(This had more structure in my head before typing it out, but I’m posting it anyways.)

theory: if you don’t embrace remote workers, your geography dictates your platform choice

— adam goucher (@adamgoucher) June 2, 2013

OK. This got enough attention that it requires some expansion.

I live in the eastern suburbs of Toronto. As part of a bit of a geography lesson, the western suburbs have been traditionally home to various large tech companies due to its proximity to the international airport (and likely tax incentives). The eastern suburbs have been more manufacturing oriented around the [once large] car plant. As such, the tech ‘community’ out here is actually pretty small and disorganized when compared to both the size of the population and even the size of the population of people in tech that commute into the city for a living. [This is in spite of large efforts by some companies around here.] The make-up of the Toronto tech scene is also interesting. Yes, there are hip tech companies like ScribbleLive around, but most of the programming jobs are dominated by the various financial and insurance organizations or provincial government agencies that are based out of the core. And with those industries come ‘safe’ development environments, and large services contracts. Both of which equate to ‘Java’ and/or ‘C#’.

So let’s pretend that I was building a ‘traditional startup’ (an interesting juxtaposition of words if ever there was one) out here. And to throw an additional twist on the scenario, that I wouldn’t be the one doing the coding, what language / platform would I use?

Its actually not so simple a choice if I want to build for the future rather than build to flip.

Actually, let’s take a step back a second. It is actually a really simple choice if I embrace remote working as a fundamental building block for my company as I would just hire the best people and have them use the best platform for the job. If I follow the 1980s approach of having everyone in the same room, 9 – 5, things harder. A lot harder.

Again, using the eastern suburbs as my example, the majority of [experienced] programmers responding to your job postings are going to be working presently in the financial industry and are interested in your gig largely because it eliminates the commute. And it brings with them the baggage of that industry; Java and C# plus a much slower release pace than what one would expect from a startup. Which isn’t to say I wouldn’t hire someone out of that situation (heck, I started at a bank), but I would be very upfront with what they are getting into. But if you want to have the largest population to pull from, then your platform is going to be one of these two.

But let’s say you want to up your change of getting funded by using some buzzwords and you decided that what you want your startup to be around is Node and Mongo. Now you have a different problem — if you are not embracing remote workers. Let’s say that there are 300 people who know Node and Mongo within an hour’s commute from my this fictitious office. Total. Of those 300, let’s say that 50 (which is likely high) are of the level you want and haven’t just run through a tutorial or two on the internets. And of those there are maybe 5 – 6 that would consider leaving their current role and if you are really lucky one might live close-ish by. Otherwise you are going to likely have to overpay to have them deal with the lack of transit infrastructure out here. But that doesn’t necessarily scale. Go back in time 6 years and see how much more Rails programmers were getting paid for just this reason.

Take a moment and read yesterday’s Do you really know the business your startup is in? as this plays into the mix here. If you decide you are a technology company then you say ‘screw it!’ and use the cool tech, overpaying for talent if necessary, but if you decide you are more of a ‘service’ company then you focus less on the platform and more about the customer experience. So if you want people in chairs in the office then you need to factor in the skills and proficiencies of the pool you are drawing talent from when choosing a platform. Of course, if you are happy to have people working from anywhere in the world, but logged into IRC when working with a VOIP line available to them at any point then your pool is nearly infinite as are your platform choices.

Here’s a theory; most startups don’t actually know what business they are in. They do likely know the business they want to be in, but unless your company is guided towards that specific goal (and you have no small bit of luck) it is your customers who actually determine what your company is an does.

In the spirit of Ninja/Pirate/Elf/Dwarf this there is two different dimensions that go into this.

The first dimension is whether you are a ‘business’ oriented company, or a ‘consumer’ oriented company. And this is not always so obvious an answer. My company for instance is pretty easy to peg as a ‘business’ one — we help businesses do automation and continuous delivery better. Which is Google? If you answered ‘consumer’, I would argue you are wrong. The consumer usage of their properties enables their real business which is advertising — which is something businesses buy. These days ‘hybrid’ companies are developing but they are going to tend to skew more to one side or the other.

The other dimension is around whether you are a ‘service’ oriented or ‘technology’ oriented. Or put another way, which part of the business is the dog, and which is the tail. A service oriented company, which I would argue Freshbooks is one, will do things differently than a technology one. For instance, all new people at Freshbooks do a stint in the call center to learn the customer’s needs, etc. I have no idea if it will scale, but it does currently and is awesome. Yes, they have a lot of tech behind the scenes, but that enables their service. Etsy, I would label as a technology focused company. I’m sure they have support people there, but hitting a person is a last gasp event and I’m guessing (I could check, but why research something for the internets?!?!) new hires don’t do a rotation through the support team and you can find their developers at various conferences explaining how awesome their tech is. (Which it is btw!)

Sitting along any point on either axis is not in itself a bad thing — unless it counters what management and investors want. (Trust me). But where you sit on them will determine a lot of subtle things in your organization;

• Do you have ‘client happiness’ people, or ‘customer support’
• Do you have all-hands cheerleading meetings at the beginning of the day, or do you have remote workers with erratic shifts
• Do you jump when a customer wants something, roadmap be damned or do you rely on your own analytics and ‘lean experiments
• Are you going to use the bleeding edge ‘cool’ thing (these days, Node and friends I guess) with the risks that come with a rapidly evolving platform or something ‘safe’ like ASP.NET MVC or Ruby on Rails.

Coming back to the original idea, I think it would be an interesting experiment for a company to reflect on this, not just by those in management but those actually in the trenches. A mismatch means there is a problem, somewhere. A further inquiry of say, the top and bottom 10% of your clients by revenue might also be interesting. Again, if there is a mismatch, there is a problem.

(Now if only I had thought of this when building my pitch deck a couple months back — it would have made a good slide I think.)

Every company has a moment where they are forced to care, I mean, really care about Security and Privacy. Sure, you should always care in the back of your mind, but at some point it comes front and center and a key pillar of everything you do. The question is whether you are going to be ready for when it comes. Odds are, no. (Because if you were, you would already be caring.) I’ve seen this moment happen a couple times, so here is my list of things you now need to pay a whole lot of attention to.

First, Security and Privacy are related, but different. In hand-wavy terms; Security is related to keeping information in and the bad guys out, and Privacy is about individual people’s information in your trust. And it is a trust, not a right.

With that, here is my brain dump on the topics.

Security

• Who is the officially designated person for security problems?
• Physical security – padlocks on the cage, etc.
• Can people wander the office without being questioned?
• Is all input validated on the backend as well as the frontend
• Disarm all input
• Disarm all output
• Don’t blindly use user provided strings in queries
• PCI compliance?
• Don’t store your password in the clear…
• The first question about any change should be ‘what are the security ramifications’?
• Do security bugs get put into the main bug database?
• How are security things disclosed

Privacy

• Who is the officially designated person for security problems?
• Do you know what PII (Personally Identifiable Information) you have
• Do you know what flows the PII participate in?
• People need to be bucketed based upon their informational needs
• Don’t clone your production database down into lower environments — without scrubbing the PII
• Do not display PII to anyone other than the owner of it. (Unless they really need it.)
• PII should be encrypted in the database
• Thou shalt not grant access to the production database
• Log the stink out of access, including read, of things that contain PII
• The second question about any change should be ‘what are the privacy ramifications’?
• Do privacy bugs get put into the main bug database?
• How are privacy things disclosed
• What legislation around PII affect you? And in what jurisdiction? Is it where the head office is? Where the app is hosted? Neither?

Writing secure code is actually not that hard. We know how to do this. At this point it is really just sloppy code. But putting the people and process around how the code that is generated is hard.

This is just a quick off-the-top-of-my-head list, but the key thing to remember is that its not paranoia if they are actually after you. And ‘they’ are.

A pattern than I unwittingly did and have noticed in others is that you need to work out what your variation of ‘consultant math’ is for yourself. No matter how many people tell you, you need to blunder through to your own understanding of it. Often by going waaaaay overboard on conferences the first year or so after going independent.

Here is my variation of ‘consultant math’ around whether I will attend your conference or not. Not that anyone will listen to this and not thing ‘hey! I’m different!’ — no, no you are not. But anyways…

The first part of the equation is whether or not I am getting paid to be there. This can either be in term of an honorarium, or flight or hotel (or all three!). Consultant math is pretty easy when this happens … in most situations. But there is the whole problem of Opportunity Cost. Let’s say you are giving me all three of these variables to attend your conference [presumably to speak] but I have to spend a complete day flying there, and a complete day back. That is really 3 days I could be billing people to work on their stuff. If the honorarium only covers 85% of a single day’s worth of billing, well, things start to break down a bit.

Fear not though, the math still might work in your favour if there is a good chance I am going to get new business from attending an interacting with the other attendees. This actually has two parts to it. First, you need to know who your target demographic is. After 3+ years at this, I have a pretty good feel on who I am targeting and the sorts of things they need to hear in order to get them to sign on. (New consultants are unfortunately at a disadvantage and I think this is a large part of the reason why they go to so many conferences in the shotgun approach.) You, as a conference organizer need to convince me that there is a match between my target demo and your audience. (Tip for consultants: check their sponsorship packages, they often include this information in there.) If there is a match, I might eat the Opportunity Cost and turn it into a Marketing Cost.

There is another variation of Consultant Math I have been tinkering with a bit recently, and that is the Sponsorship Option. This is advanced level math and should only be employed once you know your audience since it often requires more risk [in terms of outlay of money]. Upon successful completion of this math you end up sponsoring a target conference [or event association with it]. For instance, the Consultant Math with Sponsorship Option looks quite favourable for Pycon or php[tek], but is pretty abysmal for CAST, Star*, STPCon or even, ironically, Selenium Conf.

I explained Consultant Math to a could non-consultants today [including ish a conference president] which is usually a sign I need to blog about [my variation] it.

So I’m in a pitch contest right now, and easily the biggest outcome is to switch me mentally from ‘talking about my app’ to ‘building my app’. And part of that shift was to got to TorontoGROWtalks last week. Waaaay too much stuff to absorb. (And thank-you startupnorth for the discount code!). Notes are as follows…

Brant Cooper

• took 3 minutes to mention FAKEGRIMLOCK
• Innovation: Disruptive, Sustaining or Rippling?
• hockey stick growth curve is an inverse mullet
• disruptive organizations optimize for learning
• sustaining organizations optimize for execution
• pre product fit, you are pushing the product on the market. after product fit, market pulls the product from you — I’m pushing now, need to stop that, and soon.
• first vp of sales needs to be able to sell without a demo — I don’t look forward to hiring sales.
• figure out your marketing, by going out and sell
• product creates the buzz, marketing is about amplifying your buzz
• you are not ready to market until there are things in the top of your sales funnel
• how do your measure a ‘passionate user’? figure out how to measure satisfaction and then passion
• if you have sharing, measure sharing since its an indicator of passion

Mike Meltzner

• product ownership isn’t an authority, it is a responsibility
• voice decisions the rest of the group has already decided on
• Startup stages: a comparison of 3 models — Marmer stages was mentioned
• job of ceo is to say same 3 things to everyone outside of the company (Speak the vision. Keep the tank full. Build a team). job of product owner is to say same 3 things to everyone inside of the company (important thing this month, thing we’re ignoring, where we are going)
• be in everyone’s business everyday so they know to include you in decisions
• default to no — but help me to help you convince me to say yes
• no — for now…
• who is handling the project — starters, or finishers?
• unsuccessful product ownership involves handoffs
Kate Rutter
• was an updated version of
  
• 
• ux != ui
• ui = delivery mechanism
• ux
  • a mindset
  • experiential payoff
  • inspires the right kinds of ideas
  • guides decisions
• if people are not going to pay for it, you don’t have a business, you have a hobby
• make decisions intentionally
• startup workshops
• if you cant have one user being passionate about it, how do you plan on having 1000s?
• ‘you dont have a product box, you have a product sieve’
• the market will give you permission to succeed
• customers [give you money] vs end experiencers
• cohort watching is useful
• there is nothing better than being schooled by your product
• each stage of the stack has its own ways to test it
• you should be okay looking like an idiot, because this is how you learn

Laura Fitton

• is there any conceivable world where someone would pay for this blob of marketing
• be useful
  • relevant to your user
  • relevant to your product
• start responding to how they make decisions
• its not spammy feeling if it is properly segmented
• HubSpot Academy
• will customer’s thank you for your content?
• hits don’t count; conversions count
• ‘them’ vs your ‘them’
• ‘smarketing’ — combination of sales and marketing

Scott Kveton

• creation stories should all be like urban airship’s
• ‘more of a feature than a product’
• its not lying, its selling forward
• think hard on pricing and packaging
• you really /do/ need to listen to customers
• customers will direct your business far better than you can
• outgrew ec2 within 3 months
• torn down and re-built tech 3 times

Dan Martell

• to have a hockey stick you need to have the flat part
• as soon as you truly understand your core product, get rid of everything else
• you need to instrument /everything/
• know your baselinesactivity streams
  • curious
  • casual
  • core <--- pay attention to these ones
• use in app surveys
• know your funnels
• unless you have core, don’t do this…
  • opn (other people’s networks)
  • shareable moments

Michael Litt

• solve the problem the customer has, not the problem they think they have
• do not trust the data, talk to the customer
• metrics only work /after/ you have product-customer fit
• everything at the start should be outbound
• ever developer should talk to customers

Mark organ

• most important ingredient for founder or investor: conviction
• ‘how does cookie monster mean capital’?!?!?!
• you need to have deeper knowledge about things more than anyone else
• to bootstrap
  • conviction
  • microniche
  • serviceize (charge for anything)
  • high pricing (multiply by 3x – 5x your competitor)
  • systematic scale
• to raise money
  • conviction
  • beautiful story
  • abp (always be pitching)
  • get feedback and connections
  • firm, but coachable
• the best way to get a canadian investor is to get an american investor first
• you do have spend time on planes to raise money

The Call For Participation for CAST 2013 is out and includes

We would like to hear your experiences, stories, thoughts, observations, demonstrations around the lessons that you have learned in Software Testing, as well as how these could influence the way that we approach testing in the future.

And I expect a lot of interesting experience reports from interesting people (it is CAST after all). But how much really, really, new stuff will be added to the craft. I’ll be my normal cynical self and suggest that it is likely not as much as one would hope.

What I would love to see is a CFP that looks something like this.
We would like to hear your experiences, stories, thoughts, observations, demonstrations around the lessons that you have learned outside of Software Testing, as well as how these could influence the way that we approach testing in the future.

My talks a couple years ago all fell into this form. Kids in Armor and Testing Inspiration When You Least Expect It are both examples of what I want to hear people talk about. I want to hear Alan Page discuss how Orchestra composition helps him test better, how Ben Kelly tests better due to years of thwacking people in the head with bamboo, etc.

I also want non-testers to be brought in as the keynotes. Sure, I enjoy listening to James, Michael, Cem, Matt and company, but if you are keynote-ing, you are up against an astronaut in my ranking scheme (Keynote vs. Track). Now I am a baseball fan so opinions are skewed but I think the Garfoose or Shawn Green could also be great. Though they likely cost more since its an actual speaking gig rather than just the prime audience grabbing moment. Or Mary Robinette Kowal about puppetry. Or a Buddhist monk about meditation and breathing. Or. Or. Or.

That there is a speaking circuit where one can recycle topics is a smell I think …

… and another smell is that I essentially wrote this same post back in April including some of the same speaker suggestions; The ‘Un-Testing Conf’. That’s just silly. I should do some work…

Speak Up! is the latest of Jesse Noller’s community-oriented projects. [I’m pretty sure he has cloned himself to do all this stuff.] Having sat in many ‘how to give a talk’ talks (meta!) I’m amazed how often people focus on construction of slides, the narrative and practice (feh!) and forget the most important thing.

BE YOURSELF!

I’ve been saying this for a while, but this paragraph from Cheek was my hero sums thing up well.

Later that season, Cheek would give me the most important piece of advice anyone has in this business – Be Yourself. It’s easy to lose that once you get close to the pinnacle of the profession, to try to change who you are or how you do what you do in order to impress people or to make your way up the ladder that much easier, but Tom told me squarely that there was a reason I was in that booth with them and it was because of what I’d done on the way there. That my success was based on how I went about my work both on the air and off and that any future success would be based on the same thing.

Pulling it back to speaking, people want you and your content not something and someone else when they select you. Don’t change. Even if that means blatantly ignoring cultural norms. For example:

• When I did Set Course for Awesome I was warned that having cussing in the deck could be offensive to the audience. Noted. Did I change it? Not a chance.
• If I am speaking at your event I’m going to wear what I am comfortable wearing that day which is likely to include some sort of snarky tshirt. Looking at some of the photos from Justin Hunter who is over in India at a conference today the audience is in jackets and ties. I don’t think I own a jacket and tying a tie is an exercise in hilarity.

Also, if you put a mic in front of me, you are going to get snark. And opinion. Because that is me. That’s what you get. Anything else would be dishonest. And being dishonest to your audience is not a good way to engage them.

Yes, I think that slides that have fewer words are better in general and that you should tell a story when behind the mic. But if that isn’t you then don’t do it. Be yourself.

One of the traditions in my wacky subset of the testing world is not only building mnemonics, but adding / remixing them. And so, I introduce COP FLUNG A GUN. Which is COP FLUNG GUN plus Automation.

Automation

The primary way to test mobile apps remains to be, literally, by hand. But we’re starting to figure out how to do it though automation. (No thanks to the OS vendors…) As automation frameworks get more mature we should check whether automation hooks are provided by the developers. Unique ids or other identifiers for all interaction bits, synchronization points exposed, etc.

Also included in this is integration into your CI system, since that is the bit that controls the execution and reporting of the automation.

There are certain things that I can say have been with me most of my life. The Dark Tower series is one of them. (The Wheel of Time series, Star Wars, comicbooks have as well. Likely baseball too now that I think about it.)

So last night I’m paying ish attention to twitter while watching the world series and see this flit across my stream.

Wow! Finally finished Stephen King’s Dark Tower series. Those are long books. Enjoyed it, and liked the ending.

— Corey Haines (@coreyhaines) October 28, 2012

Which got me thinking about the whole series. Yes, the whole series. This is the point where my world-builder’s-disease kicks in. See, the Dark Tower is not just the ‘core’ seven books. It is in fact pretty much all his books. If it had a wizard, it is a Dark Tower book. If it had a lost puppy sign, it is a Dark Tower book. If it happened in Derry, it is a Dark Tower book. And yes, The Stand is absolutely a Dark Tower book.

But does the Dark Tower deserve all the praise it gets? I dunno.

For pure scope and intent, absolutely. But that almost feels like it was added after the fact. Or at lease communicated afterwards. Anyone reading King long enough understood his ‘universe’ held Derry near its center, but then suddenly it was part of something else. And his books would have asterisks next to books that were part of the Dark Tower. If it was always his intention to have things part of a larger epic, how come they were not there in the books that came out in the 80s and early 90s?

As for the individual books, on the whole I didn’t really enjoy them. The Gunslinger is a decent read, the next two feel like they are just moving the plot along and the flurry of the last three after his accident feel like a person’s sudden response to facing their mortality and fearing their epic won’t get completed. His inclusion of the accident in the final book seemed heavy handed and too breaking-the-fourth-wall ish. I literally stared at the page in disbelief when that happened. It is an interesting thought experiment to see how it all would have turned out had it not been for that fateful day…

‘Wizard and Glass’ however is an outstanding book and likely my favourite King book (with The Stand (unabridged) being a close second). Its enough of a standalone story that even if you have no want to read the whole series, it is worth a read. ‘The Wind Through the Keyhole’ is much the same tone and feel as ‘Wizard and Glass’ though outside of the main series. Chronologically it takes place before ‘Wizard and Glass’ but is another standalone book.

The more I think about it, the more I feel like the Dark Tower series proper has failed yet the world it has setup has succeeded greatly. ‘Wizard and Glass’ is character backstory set in the world. ‘Keyhole’ is character backstory about the world. ‘Hearts in Atlantis’ is secondary character backstory, etc. Similar to how the Terminator and Star Wars franchises were flushed out through Dark Horse Comics in the 90s, the Dark Tower is being flushed out by Marvel Comics.

What I think I would like to see however is more ancillary stories come out set in the Dark Tower world. Maybe not even by King, kinda like the Star Wars universe. (See The Thrawn Trilogy for instance to see how other can expand a universe while still staying true to it.) But now that Roland has been eight years since Roland reached the top of the tower and ‘Keyhole’ only came out this year gives me great hope that the world will continue to produce things to consume my money.

Other random bits in my head around the Dark Tower…

• ‘The Stand’ is a pretty ambitious book to choose for your first highschool book study
• I read ‘The Drawing of the Three’ (or started at any rate) during a school trip to Stratford to see ‘As You Like It’ and was listening to The Cure’s ‘Mixed Up’ so certain songs are forever associated with the book (such as ‘The Forest’)
• After reading the last page of ‘The Dark Tower’, I immediately went to this page

I’m starting to get back into comicbooks thanks largely to the reboot of the Valiant Universe. I have almost all the original Valiant books in boxes in the basement and I think the reason I was drawn to them was its tight continuity within the universe. Events in one book directly impacted, and did not contradict events in another.

The ‘big’ universes of Marvel and DC have long had issues with continuity due to their sheer size and age. I suspect managing this problem is part of the reason why DC rebooted their universe last year (dumb, dumb move if you ask me…) and Marvel is apparently thinking about it as well. That is not to say that its not possible to have working continuity in a universe that large, and with more than one title per character (Batman has 8 titles right now). My example here is the classic Knightfall crossover (which Batman Returns is loosely based). It was across all the then Batman books and, well, was excellent. And in tone for each of the books. There wasn’t much bleed over into the other books in the universe though… but oh, did Batman bleed.



(What I need right now is a word that means something like hindsight, but for recognizing the choices that might have led the the current now.)

While reading Cem’s The Oracle Problem and the Teaching of Software Testing it dawned on me that in my brain, ‘Continuity’ and ‘Consistency’ are synonyms.

Wiktionary provides a narrative device in episodic fiction where previous and/or future events in a story series are accounted for in present stories as one definition of continuity and Freedom from contradiction for continuity.

So how do I think Valiant did originally, and should now, keep their continuity in tact? Glad you asked! Or didn’t and I’ll say it anyways.

• Keep the number of titles manageable. And my manageable I mean in the area of 16. Now I have no idea the economics of the publishing world are like, but that could give a couple ‘mini worlds’ that could tightly collaborate sacrificing the whole. (Harbinger, Shadowman, X-O, A&A).
• Don’t let the groups of books dictate direction. Where the universe is heading should be a Publisher level decision. The v1 universe had that direction available for anyone to see in big-scope-terms in Rai 0 (best cover of all time)

That’s it. Only two things.

Tying this back to testing/agile, I am starting to notice [now that I am looking for it…] this sort of failings in teams. Teams trying to tackle waaay too much work in a given iteration and when features hit production they don’t work as well as they should have when get into their hands. Happens. All. The. Time. I’ve also seen when agile teams or test groups wag the product dog based on their own whims and biases. Management is supposed to do that. Not you.

(Unless the name card on your desk says ‘publisher’.)