[z3t4]

The thing with software is that it takes two weeks to get to 90% finished, then two years to get from 90% to something 100% production ready. Having worked over 5 years with ecommerce software I could write a fully working e-commerce site in two weeks, but then the marketing teem want to have it all integrated with 50 different tracking systems, and support 10 different payment systems in 30 countries, and have the correct VAT handling as well comply to local laws, integrated shipping systems, and of course 30 different versions of the site in all languages. And integrated with business software. And every moth the marketing team want to start a new compaign, buy 3 and get 20% off, buy only milk and get 30% off, buy a red shirt and get 25% off, discount codes, discount codes that includes shipping, etc. Did you know that different products have different VAT in different countries and within the same country as well? And shipping should use the average VAT from the products ordered - depending on country...

[hnbad]

Sure but the article was about a linear process with the final product being understood to be an MVP, so the remaining 10% (which really feel more like Zeno's arrow, i.e. the sooner you get to 100% the closer the time necessary to complete the remaining work approaches infinity) won't be part of the calculation and the 90% solution is the first "definition of done".

The article clearly points out discount codes as an example for a feature that could be moved out of the initial spec. And the spec itself is clearly missing various features necessary for the app to be even remotely usable.

[rabuse]

This has been the case for almost every project I've worked on. It starts out as super simple, and then the edge cases roll in, and then the feature creep, and then the bugs start piling up, and now it's been years on a "just a couple months" estimation.