An Experiment in Scotch

Ken argues that Agile (big A or lit­tle, your choice) is Gestalt. From this con­clu­sion, he says that it’s a mis­take to dog­mat­i­cally fol­low a given process or pro­scribe par­tic­u­lar tools when we’re try­ing to imple­ment Agile and that instead, we should “…help remove orga­ni­za­tional and soci­o­log­i­cal blocks that pre­vent teams from employ­ing them.” While I think he may be right that Agile done cor­rectly is Gestalt, I don’t think that his con­clu­sion nat­u­rally follows.

For those of us who haven’t been in cog­ni­tive psych too recently, the Gestalt basi­cally boils down to “the whole is greater than the sum of its parts” and is holis­tic in nature. Max Wetheimer, com­monly con­sid­ered one of the three founders of the Gestalt school, said this about Gestalt:

I stand at the win­dow and see a house, trees, sky. Now on the­o­ret­i­cal grounds I could try to count and say: “here they are…327 bright­nesses and hues.” Do I have “327”? No, I see sky, house, trees; and no one can really have these “327” as such. Fur­ther­more, if in this strange cal­cu­la­tion the house should have, say 120 and the trees 90 and the sky 117, I have in any event this com­bi­na­tion, this seg­re­ga­tion, and not, say 127 and 100 and 100; or 150 and 177. I see it in this par­tic­u­lar com­bi­na­tion, this par­tic­u­lar seg­re­ga­tion; and the sort of com­bi­na­tion or seg­re­ga­tion in which I see it is not sim­ply up to my choice: it is almost impos­si­ble for me to see it in any desired com­bi­na­tion that I may hap­pen to choose. When I suc­ceed in see­ing some unusual com­bi­na­tion, what a strange process it is. What sur­prise results, when, after look­ing at it a long time, after many attempts, I discover-under the influ­ence of a very unre­al­is­tic set-that over there parts of the win­dow frame make an N with a smooth branch…

(Empha­sis mine)

So is Agile greater than the sum of its parts? I agree that it prob­a­bly is. But here’s the kick in the pants: so is water­fall or RUP or what­ever method­ol­ogy of the week that you fol­low to write good soft­ware. The key word there is “parts” I think. As in, with­out ALL the parts, you don’t get the emer­gent fac­tor of Gestalt. Water­fall can be Gestalt, just ask the peo­ple who write the soft­ware that runs the space shut­tle. It becomes a mat­ter of dis­ci­pline in fol­low­ing a process precisely.

In my lim­ited expe­ri­ence, what I find hap­pens if you don’t pre­scribe a prac­tice in a shop that’s try­ing to imple­ment agile is that shop picks and chooses the pieces that it likes (typ­i­cally short sprints and plan­ning at the begin­ning of each sprint) and leaves aside all the pieces that they don’t like which are typ­i­cally the parts more likely to engen­der suc­cess with Agile like pro­duc­ing deploy­able code after each sprint or hav­ing access to real users, not man­age­r­ial standins. When this process fails to pro­duce the expected results, it’s the methodologie’s fault even though in real­ity, the method­ol­ogy was hardly imple­mented at all. There’s a rea­son why we call them method­olo­gies and not sug­gestolo­gies. There is a method involved and when you short cut it, you get short cut results.

I don’t think Ken Schwaber would ever sug­gest that leav­ing out pieces of Scrum would result in a bet­ter process. In fact, his book is full of case stud­ies where teams were try­ing to imple­ment Scrum by skip­ping impor­tant parts of the process.

Eric Gun­ner­son wrote about what he called “scrum­but”, that is, the prac­tice of say­ing you are doing Scrum but you’re doing it dif­fer­ently. I’ve writ­ten in the past how Scrum is anal­o­gous with a jazz musi­cian. When you are a begin­ner, you don’t have any busi­ness going off on your own and riff­ing some jazz chords because you don’t have the fun­da­men­tals nec­es­sary to do that. Scrum and Agile are the same in that as a begin­ner, you can’t decide what pieces of the method­ol­ogy are good and bad for your team because you don’t have the fun­da­men­tals nec­es­sary to make that deci­sion with­out hav­ing imple­mented the process pre­cisely. I don’t think a men­tor should ever advo­cate or facil­i­tate such behav­ior either. Instead, he should explain why all of the pieces of a par­tic­u­lar method­ol­ogy are necessary.

A process with­out ALL of its most impor­tant parts becomes anti-Gestalt because then the process not only doesn’t pro­duce the emer­gent behav­ior expected, it can also be blamed for the fail­ures that inevitably fol­low. Com­par­ing the pic­ture in Ken’s post with the pic­ture above, it becomes obvi­ous that with­out the crit­i­cal parts, the cross in the mid­dle never emerges.

Agile has the poten­tial to make writ­ing soft­ware more suc­cess­ful. But that doesn’t mean that it will ever be easy or that you can skip parts you don’t like. Imple­ment­ing agile cor­rectly involves pre­scrib­ing and then fol­low­ing an often­times dif­fi­cult process. Only by doing that can agile become Gestalt.