Week 14

Published on Author malmLeave a comment


Agile and Scrum iterative development is now so normalised in modern software engineering culture that it’s not surprising to encounter evidence of fundamentalism in job profiles such as the following for a UK engineering leadership position spotted on LinkedIn this week:


Such extremity ought to fortify doubters.  Agile approaches can certainly aid delivery status visibility of timeboxed projects such building a web site or developing a brand app.  It is, however, unwise to consider them a template for universal application regardless of circumstance.  In fact there is evidence that in certain situations the use of agile can be counterproductive.  The topic is extensively explored in the inevitably numerous, polarised responses to a recent leading question posed on Quora: “Why do some developers at companies like Google consider Agile development to be nonsense“.  The top rated response was a forensic Fisking of each item of the Agile Manifesto highlighting a variety of scenarios where the adoption of agile will not help.  The commentary on items 4, 5, 9, 10 and 11 particularly resonated with personal experience.  Distilling down to three major areas where agile proponents run into difficulty:

  1. Organisational alignment: If the environment you’re working in doesn’t practice agile across key non-engineering functions such as procurement and product management, any impact from its ‘adoption’ will inevitably be limited.  Ultimately agile is about corporate culture and mindset not a ceremonial agreement to conduct daily standups and dispense with documentation within an engineering department.   Obviously the smaller your company and the less outsourcing you do, the easier it will be to align to an agile approach.
  2. Legacy platform: If you’re responsible for evolving a complex extant platform which business critical services and features depend upon, your scope for agility is likely to be restricted by legacy maintenance considerations.  The more legacy you have, the more planning will be required to address regression testing of what may appear to non-software people to be ‘simple’ change requests.  The territory here was famously explored in Eric Lippert’s classic response to the naive question “How many Microsoft engineers does it take to change a light bulb?“.  It also explains why, as the Quora question clearly implies, agile gets a colder reception within the GAFA quartet where core platform development is a primary engineering function.
  3. People: If you have an unclear product vision and poor leadership, your return on any investment in agile will also be disappointing without a concomitant change of guard. No methodology alone can create great products. You need highly skilled people, empowered to self-organise and inspired to go beyond the confines of the ordinary as this commentator on the Quora thread points out:

“Agile/scrum promotes the false myth that process can replace product design and architecture. The management fairytale is that great products can be reliably developed simply by obsessing over process.”

The reality is that software development is ultimately about striving for cost-effective excellence in engineering.   Experience suggests that is actually pretty boring to most people who aren’t engineers. Which is good because there are lots of reasons to admire being boring in engineering.   An Amazon paper detailing the use of formal methods to develop AWS technology is an illuminating insight into what’s really involved in world class platform development; an uncompromising focus on ‘boring’ technical quality and little evidence of ‘agile’ terminology.

Moving from a cargo cult agile fundamentalism to a more nuanced and skeptical perspective requires a genuine interest in what really goes on day to day in ‘standard engineering’ starting with the recognition that capacity is limited and must be prioritised then encouraged to self-organise in order to deliver.  Often that means making tough business choices between a new feature and quality improvement which is a common source of tension between engineering and business stakeholders:

“The move from selling software products to selling and running services in high-availability environments led to a high load of operational tasks. The majority of this work is often unplanned and unaccounted … thus resulting in delays and blame wars in some organizations. The remaining minority of work is then dedicated to business and IT projects.”

Devices and Manufacturers

“Sometime in the past year, the bar for what makes a good smartphone got raised. It has to look good, it has to have a great screen, it has to last all day, it has to have elegant software, and most of all it has to have a really good camera. Miss one of those marks, and you haven’t just failed, you’ve probably lost a customer for years.”

  • They found much to admire in the S6, especially overall product design and camera capability where Samsung appear to have closed the gap with Apple: this is probably the closest any Android phone has ever come to the “just shoot” mentality [of] the iPhone“.  The Verge went on to undertake a detailed camera comparison between the S6 camera with the iPhone6 cameras to quantify their advances in smartphone photography:

“On the whole, the S6 holds its own against the iPhone, and we wouldn’t hesitate for a second to use it as our primary smartphone camera.”

Galaxy S6 review images

  • The confidence and ambition of Indian OEM Micromax was outlined last week.  Their ultimate aim is to become a global top 5 smartphone player building on their success in retailing low cost devices in India where their focus has been on creation mass consumer brand appeal using FMCG approaches rather than technology differentiation:

“The use of a sachet to sell mobile phones is to “massify” the product not unlike how India’s consumer goods companies sell shampoos and detergents in small, cheap packets to attract more cost-conscious buyers.”

  • Investment overtures from Japanese operator SoftBank suggest a Micromax valuation of $5 billion which isn’t that far shy of the $7 billion that Microsoft paid for Nokia who were themselves for many years the top low-cost consumer phone brand in India.


  • The company started by Bill Gates and Paul Allen turned 40 this week, a milestone noted by Gates, who remains actively involved, in a company-wide letter that focussed more on its future than its past:

“I believe computing will evolve faster in the next 10 years than it ever has before. We already live in a multi-platform world, and computing will become even more pervasive. We are nearing the point where computers and robots will be able to see, move, and interact naturally, unlocking many new applications and empowering people even more.”


“When a customer interested in the Apple Watch Edition enters the store, he or she will be given no-wait access to a dedicated Expert, who will provide a personalized “journey” from the beginning of the appointment until the end, as much as one hour later. There will even be a new option for at-home video conferencing rather than in-store personal setup.”

  • Many commentators like ZDNet’s Larry Dignan find the idea of spending large amounts of money on technology that will be out of date within a couple of years to be “fundamentally absurd“:

“Short of a big run in the value of gold, it’s hard to see the Apple Watch Edition appreciating. Why? The technology will be dated. Technology doesn’t hold up to the test of time.  … Nevertheless, technology vendors will soon follow Apple into the luxury device market. There’s no choice. There are margins and brands to preserve. Along the way be prepared for a healthy dose of absurdity. Apple has the brand to pull it off. Most vendors don’t.”

Apple Store Render

  • It is a gamble but a calculated one as Apple are clearly operating a barbell strategy with Apple Watch.  The majority of variants are priced under $1000 and separated from the Edition range by a gulf in price dramatically illustrated by asymco below.  It remains to be seen whether the Edition gambit works out for Apple.  If it does it will profoundly change their future product direction as outlined here a few weeks ago.  They’re likely to shift a lot of the Sports variants either way.

Screen Shot 2015-03-30 at 12.03.19 PM

  • Daniel Eran Dilger provides a unique inside view of Apple in Hong Kong and Shenzen outlining the journey the company took navigating their way through WAPI and LTE differences as well as comparing Apple Stores in the two regions.  The first image is the Shenzen store and the second is the Causeway Bay store in Hong Kong:

Causeway Bay Apple Store

  • One Chinese billionaire that probably won’t be buying an Apple Watch Gold Edition is Jia Yueting who compared Apple to Hitler because of its “very rigid control over its ecosystem“.  As The Register point out, it’s a great example of Godwin’s Law in action.

Google and Android

  • Google have published a white paper outlining their work on improving Android security in 2014 both on the handset and backend in terms of Google Play app scanning where:

“To identify PHAs (potentially harmful apps) before they even appear in Google Play involves a machine learning system that apparently pokes at millions of data points and relationship graphs for its security detection screen.”

  • The paper reports that fewer than 1% of Android phones are carrying PHAs.
  • This post suggests that Android is “winning out over iOS” because it offers OEMs more scope for UI differentiation through integration of alternative launchers.  The definition of winning here is somewhat context-dependent.  First of all Apple are definitely winning out in terms of profit share and secondly as we have noted many times before, Google doesn’t benefit from non-Google variants of Android products which is where much of this UI differentiation effort is focussed:

“By default, the various launchers, homescreens, overlays, and other Android interface solutions have to provide a unique take in order to stand out from the default Android or OEM offering.”

  • Helping to emphasise that point, Chinese OEM OnePlus have released OxygenOS, their own take on Android Lollipop albeit without any Google Services involved:

  • Laszlo Bock, SVP of People Operations at Google, posted on LinkedIn about how Google employees are consuming less calories as a result of changes in environmental cues and nudging towards healthier options.   It’s all good of course but still somehow faintly reminiscent of Fitter Happier.

Apps and Services

  • Feedly continues to grow from strength to strength as the de facto replacement for Google Reader.  More evidence of their status as essentially the sole player in this field came this week in the announcement that Taptu is to shut down with the recommendation that its customers adopt Feedly instead.
  • In a new twist in the increasingly crowded streaming music scene, Jay-Z announced that Tidal will become an “artist owned” platform.  To underline his point, the press conference where he made the announcement was a surreal affair attended by Madonna, Daft Punk, Jack White, Kanye West and other notable luminaries who all support the move to a full subscription model:

“Tidal, which makes millions of songs and thousands of high-definition videos available in 31 countries, will have no free version. Instead, it will have two subscription tiers defined by audio quality: $10 a month for a compressed format (the standard on most digital outlets) and $20 for CD-quality streams.”

  • Many were distinctly unimpressed from a consumer perspective with the prospect of a Spotify clone that costs twice as much even if it offers the prospect of hi res audio.  Ultimately Tidal will rise or fall on the extent of truth behind the assertion that “people perceive music as something you no longer have to pay for“.

  • This excellent Eugene Wei post on “platform risk” contrasts the inherent risk of startups relying on little ‘p’ platforms vs. the relative security of building on top of ‘P’ platforms.   Amazon Web Services is an example of a big ‘P’ platform proposition. Twitter is a prime example of little ‘p’ as amply demonstrated by the recent Meerkat-Periscope saga:

“one must be willfully naive to consider ad-monetized social graphs like Facebook and Twitter to be capital P Platforms. I prefer to call them little p “platforms” … Developer beware. Unless they change their business model, any developer trying to build some other graph off of Facebook or Twitter should have a second strategy in place in case of explosive growth because access won’t persist.”

“if you can understand the user’s language and you can understand how they express themselves, there are opportunities to build really high context personalised prediction experiences that go beyond just text prediction”

Smart Retail


“It was also a clearly orchestrated affair, as it required “privileged access to backbone routers within its borders to modify the Baidu resources,” wrote Bill Budington, an EFF software engineer.”

Wearables and the Internet of Things

  • VisionMobile point out that today the Internet of Things is evenly split between standalone vertical manufacturer platforms like that of Garmin and more general purpose platforms such as Nest that cater for a range of different devices.  Innovation and progress will require cooperation between device manufacturers leading to a VisionMobile prediction that:

“the top IoT platforms in 2020 will be cross-vertical. Sector-specific platforms will be niche or in decline.”

  • The Olio smart watch is a unique high end proposition built on top of custom hardware and cloud-based software platform. Available in steel or PVD black, it will start at $595 and offers  the user the holy grail, “a smartwatch design to combat immediate obsolescence“.  It will be interesting to see how it does relative to the Apple Watch Edition:

“the Olio has a cloud-based system that only surfaces important messages as defined by the people you interact with the most. There are only a few screens available and the watch face itself displays the number of interactions per set interval of time using fanning lines of differing length.”

Olio Creates A Homebrew Smartwatch With Some Amazing Features

  • Xiaomi debuted smart scales that will cost just $16.   It underlines the company’s push to be a major player in the Chinese smart home and the importance it attaches to data over hardware margin which must be negligible at that price.

Raspberry Pi


  • Drunk user testing missed the cut last week.  It’s too good to pass by during April Fools week:

WHAT: I’ll get very drunk, and then review your website. I’ll send you a document outlining where I thought the website needed help, and a screencast of me going over the website.

HOW MUCH: $500 per site.

  • User Testing appeared to channel some of that alternative humour for their April Fools joke, a Hipster Testing service:

Software, Toys and Games

“In the end, while you may disagree, I place more value on writing code that is easily understandable (and avoids redundant computation) than on a correct implementation of a textbook algorithm dating from 1968, so you’re getting a -1”

  • Bayes’ Theorem in lego with Count Bayesie introduces the concept of a lego stud-based ‘probability space’ as a teaching aid for conditional probabilities as demonstrated below.  It turns out that working out the answer to the question “If we know we are on a yellow space, what is the probability it’s red underneath?” or P(red|yellow) can be intuitively gathered from observation – 2/3 of the yellow blocks sit on red in the probability space.  The formalisation of Bayes’ Theorem in terms of P(A) and P(B) is not necessarily so obvious though Count Bayesie gives it a go by counting up the various lego stud fractions for the corresponding rhs entities in the Bayesian equation P(A|B) = P(B|A) x P(A) / P(B):

“The probability of getting a yellow brick is conditional on whether you are on a blue or red space. In probability theory we express these conditional probabilities as P(yellow|red) which is stated “The probability of yellow given red”.

    1. P(red|yellowP(yellow|red) x P(red) / P(yellow)
    2. P(red|yellow= (4/20 x 20/60) / (6/60) = 2/3

Lego Brick Probability Space

  • Mattel have been under assault from a variety of Lego innovations in the last decade.  Hello Barbie represents an attempt to fight back with an app-powered product of their own. Billed as “the world’s first interactive smart doll”, Hello Barbie integrates technology from “conversational apps” startup ToyTalk and should launch later this year in what is already proving for some a controversial move:

This fall, Mattel plans to introduce Hello Barbie, a Wi-Fi enabled version of the iconic doll, which uses ToyTalk’s system to analyze a child’s speech and produce relevant responses.

Dr Richard Graham, a consultant child and adolescent psychiatrist who runs a technology addiction unit at the private Nightingale Hospital in London, sees children with a serious Minecraft habit. He talks about the game’s “hyper-reality” which he says makes the external world “slower, paler, less stimulating”.

Children playing Minecraft

Work and Culture

  • Continuing the theme of technology warping perceptions of reality, Nick Carr writes on evidence suggesting that prolonged human interaction with search engines literally changes the way your brain works and results in a unjustified overconfidence in one’s own abilities:

“searching the web gives people an “illusion of knowledge.” They start to confuse what’s online with what’s in their head, which gives them an exaggerated sense of their own intelligence. The effect isn’t limited to the particular subject areas that people explore on the web. It’s more general than that. Doing searches on one topic inflates people’s sense of how well they understand other, unrelated topics.”

  • Carr suggests the phenomenon is one of the consequences of an internet age where “processes of human memory are adapting to the advent of new computing and communication technology“. A rapid global move from “biological information storage” toward “digital information storage” may have further profound impacts on human memory.  Carr cites a study claiming that:

“simply taking digital photographs of an experience will tend to reduce your memory of the experience”


  • Engadget discuss super intelligent AIs with Ex-Machina director and noted author Alex Garland.  He appears to have read widely on the subject and reached the conclusion that with appropriate “checks and balances” conscious AI need not be feared.  He makes the point that there have been enough appalling ‘real’ humans throughout history without worrying about AI malevolence.

  • 80’s Alternative Intelligence Max Headroom hit the US 30 years ago.  The Verge looks back on a cyberpunk icon of the 80’s who disappeared almost as quickly as he arrived.


  • Stanford graduates are in such demand that some are being offered $150k entry-level salaries.  And then turning them down in favour of self-actualization considerations that range from “making an impact” and “doing good” to “working on interesting problems”.

“after six decades of independent scientific research, there is very little evidence for the idea that brainstorming produces more or better ideas than the same number of individuals would produce working independently. In fact, a great deal of evidence indicates that brainstorming actually harms creative performance, resulting in a collective performance loss that is the very opposite of synergy.”

  • Another HBR piece on the 5 biases pushing women out of science and technology points to evidence suggesting that gender bias not “pipeline” issues or personal choices push women out of science.   Rather depressingly gender bias still manifests in different ways for different racial groups:

“notably, nearly half of black women (48%) and Latinas (47%) report having been mistaken for administrative or custodial staff, an experience far less common for white (32%) and Asian-American (23%) women scientists.”


Leave a Reply