<?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 Michael Rose on Medium]]></title>
        <description><![CDATA[Stories by Michael Rose on Medium]]></description>
        <link>https://medium.com/@_rose_m_?source=rss-8842af9faec7------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*00doR6j9jjZFpOTXEY8ePQ.jpeg</url>
            <title>Stories by Michael Rose on Medium</title>
            <link>https://medium.com/@_rose_m_?source=rss-8842af9faec7------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 05:54:36 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@_rose_m_/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[README]]></title>
            <link>https://medium.com/@_rose_m_/readme-7297d9c1db62?source=rss-8842af9faec7------2</link>
            <guid isPermaLink="false">https://medium.com/p/7297d9c1db62</guid>
            <category><![CDATA[life-lessons]]></category>
            <category><![CDATA[tech-culture]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[management]]></category>
            <dc:creator><![CDATA[Michael Rose]]></dc:creator>
            <pubDate>Tue, 20 Nov 2018 10:35:33 GMT</pubDate>
            <atom:updated>2018-12-04T07:21:13.966Z</atom:updated>
            <content:encoded><![CDATA[<p>Inspired by some of the (maybe) well-known <a href="https://hackernoon.com/12-manager-readmes-from-silicon-valleys-top-tech-companies-26588a660afe">manager READMEs</a>, I have chosen this form to give a brief account of who I am and what you should know if you are going to work with me.</p><h3>About Me</h3><ul><li>Right now I work at collaboration Factory AG where I do software development, recruiting, and other organizational tasks</li><li>I am a nerd at heart, interested in all kinds of technology, but especially programming, and I do like programming <em>a lot</em></li><li>I like to share a laugh — life is serious enough by itself so I like to lighten it up when possible, and I definitely can laugh at myself, too</li><li>In my free time I love doing sports — especially a nice, hard circuit training session helps to keep me balanced</li></ul><h3>Principles &amp; Values</h3><p>Those are things that are very important to me and influence what I do — which doesn’t mean I expect everyone else to be an exact match in every respect, but I do my work based on them.</p><ul><li><strong>Loyalty and having each others back</strong>: if things go wrong or help is required I’m not just going to quit or leave easily — I’ll be there and try my best to solve the issues. On the other hand, I might come and ask for the same one day. <em>Hint: this is not unlimited — if goes too far you’ll lose me, but I’ll warn you before.</em></li><li><strong>Honesty &amp; candor</strong>: I’m not going to lie to anyone, and expect it the other way around. I do not like being told what I presumably may want to hear when isn’t the full truth, especially if someone else has received a different answer to the same question. I am open to new and different approaches, solutions, or ideas that I haven’t thought of — and I ask you to think so, too. “We’ve always been doing it this way” is not an argument I accept.</li><li><strong>I’m hands-on and judge actions not titles</strong>: I like a hands-on approach after giving proper thought on how to proceed — but I’m not a fan of drowning in thought. Just do it, give it a try, then gather feedback and experience and evaluate that. That also translates to my view on others: I am not impressed by fancy titles or seniority of age but by how you interact with me or others and what you <em>do</em>.</li><li><strong>I’m ambitious and don’t follow orders blindly</strong>: I want to push things and make an impact. I’ll also happily take ownership. Therefore I need to know the bigger picture but I’m bad at just following orders. A classical coercive power hierarchy definitely doesn’t suit me.</li></ul><h3>Feedback</h3><p>Giving and receiving feedback is very important to me. It is the only way to know whether my behavior and actions are correct or can be improved.</p><ul><li><strong>I really like getting feedback</strong>: share it with me whenever you want and whoever you are. The closer you are to me the more important your feedback will be.</li><li><strong>Give open, honest feedback</strong>: don’t play politics or try to say what I want to hear. Give criticism if required — but be prepared for follow-up since I’ll probably ask some questions to know the details of what I did wrong.</li><li><strong>Don’t wait for a special occasion</strong>: if you have feedback to give just hit me up and I’ll be able to spare some time.</li><li><strong>If you want feedback from me just ask</strong>: I’ll try to give feedback proactively to those I feel close to but I won’t volunteer it at the company. You are always welcome to ask.</li></ul><h3>Expectations at Work</h3><p>Apart from everything outlined above, I do have certain expectations in a working environment that I want to point out.</p><ul><li><strong>Awesome team — great talent, great people</strong>: I want to work in a team of very talented people who inspire me to try new things, learn something new, and share their experience with me. Great people also means a friendly, familiar atmosphere where I can feel at home — no fights, no elbow-mentality.</li><li><strong>Draw a big, ambitious picture and let me focus</strong>: Since it’s important to me to make an impact I’d like to know the big picture — where are we going, what is the strategy. I want to be provided with ambitious goals, not just the ordinary; and to achieve them I need to focus on only a small number of things at a time. Don’t make me change focus every week — and be consistent in the big picture you share.</li><li><strong>Challenge me and be ready to be challenged</strong>: I’m bad at doing the same thing in the same context for a long period of time. Give me new challenges from time to time — from technical to organizational to strategic. I will help everywhere I can but what really drives me is working on and solving complex challenges. Since I’m bad at just following orders, I’ll also challenge you to show me why things should be done.</li><li><strong>I do have an opinion and want it to be valued</strong>: I definitely don’t know everything better than anyone else, but I do have an opinion on a lot of things. When I don’t have any solid reasoning to support my views I won’t share them; but when I see something isn’t right I’ll speak up, ask questions, and say what I think. I expect to be heard and want it to be valued; and I will always accept a plausible explanation why my view is wrong.</li></ul><p><em>Originally published at </em><a href="https://www.coderose.io/readme"><em>www.coderose.io</em></a><em> on November 20, 2018.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7297d9c1db62" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[WiFi AP with Raspberry Pi 3]]></title>
            <link>https://medium.com/@_rose_m_/wifi-ap-with-raspberry-pi-3-c6a2bfad62f4?source=rss-8842af9faec7------2</link>
            <guid isPermaLink="false">https://medium.com/p/c6a2bfad62f4</guid>
            <category><![CDATA[raspberry-pi]]></category>
            <category><![CDATA[security]]></category>
            <dc:creator><![CDATA[Michael Rose]]></dc:creator>
            <pubDate>Mon, 29 Oct 2018 20:08:12 GMT</pubDate>
            <atom:updated>2018-11-22T08:39:34.983Z</atom:updated>
            <content:encoded><![CDATA[<p>In order to have some rogue WiFi networks for playing around with security tools and techniques at home, I wanted to set up my Raspberry Pi 3 as an Access Point. Turned out, that trying this with two WiFi adapters — the builtin one plus a USB adapter — can be a huge pain…</p><h3>Overview</h3><p>To give a quick overview of what scenario I want to create check out the picture below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/601/0*j06WEa57dW3ZSULe.png" /></figure><p>In the end we should be able to spawn a WiFi for our testing purposes as shown here being the BadWiFi. The main machine will be able to connect to the Pi itself via SSH over the regular WiFi. The attacker will just connect/sniff the BadWiFi.</p><h3>Hardware Setup</h3><p>Since I have an additional WiFi adapter the Pi registers two different interfaces:</p><ul><li>wlan0 is the Pi&#39;s builtin WiFi interface</li><li>wlan1 is the USB adapter</li></ul><p>For the purpose of creating my own AP I will use wlan0 as the AP interface and wlan1 remains connected to my regular home WiFi.</p><p>If you want to set this up yourself, too — I had <strong>no luck</strong> using the interfaces the other way round (wlan1 as AP). I wasted around 12 hours trying, so just don&#39;t (or you tell me why I failed miserably). The issue I had with using wlan1 as AP was that all clients were always disassociating from the WiFi and could not establish a stable connection.</p><h3>Preparing Configuration</h3><p>Usually the wpa_supplicant will try to control all our WiFi interfaces on the Pi. In my case it would always make both of them immediately join the configured home WiFi. One neat thing I found after lots of googling was that it is possible to separate wpa_supplicant configurations by interface.</p><p>First, let’s use the main configuration as base for both individual configs.</p><pre>$ cd /etc/wpa_supplicant<br>$ sudo cp wpa_supplicant.conf wpa_supplicant-wlan0.conf<br>$ sudo cp wpa_supplicant.conf wpa_supplicant-wlan1.conf</pre><p>Since we will be using wlan1 as our <em>normal</em> WiFi interface, all the network configuration blocks should go into wpa_supplicant-wlan1.conf. Make sure both the regular wpa_supplicant.conf <strong>and</strong> wpa_supplicant-wlan0.conf <strong>do not</strong> contain any network blocks. In my case I only had a single line with a country setting in the /etc/wpa_supplicant file.</p><p>Verify that everything is still working and you can connect to the Pi using ssh by rebooting it. You should see when running iwconfig that only wlan1 is connected to the regular WiFi.</p><h3>Software Setup</h3><p>To be able to create the AP via software on the Pi I used hostapd. In order to provide the DHCP server required to automatically hand out IP addresses I relied on dnsmasq. To install both packages on the Pi just do:</p><pre>$ sudo apt-get install hostapd dnsmasq</pre><p>After installation we just make sure to have both services stopped so that we can play around with the configuration files:</p><pre>$ sudo systemctl stop hostapd<br>$ sudo systemctl stop dnsmasq</pre><p>Since I don’t need the AP all the time and only activate it when required, I decided to completely disable both services to not have them started on boot:</p><pre>$ sudo systemctl disable hostapd<br>$ sudo systemctl disable dnsmasq</pre><h3>Configuring DHCP</h3><p>For DHCP to be fully operational both the dhcpcd as well as dnsmasq have to be properly configured. The former needs to be used to assign a fixed IP to the wlan0 interface which will then hand out DHCP leases to the WiFi via dnsmasq.</p><p>I decided to give the BadWiFi the 192.168.1.1/24 IP subnet since my regular WiFi has 192.168.178.1/24. As such wlan0 should get the address 192.168.1.1.</p><p>Just edit /etc/dhcpcd.conf and at the end of the file add the following (check there is no other interface wlan0 block):</p><pre>interface wlan0 static ip_address=192.168.1.1/24</pre><p>The original configuration of dnsmasq that is provided after installation at /etc/dnsmasq.conf contains a huge load of possible configuration options and comments. It&#39;s best to just start with a fresh one and keep the old as backup:</p><pre>$ sudo cp /etc/dnsmasq.conf /etc/dnsmasq.conf.orig<br>$ vim /etc/dnsmasq.conf</pre><p>The required content for DHCP to work in our AP is:</p><pre>interface=wlan0<br>no-dhcp-interface=wlan1<br>dhcp-range=192.168.1.100,192.168.1.200,24h<br>dhcp-option=option:dns-server,192.168.1.1</pre><p>We set to work on wlan0 but exclude wlan1 (my regular WiFi also does DHCP and I don&#39;t want any conflicts there). The DHCP range itself is limited to the addresses 100 to 200 and wlan0 will also be the DNS server for all clients.</p><p>After all this configuration once again make sure to reboot your Pi and verify that everything else is still working as expected…</p><h3>Configuring the AP</h3><p>Finally — it’s time to configure hostapd and get the software AP up and running. We will place its configuration file at /etc/hostapd/hostapd.conf:</p><pre>$ sudo vim /etc/hostapd/hostapd.conf</pre><p>The required content to create a simple WPA2 network is as follows:</p><pre>interface=wlan0<br># country_code=DE - adapt if needed</pre><pre>ssid=BadWifi<br>channel=9<br>auth_algs=1</pre><pre>wpa=2<br>wpa_passphrase=YourSecret!<br>wpa_key_mgmt=WPA-PSK<br>wpa_pairwise=TKIP CCMP<br>wpa_group_rekey=86400</pre><p>The options should be self explanatory. I only added wpa_group_rekey=86400 to make sure the rekeying takes place only once per day.</p><p>A great resource to find out all those settings and see what is possible has been <a href="https://w1.fi/cgit/hostap/plain/hostapd/hostapd.conf">that dummy file</a>.</p><p>The last thing to change to get hostapd to pick the right config is to edit /etc/default/hostapd and make sure the following line is in there:</p><pre>DAEMON_CONF=&quot;/etc/hostapd/hostapd.conf&quot;</pre><p>Finally for IP forwarding to work and be able to share the WiFi connection upstream we need to change one setting in /etc/sysctl.conf by uncommenting the following line:</p><pre>net.ipv4.ip_forward=1</pre><p>Just reboot a last time without enabling any services and make sure everything else works as expected.</p><h3>Starting the AP</h3><p>After all the changes and configuration hurdles we can now try and start the Access Point at last:</p><pre>$ sudo systemctl start dnsmasq<br>$ sudo systemctl start hostapd</pre><p>A minor little thing I found when starting hostapd is that sometimes it doesn&#39;t have enough entropy to be able to start the WPA encryption required for all handshakes. The easiest workaround: just run find * / for a few seconds at that will be solved.</p><p>And after a few seconds you should finally be able to see the new BadWiFi and join it with the provided password from above!</p><h3>Conclusion</h3><p>Despite following the <a href="https://www.raspberrypi.org/documentation/configuration/wireless/access-point.md">official documentation</a> I could not get the AP up and running. It really took me some hard hours of research, pain, and failure to finally succeed. I hope my description above does help you out when trying to get it running yourself. BUT IT WORKS!!</p><p><strong>Resources:</strong></p><ul><li><a href="https://gist.github.com/kimus/9315140">https://gist.github.com/kimus/9315140</a></li><li><a href="https://wiki.ubuntuusers.de/Dnsmasq/#Konfiguration-als-DHCP-Server">https://wiki.ubuntuusers.de/Dnsmasq/#Konfiguration-als-DHCP-Server</a></li><li><a href="https://www.elektronik-kompendium.de/sites/raspberry-pi/2202031.htm">https://www.elektronik-kompendium.de/sites/raspberry-pi/2202031.htm</a></li><li><a href="https://forum.turris.cz/t/solution-found-for-wifi-clients-disconnects/6065">https://forum.turris.cz/t/solution-found-for-wifi-clients-disconnects/6065</a></li><li><a href="https://www.raspberrypi.org/documentation/configuration/wireless/access-point.md">https://www.raspberrypi.org/documentation/configuration/wireless/access-point.md</a></li><li><a href="https://w1.fi/cgit/hostap/plain/hostapd/hostapd.conf">https://w1.fi/cgit/hostap/plain/hostapd/hostapd.conf</a></li></ul><p><em>Originally published at </em><a href="https://www.coderose.io/wifi-ap-with-raspberry-pi-3/"><em>www.coderose.io</em></a><em> on October 29, 2018.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c6a2bfad62f4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[JS Kongress 2017 Roundup]]></title>
            <link>https://medium.com/@_rose_m_/js-kongress-2017-roundup-b37f4654be05?source=rss-8842af9faec7------2</link>
            <guid isPermaLink="false">https://medium.com/p/b37f4654be05</guid>
            <category><![CDATA[conference]]></category>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[js-kongress]]></category>
            <dc:creator><![CDATA[Michael Rose]]></dc:creator>
            <pubDate>Thu, 23 Nov 2017 19:40:39 GMT</pubDate>
            <atom:updated>2017-11-28T13:40:38.181Z</atom:updated>
            <content:encoded><![CDATA[<p>A little while ago on the 13th and 14th of November I attended the <a href="https://2017.js-kongress.de/">JS Kongress</a> themed <em>The Future of JavaScript</em> here in Munich. It has been my first time being to this conference — and up forefront: it has been a great and very pleasant experience. Thanks a lot to all the organizers for making this possible! But let’s get into some more details…</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sfc6xTVJQLcywyfXIt7k0Q.jpeg" /></figure><h3>Overview</h3><p>The venue for this year has been the <a href="https://www.google.com/maps/place/Alte+Kongresshalle">Alte Kongresshalle</a> which is very easy to reach. There were 21 speakers from all around the world with a very wide range of topics for talks. Quite a few sponsors supported the conference but it was good that they were not pushed in any sorts onto us attendees. Most interesting definitely was the existence of another track besides the regular talks — the so-called <em>In-Deep</em> track. There, e.g. members of TC39 (like <a href="https://twitter.com/bterlson">Brian Terlson</a>), npm (<a href="https://twitter.com/ag_dubs">Ashley Williams</a>), or the V8 team (like <a href="http://benediktmeurer.de/">Benedikt Meurer</a>), (and lots more) hosted discussion rounds and provided deep insights into various topics — the best part of it really being the familiar and close atmosphere.</p><h3>Highlights</h3><p>I just want to share some of my personal highlights — of course based on my personal interests — of the two conference days, my top 3 of each. Have a look at the <a href="https://2017.js-kongress.de/schedule-js-kongress-2017/">whole schedule</a> to see which topics were covered during the regular talks. <strong>Best thing ever</strong>: the talks were recorded and uploaded to <a href="https://www.youtube.com/channel/UCe2uw4y9oTxfq-Z61ODWRhg">YouTube</a> — so go watch them!</p><h3>Day 1</h3><h4>Reactive Brain Waves — Uri Shaked</h4><p><a href="https://urish.org/">Uri Shaked</a> held a cool talk featuring an EEG headset (yes, fancy brain reading at work!). He demonstrated a simple <a href="https://angular.io/">Angular</a> app which used the <a href="https://developer.mozilla.org/en-US/docs/Web/API/Web_Bluetooth_API">Web Bluetooth API</a> to connect to the EEG headset. The API he adapted himself streamed the received data using <a href="http://reactivex.io/rxjs/">RxJS</a> and provided a great example to demonstrate the use of operators. The final result was an eye icon opening and closing when Uri blinked. Overall the talk was a great walkthrough as what is already possible with external devices and tools at hand. Uri also made a writeup and follow-along Gist with more details — <a href="https://gist.github.com/urish/f641c126ab45e8abefa3a9ab3d3575ca">check it out</a>.</p><p><a href="https://www.youtube.com/watch?v=hYiQ5svNjtw">Watch on YouTube</a></p><h4>A Tale of TurboFan: Four years that changed V8 forever — Benedikt Meurer</h4><p>As a member of Google working on the V8 Engine, <a href="http://benediktmeurer.de/">Benedikt</a> showed to what lengths the team goes in order to continuously make V8 the fastest JavaScript engine out there — check out <a href="https://docs.google.com/presentation/d/1UXR1H2elTdAYJJ0Eed7lUctCVUserav9sAYSidxp8YE/edit#slide=id.p">his slides online</a>. By showing some examples of JavaScript and how it translates to later machine code he elegantly and understandably explained common patterns in engines to run code as fast as possible. Benedikt also detailed the architectural structure of V8 and its TurboFan optimizing compiler. For me this definitely helped to get a deeper understanding of how my (maybe not so performant) code gets executed — and to loosely quote Benedikt: “The JavaScript code is written for developers, not for machines.”</p><p><a href="https://www.youtube.com/watch?v=cvybnv79Sek">Watch on YouTube</a></p><h4>Building a radio data network with Node.js — Thomas Watson</h4><p>Probably my favorite talk of the day where <a href="https://wa.tson.dk/">Thomas</a> demonstrated how to use a Node.js application to send and receive data via radio waves. He used a <a href="https://greatscottgadgets.com/hackrf/">HackRF One</a> for transmission — due to its wide frequency range he also showed how to remotely control a kids RC car. His talk also included background information explaining the theory of radio waves and how data can be transported using them. During one of the coffee breaks I also had the chance to have a nice chat with Thomas, very friendly and interesting sport!</p><p><a href="https://www.youtube.com/watch?v=5Xz6LJuvHCc">Watch on YouTube</a></p><h3>Day 2</h3><h4>Tomorrow’s JavaScript debugger — Amit Zur</h4><p>In my view the best talk of the conference in which <a href="http://amitzur.com/">Amit</a> showed the new <a href="https://github.com/devtools-html/debugger.html">Mozilla Developer Tools</a> and what he has already contributed during their complete rewrite. For me it was a pleasant surprise to see how far the tools have been developed and great to hear that they are based on standard and modern web technologies. Amit also made sure to underline the importance and simplicity of contributing to open source projects — like the <em>debugger.html</em> project. I also spoke with Amit about his next ideas of features for the debugger — <em>Debug Trails</em> and potential ways of supporting some kind of reverse engineering or debugging without proper source maps…</p><p><a href="https://www.youtube.com/watch?v=Rop3EgPvBMw">Watch on YouTube</a></p><h4>OUTBREAK: index-sw-9a4c43b4b4778e7d1ca619eaaf5ac1db.js — Alexander Pope</h4><p>Cryptic title but a haunting and entertaining story-telling talk by <a href="https://twitter.com/popeindustries">Alexander</a> showing the (potential) dangers of <a href="https://developer.mozilla.org/en-US/docs/Web/API/Service_Worker_API">Service Workers</a>. It was filled with valuable tips and tricks on how to properly use them — and especially avoid common mistakes. Otherwise you might be haunted by those errors for a very long time… The talk helped me to get some overview of an API I have never used so far.</p><p><a href="https://www.youtube.com/watch?v=cHDhGG-wPkk">Watch on YouTube</a></p><h4>The parallel future of the browser — Lin Clark</h4><p>The closing keynote on day 2 was given by <a href="https://code-cartoons.com/">Lin</a> and held what the title promised. Slides filled with self-drawn illustrations outlined the path of the — at this day — released Firefox Quantum and its huge technical changes in its core. The new Firefox is able to process even single tabs in multiple parallel tasks and Lin presented the challenges the team had to face in order to reach this goal (shaving some Yaks here and there…). For me the talk was really revealing in terms of how browsers need to continuously evolve and invent to keep up with the ever changing demands (from us developers, too).</p><p><a href="https://www.youtube.com/watch?v=xCBWS8G5Cm0">Watch on YouTube</a></p><h3>In-Deep Track &amp; Office hours</h3><p>Every day at 10:40am the organizers invited everyone to come and hand in their topics to make the schedule for the parallel <em>In-Deep</em> track. It comprised a very broad selection of topics ranging from discussion rounds to presentations of certain technologies. To be honest, the quality did vary in the ones I attended but nevertheless it is great to have people come up and present — there is always room for improvement. Moreover, besides the <em>In-Deep</em> track there were also so-called <em>Office Hours</em>, where the speakers or other community members were available for questions or discussions so one could just grab the expert he needed.</p><h3>Overall Impression</h3><p>To sum up it has been a very good conference from my point of view. There were awesome talks, great people, and all this in a really friendly atmosphere. I met new faces and made interesting contacts — apart from learning new things! I can only recommend paying close attention and getting a ticket to <em>JS Kongress 2018</em> when they will announce it since I will definitely be looking forward to it :)</p><p><em>Originally published at </em><a href="https://www.coderose.io/js-kongress-2017/"><em>www.coderose.io</em></a><em> on November 23, 2017.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b37f4654be05" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Evolving Processes — Part 1]]></title>
            <link>https://medium.com/@_rose_m_/evolving-processes-part-1-d58187820e66?source=rss-8842af9faec7------2</link>
            <guid isPermaLink="false">https://medium.com/p/d58187820e66</guid>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Michael Rose]]></dc:creator>
            <pubDate>Wed, 01 Nov 2017 07:53:27 GMT</pubDate>
            <atom:updated>2017-11-21T06:45:45.846Z</atom:updated>
            <content:encoded><![CDATA[<p>Since I started working at <a href="https://www.collaboration-factory.de/">collaboration Factory AG</a> in the beginning of 2015, our software development process has changed quite a bit. In this post I want to outline where we started, what challenges we faced, and which solutions we tried to employ to solve those. Very likely, lots of companies have come across these issues, too, and we are still far away from perfect as you will see in the end. I would love to hear your feedback and experiences as well as the measures you have taken to improve and streamline your workflow.</p><h3>Starting Up</h3><p>By early 2015, the company in total comprised less than ten people. As such, the development team was small with only 4 people, too. Everyone of us basically knew our product very thoroughly — there were some but comparably few core components, projects, and especially only a few customers to deal with. Therefore, we were in a sense <em>omniscient</em> — we knew everything about our code, features, and requirements.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/521/1*uQjVJfrVCWWaGre3NnQYZg.png" /></figure><p>It looks — <em>and really was</em> — simple and allowed us to very rapidly develop the product, implement new features, or react to changing customer requirements and feedback. Since the team fit in a single room and of course talked with each other during the day, every piece of information like what had just been implemented immediately spread and was shared. At the core the process solely consisted of a very flexible, self-organizing team.</p><h3>Then came Growth</h3><p>However — or better: gladly — growth set in, as a small, startup-like company eagerly hopes:</p><ul><li>new developers and consultants came on board of the company</li><li>the product grew with a rapidly increasing feature set</li><li>the number of core components, projects, and especially customers largely increased</li></ul><p>Apart from all the positive factors this growth caused some problems regarding our so-far lightweight development process. Team members worked on various numbers of projects and core components in parallel to increase the number of out-of-the-box features as well as adapt our product to specific customer requirements. Would you have drawn a diagram — it probably would have looked like below, <em>not simple anymore</em>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/658/1*tp1qa8lTXY1hTNw31VouDQ.png" /></figure><p>Quite quickly, one of the key advantages of the small, coherent team started to fade: it became <em>a lot</em> harder to share knowledge among the members of the team. A main driver of this problem were the customer projects since they typically require a lot of domain specific knowledge. It would not have been efficient — if at all feasible — for every team member to know every detail from a customer point-of-view.</p><h3>Real Agile to the Rescue (?)</h3><p>To tackle these issues we decided to take a more or less proper and <em>Scrum-like</em> approach. Why just <em>Scrum-like</em>? Despite introducing the <em>Sprint</em> concept, tracking a <em>Backlog</em>, and doing a <em>Planning</em> every three weeks, we did not fully <em>enforce</em> it because it was not really possible.</p><p>What is not enforcing a <em>Sprint</em>? In our case, it was hardly possible to properly plan the upcoming Sprint in the Planning meeting. The development cycle and deployment dates for customer projects were not in sync with our internal Sprint track. Thus, requirements changed frequently — especially in prioritization — bugs had to be fixed or deployments had to be built and delivered. The core concept of having an un-interrupted, committed to Sprint with focused work was constantly violated. Furthermore, developers were still separated project-wise and such information did not spread as expected.</p><p>While providing a good framework with rough rail guards, Scrum itself just did not fit for us at that time. Really enforcing it would have cost flexibility that was necessary business-wise to fulfill customer requirements to get and keep them onboard.</p><h3>Battling the Odds</h3><p>Our decision was to keep the idea of a Sprint — the fixed time frame and a regular Planning meeting every three weeks. The focus of Planning shifted to a couple of things:</p><ul><li>a quick recap of the last Sprint and how every project had evolved — i.e. some kind of Retrospection</li><li>prioritizing core issues of our standard platform</li><li>doing a rough resource planning to decide who works on which project to what extent</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/766/1*RaDHUuOlEQIS8bF0ttgZbQ.png" /></figure><p>Project specific details were discussed face-to-face with our consultant working together with the customer. If new issues came in with a higher priority they were implemented first in a Kanban-like fashion. As such we could drastically reduce the amount of interruptions. This approach also allowed us to handle the oncoming wave of challenges.</p><p>In the sake of these changes we also introduced a new development workflow regarding how we deal with code changes of our core platform in general — our new approach to Pull Requests on GitHub is <a href="https://www.coderose.io/life-of-a-pull-request/">an article on its own</a>.</p><h3>Remaining Challenges</h3><p>The little overhead in terms of unnecessary communications — “need-to-know” — meant that knowledge sharing did not improve. Without a proper Backlog and Planning, forecasting which features and functions would be included in the next release was nearly impossible. Moreover, releases were still driven by customer projects, i.e. whenever a customer milestone was up ahead, a release branch was created and deployed. This also meant that features and breaking API changes were built on top of a published release branch (<em>insert crying kittens</em>).</p><p>We are still finding our way and clearly see the need of improving our development process to fit our current state — in terms of how we work as well as how large the company has grown now. It will be a long one to go but an interesting one with lots of things to learn. I will continue to post about our evolvement and look forward to reading your feedback.</p><p><em>Originally published at </em><a href="https://www.coderose.io/evolving-processes-part-1/"><em>www.coderose.io</em></a><em> on November 1, 2017.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d58187820e66" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Life of a Pull Request]]></title>
            <link>https://medium.com/@_rose_m_/the-life-of-a-pull-request-3a5d62de1ebc?source=rss-8842af9faec7------2</link>
            <guid isPermaLink="false">https://medium.com/p/3a5d62de1ebc</guid>
            <category><![CDATA[github]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[pull-request]]></category>
            <dc:creator><![CDATA[Michael Rose]]></dc:creator>
            <pubDate>Thu, 10 Aug 2017 15:41:00 GMT</pubDate>
            <atom:updated>2017-11-18T12:33:21.516Z</atom:updated>
            <content:encoded><![CDATA[<p>Everyone of you is most likely familiar with <em>Pull Requests</em> (PRs) on <a href="https://github.com/">GitHub</a>. They are an essential tool for collaboration on projects — be it open source or commercial ones. As such they are a key utility for our daily work at <a href="https://collaboration-factory.de/"><em>collaboration Factory AG</em></a>, especially concerning the development of our core framework — the platform <em>cplace</em>.</p><p><strong>Disclaimer:</strong> Since we had some issues at the beginning with how to best deal with PRs in our development process I want to describe the way we handle them currently. This post therefore is an opinionated and maybe flawed way of managing them — so if you have any suggestions for improvement just leave a comment below and I’ll be happy to follow up.</p><h3>The original Problem</h3><p>As mentioned above PRs are one of the key utilities to manage changes to the core part of <em>cplace</em> — the so-called <em>platform</em>. Basically every change needs to be reviewed by at least one other team member in order to increase quality and reduce error rates. GitHub provides a very nice and practical UI as well as the tools necessary to efficiently deal with PRs: review, comment, discuss, and finally close them.</p><p>The main issue we had however was an ever growing list of outstanding PRs. States were not clear, i.e. it was not immediately visible which were awaiting to be reviewed, who was responsible for asking for help or closing the ones already completed. As a result we had some dead PRs lingering around for months and cycle times measuring the duration from opening to closing were horribly high.</p><p>We therefore set out to tackle this as the main issue: reducing cycle times of PRs in order to make new features and bugfixes available to customers and partner developers quickly while maintaining high standards.</p><h3>The Labeled Lifecycle</h3><p>One of the features GitHub provides was used by us only to a marginal extent — labels. However, they are easily filterable and allow to quickly categorise PRs. Our new workflow consequently makes good use of them. Let‘s take a look at the complete lifecycle of a PR as shown in the image below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/926/0*20TI-Cs4HZfzGCbL.jpeg" /></figure><h3>1. Creating the PR</h3><p>When work on the PR starts it is initially created with the label working. The <em>Assignee</em> of the PR is set to and will always stay its original creator. As such the responsibility of working, tracking, and following up on the PR is defined clearly.</p><h3>2. Requesting Reviews</h3><p>Once the implementation has been finished the creator sets the label to needs review. Thereby others know which PRs are awaiting another pair of eyes for cross-checking. We don‘t really make use of the <em>Reviewers</em> on the other hand since we want to encourage all of our developers to review as much as they like and especially from all areas of the system, i.e. not limited to their special domain. This boosts confidence and knowledge across the whole platform.</p><h3>3. Conducting a Review</h3><p>Developers pick any PR labelled needs review and make a thorough review of the changes. As a result the PR can be either <em>approved</em> or <em>changes requested</em>. We do not use <em>comment</em> as the outcome of a review which I will explain further on.</p><p>Approval is the easiest workflow — what remains for the reviewer is to set the PR‘s label to ready to merge. The <em>assignee</em> - i.e. the original creator - thereby knows the review has been conducted and completed successfully.</p><p>If the reviewer on the other hand has any remarks he marks his review as <em>changes requested</em> and assigns once again the label working. Every developer then can filter all PRs with the label working and state changes requested to quickly find the ones which require additional effort and changes by him. We decided to ditch <em>comment</em> as review result in order to make clear that an additional discussion is required - we further add an additional question label (not shown in the image) to these PRs.</p><h3>4. Follow Up Work</h3><p>Every developer checks his PRs according to their states. The ones with ready to merge can now be immediately merge into their destination branch adding a proper detailed commit message to be used for changelog generation. The only exception is if the creator wants other developers to do an additional review - in this case he resets the label to needs review and awaits another review.</p><p>All PRs which are labelled working and have <em>changes requested</em> need to be tackled as soon as possible and the required changes must be completed in a timely manner. Once these are done he changes the label back to needs review as described in step 2.</p><h3>No Improvement without Process</h3><p>Of course changing and deciding on a new workflow did not immediately change anything by itself. They key point was to introduce it as <em>process</em> or <em>discipline</em> in the daily workload.</p><p>We agreed to dedicate two separate time boxes per day of 30 minutes each to reviewing and follow up work — directly at the beginning of a work day as well as after lunch time.</p><p>The priorities are clearly defined as follows:</p><ol><li>Merge when possible — Check if you have any PRs ready to merge. Decide whether to request additional reviews or otherwise immediately merge the completed feature / bug / etc.</li><li>Follow up work — first everyone has to check whether there are any of his PRs in working with <em>changes requested</em>. If so take the time box to work on those.</li><li>Review — when there‘s time left pick any PR with needs review and start your review. Here PRs you already reviewed before and had re-work done have priority over completely fresh ones.</li></ol><h3>Conclusion</h3><p>By better leveraging the tools provided by GitHub and introducing our own lifecycle we could greatly improve cycle times for PRs. There are hardly any — yes, sometimes they still exist — PRs which await completion longer than a few days. It also helped to keep the number of open PRs down and constant at around 20 to 25. In combination with the assignment of <em>Milestones</em> we are now able to prioritise and focus our work on what is necessary.</p><p>Are you using your own labels for lifecycle and workflow management of PRs on GitHub or are you solely relying on the standard review states (<em>approved</em>, <em>changes requested</em>, <em>commented</em>)? Let me know if you have any questions or suggestions for improvement in the comments below and I will follow up as soon as possible.</p><p><em>Originally published at </em><a href="https://www.coderose.io/life-of-a-pull-request/"><em>www.coderose.io</em></a><em> on August 10, 2017.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3a5d62de1ebc" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>