Why do software projects take so long? At the agile lunch Austin this question was raised. Look at what stack overflow has accomplished in roughly 2 years with a similar team size.
(ugh this site's rich editor is brokey for my iPad.) How much does it cost to add another field to the system. Do IOC containers solve every problem? Do we overuse them? Is it really necessary to mock 20 things to test one little piece of code? Are we throwing tools to solve these problems rather than find better designs that have less friction?
One of the problem is under architecting, but over architecting is just as bad. How many people are involved in making decisions? For stackoverlow, there were many techical & architectural mistakes, but they shipped. . Customers (and managers) don't care about perfect architecture. As developers we do, because we want a good system that is easy to extend and that we have High confidence that it is correct. Where is the value?
Having access to the users is imperative to make sure you're building the right product.
Tom's Notes:
Does the complexity of a project meet the needs? Are the tools right for the job?
MS Dynamic Data - a solution? (folded into MVC)
Inhibitors
- Politics
- Over-Testing (ex: text domain logic, not implementation)
- Accessibility to the Customer/Business User/End User
- Purism
- Novice Developers
Quotes
- "Sucks" is relative.
- When adapting Agile, it's like changing oil... you first have to remove the dipsticks.
Senior developers shouldn't just give answers to junior devs but present opportunities to learn.
Takeaways:
- You and your customer should share the same values.
- Developers need to understand why they follow or use design/testing patterns.
- Where is the pit of success?
- Is the "benevolant dictator" the answer?
Comments (0)
You don't have permission to comment on this page.