Intentional Technical Debt
This goes along with the entry I wrote called A Big Ball of Mud. Often when you work for someone else, even though you’re the pattern maker, you don’t get to decide how to do things. Some decisions are taken out of your hands for whatever reason. Here’s an example that annoyed me way back when.
This coat (#72558) was a unisex wool bomber type jacket. The point of contention being “unisex”. In my opinion, there is rarely such thing. What “unisex” must mean by default is “sized for men” but with the implication that “we’re optimizing our chances by saying women can wear it too”. That’s the reality. Were it not and sized for women, it wouldn’t be called unisex because few men could wear it. In the photo above, you can see the woman’s sleeves are much too long and the neck way too large. In this case, the technical debt of this jacket was incurred by a purchasing decision. In this plant, they only wanted to buy three sizes of separating zippers. No lie.
In this case, the big ball of mud -”a haphazardly structured, sprawling, sloppy, duct-tape-and-baling-wire, spaghetti-code jungle” were the grade rules, haphazardly cobbled together to fit purchasing specs. This made for all kinds of fun and games with grade rules. Normally length increases or decreases with sizes but with only three size zippers, the grade rules had to be designed with size breaks. XS/S took one size, M another and L/XL yet another. This rule library was called SMLZip. Jackets that had button or snap closures took another grade rule library, RulibSML. This meant that an otherwise identical jacket but with a button closure, would not be the same length for all sizes for the version that came with a zipper.
Complicating matters rule libraries were shared between men’s and women’s even in non-unisex goods. The work around -which actually worked pretty well as a strategy, the practical expression being the difficulty- was to name sizes (with numbers). Since they’d run out of size names long before I showed up, the RulibSML names worked like this:
- 92=women’s small
- 93=women’s medium
- 94=men’s small
- 95=women’s large
- 96=men’s medium
- 97=men’s large
- 98=women’s xlarge …and so on.
And yeah, I realize that women’s extra-large at the outside of the men’s large looks like a glaring problem. It did in real life too. Any pattern maker or grader worth their salt has gone into dry heaves at this point. How does a company decide to use the same grade rules between men and women is beyond me too. By the way, the company was sufficiently successful, they weren’t in debt or anything or bleeding money from every orifice. As there was no crisis, there was no sense of urgency to repair what probably would have been an undiagnosed problem.
Anyway, with the styles that took a zipper (remember management only wanted to buy three sizes) you had sizes grouped in three length size breaks too, exacerbating fitting problems from using the shared rule libraries. SMLZip worked like this:
- 92, 93, 94 =zip size short
- 95, 96 =zip size medium
- 97, 98, 99 =zip size long
Since the company had a Gerber CAD system, resolution for better fitting patterns wasn’t insurmountable. My suggestions would have been to separate men’s and women’s grade rules, developing separate rule libraries for each. Unlike with hand made patterns, the new grade rules could have been implemented with the click of a button, pulling down the new library substituting for the old one. It just seemed so pointless. I can only imagine this system was implemented way back when, when the person who was then in charge of patterns and didn’t know much (she didn’t, a lot of her work was liege and legacy around there) thought this would make a quick fix solution. And maybe that’s all it was ever intended to be. A quick fix solution in a novel situation became policy. And that brings me to that concept of intentional technical debt; “a conscious decision to optimize for the present rather than the future”. That’s the problem with short term fixes, they tend to become long term solutions. No one ever revisits the problem, it just gets buried with other “solutions” cobbled on top of those.
As far as a solution to the zipper quandary, with a closer analysis, there could have been some overlap in purchasing between sizes. Better still, they could have purchased by the yard and set stops and sliders too. This would have allowed a lot more variety in styling since fabrications were repeated across styles. Having zippers by the yard would mean there could be five different sizes for women’s short jackets, or long ones as well as those for men (although their jacket lengths didn’t vary). The end result being, the women’s line was being impinged by purchasing requirements for the men’s line. Who’s to say what the effects of these practices were?