Not surprisingly, the various teams that contribute to the Windows product have been hard at work planning what goes into potential future versions of Windows.
As a part of that planning process, we've been collecting all sorts of data, and sifting through it to try to figure out what are the most compelling scenarios for the myriad of customers of Windows.
I was looking at the feature planning timeline and it had something like:
When I saw "T-Shirt Sizing", I got really excited. There's an old adage that says something like "A new product isn't done until the T-Shirts have been ordered", so I thought it was cool that the planning process had integrated ordering the group/feature T-Shirts right into the process.
Unfortunately that wasn't the case :(. Instead, what the people who designed the planning process meant by "T-Shirt Sizing" was categorizing the effort to implement the feature into "S, M, L, XL, XXL, etc". The basic idea was to get a rough feel for the amount of effort associated with each feature.
T-Shirt Sizing is an "interesting" metaphor for development effort. Normally metaphors speak to fundamental truths - "All the world's a stage, And all the men and women merely players, They have their exits and their entrances" is a metaphor that maps life to the actors in a play. But I'm really not that sure what T-Shirt Sizing has to do with the amount of effort involved in a software project. You might as well use "Ice Cream Scoops" for your metaphor - single scoop is a small amount of work, double scoop is more, triple scoop is still more, banana split is really big.
Personally if I were doing the planning, I'd not bother with the cute name and simply describe it for what it really is "Feature Size Guestimate". But that's just me, and I don't get to decide these things.
Obviously this is my opinion, the opinions of others may vary.
T-Shirt Sizing is not intuitive. They should have went with units of work measured in Volkswagens per Football Field. I.e., it's harder to stuff 2000 VWs in 12 fields than it is to fit 100 VWs in 1 field.
I'm taking bets on how long it will be until "t-shirt sizing" is verbified into "t-shirting".
"Have you t-shirted that feature yet?"
"We had to defer that, it t-shirted to XXL."
And then a few months later, Raymond will write a post titled "Microspeak: T-shirting"
There is another way to look at it: you should only wear a T-shirt that's suitably sized for you (unless you want to look ridiculous), so one way of stretching the metaphor to fit (:P) is to say, this feature should be done by John since it's big, this by Joe since its small... but of course that involves a lot of judgment and possible ego damage, so it's still impractical.
Nice observations, Larry.
IMNSHO, "T-shirting" is not actually much less repellent than the alternative use of the term "costing."
Both are verbified exercises in guestimation based upon instinct and experience -- I'm not sure the poetry of the metaphor actually matters much given how ultra-fuzzy (and extra-large) the final team estimate always is at the end of a long list of (T-Shirted) possible features.
The 'T-Shirt' metaphor does have one other thing going for it though: in terms of shelf-life, these estimates are about as valuable as a T-Shirt.
* * *
I don't hate this idea. At least it forces everyone into a constrained set of estimate sizes, similar to the agile planning game. Also there's no chance someone can wrongly infer some high level of precision later on in the lifecycle.
And what about confusion because of different T-shirt sizes in Europe and US?
The point of using T-shirt sizing (or any other arbitrary set of sizes) is to make it clear that what comes out of it isn't really a detailed costing that means anything.