. I hope that we can equally understand that the programming world is much bigger and more complicated than warring groups of formalists and hermenauts, one to be rejected and the other to be embraced. Each group has something important to tell us. I think that the best outcome from all of this is not to pick sides, but rather to develop a new willingness to listen, to critically assess, and to use whatever we might hear to become better programmers, no matter what label we might apply to those who say it. But maybe that’s just the formalist in me talking.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=726926b1cafe" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>