Confessions of a Lapsed Developer

I think it happens more often to developers who become development managers, but that’s not the route that I took.

The world is suffering through the use of software designed or, to verbify a noun, ‘architected’ by former developers who ascended to the rank of software architect and decided coding was beneath them after that.  I didn’t go that way either, although I did occupy both of those roles, I continued to code while I held them.

No, my route to non-coding started when I decided that I had to be the one to try to make it possible for developers in a certain very large bank to practice Agile software development.  I spent four years in the corporate change group educating, explaining, presenting and otherwise promoting Agile methods as a valid and safe methods of developing not only software, but for performing any change.  My team and I made a lot of headway and now, Agile methods are considered an alternative in enlightened areas of the bank.  I feel that I was successful at that.

There’s another part to the confession.  While I was still coding and practicing Agile techniques, I was pretty much a purist in that I stood by the idea that simplicity should and would win over adopting frameworks like WPF, or WCF (I was primarily in the Microsoft world at the time).  I disdained data binding where custom controls hid all the ‘magic’ required to update a view.  I preferred MVP (Model-View-Presenter) later re-christened Passive View pattern, not only because of its inherent separation and ease of testing, but because updating the view manually allowed one to have literally Web, Windows or Command Line interfaces to the model and use the controller (or presenter).

These two elements of the confession mean that I probably have been outside of the mainstream of development for six to eight years – a lifetime.

I now find myself in a situation where understanding not only Javascript but AngularJS will be tremendously helpful and I am lost.  These new frameworks are full of magic, so I guess I’ll just have to get used to not knowing exactly what is going on.

I can still code in C#, and I can even push by in Python if I have to but these new frameworks in their new environments have me stumped – at least for the moment.

I have to relearn a lot of stuff, and re-orient myself to the new environments and I’m finding that summoning the energy necessary is extremely difficult.

If you have any shortcuts to knowledge and understanding for guys like me, I bet you could make a fortune.  I can’t be the only one who’s out here struggling..

If I figure something out, I’ll put it here.  I’m hoping you’ll do the same.

Fess up Apple Fanboys..

It’s a long story* but after many years of resisting the siren call of buying a Mac when all the cool kids were doing it, I have recently come into possession of a 2 year old Macbook Pro.

So here’s where I expect a confession from all the Apple fanboys.  I’ve been getting familiar with the Mac, adjusting preferences, installing Office:Mac, etc. and you know what happens?  Every time there’s a change to anything ‘important’ and even though I’m logged into an administrator account, I get prompted for my account password every time.

Weird.  Microsoft’s Windows Vista was raked over the coals for exactly the same behavior.  That was the beginning of 2007, and Windows 7 made that behavior a little more intelligent (Windows 8 makes the experience a little different but that’s another story).  This is OS X 10.6.8 which presumably is newer than that, but hey, I’m not a fanboy, so I can’t give chapter and verse.. and yeah, I meant that analogy.

Fess up fanboys; OS X isn’t that fantastic.  In the words of one of my old friends, “It’s just another operating system.”

———

* As a result of a relocation, we are moving out of our 3,000 sqft single family home into a 1,500 sqft apartment and the lawn tractor had to go.  A local craigslist listing was looking to barter a Macbook Pro for some lawn equipment and voila, I now have a much smaller package of equivalent value.

Two Sets of Books – Wrong Metaphor

You may have heard folks talk about having two sets of books; usually they say that conspiratorially and with a wink.

Why? Because when you have two sets of books, that means you are deceiving someone. You are literally telling one story to the authority (a false story) and another to yourself, or whomever you want to tell the truth to.

Gangsters have two sets of books. Long haul truckers who don’t like maximum drive hour restrictions (and violate them) have two sets of books.

You get my point. “Two sets of books” is synonymous with deception.

Why am I going on about this? I’ve heard several Agile advocates talk about this idea over the years. Early on, the idea was that management would be hostile to Agile and therefore, for the good of the company (and the team), we should go ahead and keep our Agile information to ourselves and tell management we’re doing something different.

I’ve never been comfortable with that approach.

One aspect of becoming Agile is being transparent, and two sets of books isn’t being transparent. I feel a lot better about telling it like it is, and educating folks about the data we have about progress that they never had before.

There’s a difference, though if one is talking about having a translation available to help guide folks to a new understanding, but that’s not having two sets of books. That’s having one book, with a translation.

So, for me at least, ‘two sets of books’ is a bad metaphor. Be transparent and say what you mean.

We’ve Heard This Before, but Where is the Value?

This morning I was fiddling around with TweetDeck (I’ve been neglecting my feeds) when I came across a tweet from Scott Ambler (@scottwambler) referencing a tweet from Tom Gilb (@imtomgilb), the guru of requirements, wherein Mr. Gilb exclaims:

Massive €142 billion IT failure. Agile did not solve it. Coders cannot solve it. Requirements cited. Wake up!

Here is the article, by-the-way: http://www.bcs.org/content/ConWebDoc/19584

The only trouble is that Agile is not even mentioned in the article, so the assertion that “Agile did not solve it” is disingenuous.

The article mentions that 7 of 10 projects were likely using Waterfall, based on some previous research, but there’s no mention of what the other three were – not even speculation.  The analysis goes on to explore multiple angles about where development efforts were stopped or not, leading to the conclusion:

Developing an alternative methodology for project management founded on a leadership, stakeholder and risk management should lead to a better understanding of the management issues that may contribute to the successful delivery of information systems projects.

Sounds like the way I describe Agile to folks.  No, Seriously.

But here’s the big miss as far as I’m concerned – and this goes for the various Standish CHAOS Reports that we hear about from time to time – the definition of failure is this false triangle:

…projects that do not meet the original time, cost and (quality) requirements criteria

Where is the value proposition? None of those three data points are important if the customer received enough value from the solution to make the investment worth it.

When will they add the question, “Did the resulting solution deliver enough value to warrant the extra investment in time/money?”

I’m not saying that any of those projects could be cast as successes if the value question were included in the analysis, but there might be some.  Do you think all those expensive and/or late projects were useless?

I doubt it.

Even leaving Agile out of it – although I think it would be hard to deliver software consistently in the following examples without using some elements of Agile – let’s look at some possible outcomes using strictly scope/cost/time:

• Your customer decides that you have delivered enough value after two thirds of the allotted time and don’t need to go further.  Failure (scope and time)
• Your customer realizes his/her initial idea had some flaws and changes direction to meet reality: Failure (scope)
• Your customer discovers that the competition has already delivered similar functionality and we need to add some new abilities: cost is not a consideration.  Failure (scope and cost)

I could go on, but you get the drift here: Scope aka requirements are the problem.  Or rather, signed-off, big-requirements-up-front contracts are the problem.

Yes, you can manage all this in a Waterfall environment, sort of.  But if you really could, you wouldn’t be using classic Waterfall, would you?

Agile methods provide an answer for these challenges because we concentrate on delivering value, close involvement with the stakeholders, and regular feedback at short intervals.

Value is the missing metric.

It’s Not That Simple..

In a recent post, Tim Ottinger & Jeff Langr gets to the inevitable snap-back point for all agilistas; “KNOCK IT OFF WITH THE TOOLS!” (I’m paraphrasing).

I understand the sentiment, but it’s not that simple.

In cases such as ours, we have heavy regulation, we have auditors, we have distributed employees all over the world.  Heck, it’s hard to find a US-based team that isn’t distributed.  I work from home myself, as do something like 18,000 of my co-workers.

We need tools.  Not one, but several to accommodate different technology platforms.  We need to have a centralized repository of stories and reports, like burn charts, velocity reports, test coverage, etc.  We need all those IDEs and refactoring add-ins and Group Chat clients and virtual pairing stations.  All that stuff.

I get the argument from some of the agile practitioners in our organization, “Look, open source projects have governance and distributed team members and all they use is e-mail and maybe Excel.”

Yes they do.

The key differentiator is that an open source team (even a foundation) is governing only one product, or one suite of products, not thousands or tens of thousands of applications.  Did I mention we have something like 100,000 technology staff?  I’m guessing we have 20-30 thousand developers.  Manage a portfolio like that with e-mail and Excel. 

Good luck.

There is no way that we can afford to send auditors to each team room wherever it might be; it’s just not realistic.  If we’re lucky (and sometimes, we are) we can get the whole project team together for initial story workshops and release planning.  The rest of the time it’s conference calls and LiveMeeting.

So, yeah, if you are working in a small company with a small team, or even a large company with a small suite of custom applications and a few teams of developers, use a whiteboard and a bunch of index cards.  It’s way more fun; I’d do it again in a second..

Reality for me and my co-workers is that we don’t have the luxury of that style of intimacy.

It’s just not that simple..

What is the Agile Sweet Spot?

An interesting question came to us from an executive sponsor of our initiative: “What is the sweet spot for Agile?”

My initial response was unspoken; “What do you mean, ‘sweet spot’?  Agile is best for all projects!”

Then, I thought better of that response.  He was, after all, an executive sponsor.  Too much passion may be passion, but it’s still too much.

So, I went away with this question in mind, and I’ve been thinking about it pretty much every day ever since.

I saw an article by Tom DeMarco in the IEEE Software magazine, entitled “Software Engineering: An Idea Whose Time Has Come and Gone?”.  In that article, he makes a great point: There are essentially two kinds of projects, ones that have significant value, and ones that have marginal value.

Tom makes the argument that marginal projects require greater control over the lifecycle in order to manage the expense versus the returned value.  Whereas, projects with a greater or more significant value relative to cost require less control.

This sets up a paradoxical situation.  Quoting Tom:

This leads us to the odd conclusion that strict control is something that matters a lot on relatively useless projects and much less on useful projects.

Nice!  That appealed to the elitist agilista in me.  But how was I supposed to go to the executive sponsor and say, “Well, Agile projects should be ‘restricted’ to those projects that matter most.  Use Waterfall for the projects that don’t matter.”?

No way.

But, what did dawn on me was that ‘projects that matter most’ is a good analog for strategic projects.  So what is the opposite of strategic..

I was watching some conference videos this evening, and one that showed on my radar was a talk by Martin Fowler in Australia at the Amplify ‘09 conference.

He makes an excellent case for strategic versus infrastructure.  And he gives the ratio of 80/20 or higher (i.e. 85/15 or 90/10) infrastructure to strategic.

Perfect.

That’s what I’ve been feeling, but didn’t have a way to say it.

Apply Agile to Strategic projects.  Apply Waterfall to Infrastructure projects.

I know.  You’re saying, based on my previous discussion above, that infrastructure doesn’t matter.  That’s not what I’m saying at all.

What  I am saying, echoing Martin’s presentation, is that infrastructure projects are more like commodities; there’s value, but it’s low margin.  And that’s okay.  In fact, we couldn’t do our business without it.

But, if you want bet your business projects to get the most value faster, you need to do it in an Agile way.

And it’s a lot easier to scale 20 percent, or less, of the project portfolio than a higher ratio.  And to get business buy-in and their dedicated time.  And justify cross-functional teams.  And justify co-location.  Etc.

I feel like I have something I can work with to answer the ‘Sweet Spot’ question.

Let me know if you have a counter to this, but I’m feeling pretty optimistic..

No, YOU don’t get the point..

In this post, Stephan Schmidt says Martin Fowler misses the point because Scrum is not about engineering practices (see here).

Actually, Stephan, you missed Martin’s point.

Martin’s post says essentially, that Scrum is gaining a reputation as an agile process that doesn’t work.  That is happening because the people who are adopting Scrum as their process believe that the management processes provided are enough and/or they aren’t aware of the engineering practices that are required to support it.

Professional developers do get it. They get that they need the discipline of the engineering practices.  They need that discipline regardless of the process by-the-way.

Unfortunately, the IT world is full of developers who don’t get it.  And their managers know they need to do something to try and change things.  Scrum has a corporately acceptable brand because it sounds like fast waterfall.  And, by-the-way, it doesn’t specify any engineering practices to help support the process.

Martin doesn’t miss the point.  That’s exactly his point.

Yes I Can Get Some Support..

The day after my last blog post about the standards group’s seemingly knee-jerk reaction to our CI choice, the group collectively got together and decided that encouraging CI was more important than the tool being used.

They got it.  All I had to do was point it out, and after some thought, they agreed.

I feel better about this process..

Can I Get Some Support Here?

I’ve been chasing my development comrades around for a year or so, convincing them that test-driven development is a good thing.  That continuous integration is a great way to show what  a great job we’re doing by being transparent – posting our success and coverage numbers for everyone to see.

All the developers I can get to try it come back saying, “You know, I wouldn’t have caught this bug before it went to production without TDD.”

Score!  A convert!

We set up a CI server running JetBrains’ TeamCity.  Dead easy and it’s been working great.  We have about four development teams using CI this way now – we’re still working on the testing – and even the configuration management group likes what they’re seeing.

So what’s the problem?

Memo from the standards group.  “Uh, we don’t have TeamCity on our list.  Your choices are, CruiseControl, Rational Buildforge, or Team Foundation Team Build.”

My response was, “Okay, but we’re in the middle of a transition here.  Do you suppose we could just go ahead and change over later (once the habit is good and engrained)?”

Their answer?  “We’re not comfortable with that”

I am getting a lot of support from the local management, so hopefully things will brighten in the next day or so.

Next.. Can I get you guys to estimate in some TDD time?

Donald Knuth Doesn’t Get It?

I was surprised to read excerpts from an interview that Andrew Binstock did with the pre-eminent Donald Knuth regarding aspects of software development methodologies and the state of programming generally.  The summary was provided here: http://www.artima.com/forums/flat.jsp?forum=276&thread=229705

The actual interview is here: http://www.informit.com/articles/article.aspx?p=1193856

What I found interesting is that, perhaps not surprisingly, the reviewer (Frank Sommers) doing the summary picks out supposedly controversial topics from the detailed interview leaving the impression that Dr. Knuth is out of touch with software development topics — at least from where I stand.

So I went and read the entire interview.  What I was left with was a brilliant, opinionated academic computer scientist who by his own admission is more interested in the science of computing than in software development per se.

He doesn’t elaborate on what in eXtreme Programming that he finds to be ‘exactly the wrong way to go’ which leads me to assume (perhaps incorrectly) that his objections are more due to the hype often associated with XP than with proper practise of it.  This article comes to mind: http://www.xprogramming.com/xpmag/jatBaseball.htm

There are a number of things on Dr. Knuth’s website that are interesting.  I’ll be playing with MMIX simulators well into the night..

Then I went to Dr. Knuth’s homepage and discovered his “Infrequently Asked Questions” page: http://www-cs-faculty.stanford.edu/~knuth/iaq.html.  Dr. Knuth is thinking about other things in a rather lucid way..

Thanks for looking in.