XP and the Dot-com Boom

At around the time that XP was being introduced (1999 and early 2000), the much-derided dot-com boom was in full frenzy. It s just possible that this was a major factor in XP s sudden rise in popularity. At the time, investment money flowed like water, straight through the fingers of overconfident CTOs wielding outrageous business plans in the faces of starry -eyed venture capitalists. Stressed project managers, late-night hacking sessions, stale pizzas, and bug-infested source code became a cultural de jure, reflected by the emergence of Web sites such as NetSlaves ( http://www.netslaves.com ) in which self-proclaimed IT-workhouse victims described how they were regularly forced to work all-nighters and weekends to hit some earth-shattering deadline that could make or break the company. Though the staff generally hated this unsustainable pace, the prevailing attitude (however erroneous) was that working late was a necessity in order to hit the insane deadlines.

Soon, of course, the boom swiftly turned to gloom, and the cloud of insanity brought on by the promise of a new tech horizon began to dissipate. People had begun to notice that most, in fact almost all, of these projects had resulted in break rather than make for the start-up companies involved. The dot-com generation had engineered a bad name for itself: Dot-commer, thy name be cowboy. ^[5]

Most of these failures were due to either the lame, ill-conceived, and badly thought-out nature of the aforementioned business plans ^[6] or an inadequate software development process (and a butt-kissing staff that should have known better). Usually, there was a combination ”a deadly cocktail of problems. Tech start-ups began with a high-level vision of what their product or Web site should do, but with very little in the way of concrete requirements. The requirements (i.e., what the product should actually do) would be fleshed out as the team went along. The phrase Internet time was invented to justify the unrealistic deadlines and ever-changing corporate strategies. Internet time was basically 7 to 1 (i.e., the same as dog years ). Not coincidentally, most of the companies died like dogs.

Internet Years

Do XPers really use Internet time to justify not having a clue? In response to an online article about controlling requirements churn, XP author Ron Jeffries posted this message:

GROUCHO  

Have internet years been invented where you are? People want to change their minds. People want to start projects before they know everything about what they want. People want to say ˜I ll know it when I see it . . . ^[7]

It is, of course, useful to be able to move forward with only partial requirements. Sometimes this is unavoidable. However, there s a difference between this and positively encouraging a customer to move forward with only partial requirements, when doing so is avoidable, when it is possible to gather and prioritize all the requirements first.

Embracing change is a key aspect of XP that made it such a good fit for dot-com culture.

 

It was obvious that traditional software development life cycles couldn t cope with such a volatile way of working. The industry, having regressed into a sort of facile, fickle-minded juvenile state, wasn t about to mature and sort out its priorities ”so in the meantime something more agile was needed.

XP emerged at around the same time. Its origins weren t in the dot-com arena, but in a failed payroll system for Chrysler Corporation.

As we explored in Chapter 2, the C3 project was cancelled acrimoniously in February 2000, and later the same year the first XP book, Extreme Programming Explained , was published. (Hey, why stop the presses just because the project that begat the book contracts failed dismally?)

 

However, XP s appearance and Extremo message were perfectly timed with the cultural forces of the dot-com boom ”enough to capture the industry s imagination and gain XP more mind share than it really deserved. Many book deals, a surprise RUP tie-in, and lots of dedicated conventions later, XP continues to gain momentum.

Free Pizza!

This story was told to me by a programmer friend who did some work at LastMinute.com (a successful UK-based dot-com start-up specializing in cut-price vacations ). The company wasn t doing XP (it s painfully obvious that it wasn t following the sustainable pace and planning game practices), but this story does give some insight into the dot-com mentality :

When I was at LastMinute, the programmers there bragged about how they had all lost their girlfriends in the first month, and how they programmed into the nights, because they could all get free pizzas. I left in under a month, as they were simply not going to go live in a month, like they believed, like they had promised . They refused to simply stop programming, redefine the schedule with some reality in mind, and work out why they had so many bugs . They eventually went live eight months later.

 

As a software process, XP was remarkably well suited to the dot-com world. In fact, many of the XP success stories involve small-scale Web projects. Many of the dot-com projects (disasters and success stories alike) involved small tech start-up companies with development teams of 5 to 15 programmers and testers, and vague ideas of what they wanted to produce, just that whatever it was should involve aWeb site, should provide a service (e.g., Pets By Mail), and should be ready in Internet time. First to market with BrusselsSproutsHomeDelivery.com was all that mattered, regardless of how many people really wanted to buy their sprouts online.

As a culture, XP could have been tailor-made for generation dot-com. Foosball-playing and Nerf gun “toting cowboy programmers love the whole idea of oral documentation, constant snack food and, of course, knocking off at 5:00 PM every day (no more unsustainable pace ”besides, working til 3:00 AM sucked anyway!). The idea of constant pair programming probably also appeals, young socialites that they are. But especially no up-front design and no documentation. Absolute nirvana for the cowboy coder . Though XP wasn t dependent on the dot-com boom (XP has continued to grow after the crash), both the timing and the way in which dot-com mentality quickly spread to the rest of the computing world were perfect to accelerate XP s acceptance.

In fact, the whole people process aspect of XP virtually guaranteed its acceptance in the dot-com culture that still lingers uncomfortably in the post “dot-com world.

^[5] Of course, XP does warn against the cowboy coder ”for example, see http://c2.com/cgi/wiki?CowboyCoders ”but this doesn t stop such programmers from being attracted to XP, for the reasons that we state.

^[6] Which, astoundingly enough, begat lame, ill-conceived, and badly-thought-out software designs, which begat a process ”hey wait . . .

^[7] Ron Jeffries posting to the Software Reality Web site, http://www.softwarereality.com/lifecycle/tenrules.jsp, August 6, 2001.