Starting From Scratch

I believe that everyone hates the feeling of starting from scratch, regardless of age or experience. This will cause people who should know better to make bad decisions, even misleading people who are conscious of the effect.

In my mental model of knowledge, I split skills between three major mastery levels:

• Those we can use professionally. I program professionally, and I practice daily to strive for mastery. I’m getting to the point where I can handle any task that is thrown at me.
• Those we can use. I can cook without killing myself of burning down my apartment complex, even though I can’t make many types of food and wouldn’t recognize a Mother Sauce if it was poured down my shirt.
• Those we couldn’t really use if our life depended on it. I’m not good at drawing. I’ve never spent time on it, and I have no motivation to learn. I can “draw” diagrams and get my point across, and this is enough for me.

All people treat the skills from the first category differently than the other two: we use them with confidence. Some people use them with arrogance. We know these fields well enough to absorb innovations and new ways of thinking. We can confidently discuss them with other people.

In my mental model, starting from scratch in our expert area causes people to reject a change. For instance: an Excel or Word guru would never willingly switch to a completely unrelated product. Relearning things line inserting line-breaks and adding pictures to a text document is completely orthogonal to anything they actually have to do on the job, much less something more complicated like Mail Merge. Instead of using it like it should be used, they would likely try using it like they used their old product, declare that it can’t ramrod the flimjam, and switch back to their old way of doing things, regardless of which is really better.

Interfaces and APIs

Altered interfaces are difficult for the end user. This isn’t just about GUIs or the placement of buttons in a menu; this extends to software APIs as well.

An altered menu causes all users to waste time looking for new ways to do something they could do 5 minutes ago. Changes to an API cause the programmers relearn a library they already knew — not to mention the laundry list of changes that need to happen to the new stuff!

APIs and Interfaces

The interface for a library or class develops massive unavoidable dependencies with use. I’ve called this “code inertia” in the past. Joel and Jeff’s recent discussion of dependencies between unit tests and code is a perfect example of code inertia: making nontrivial changes to an application can force a nontrivial number of changes to your automated testing environment.

This has not only a physical effect on the code, but also a mental effect for programmers! I’ve actually heard arguments from programmers based on the IDEA of possible future change: we shouldn’t use this library because it’s changed significantly in the past, and it might change again! Having to relearn the API combined with the frustration of updating broken code leaves a giant red flag in a programmer’s mind, even though any codebase that is being properly maintained will go through interface changes as new versions are released.

Excel

I was fortunate enough to take a Numerical Methods and Analysis class during my senior year of college.

If you’re not familiar, it’s the mathematics of computation: finding efficient approximations, determining rate of convergence, and knowing the margin of error at each stage.

This was my only college class where a calculator was ever useful. Most of our homework involved finding the values and error in iterative processes, so the professor suggested a “complex math package” for our homework: Excel.

Despite my initial skepticism, Excel is AWESOME with iterative numerical computation: type in a relative formula, highlight 50 rows, and click “fill down.” Voila! You can put a bunch of results in adjacent columns and compare convergence rates visually before you even prove that it converges at a certain rate. It’s quick, it’s interactive, and it supports easy experimentation.

Why is this relevant? My college upgraded to the Microsoft Office 2007 suite the previous summer, toolstrip and all. Every single class was punctuated by my professor loudly complaining that she HATED the new Excel because the interface was WORSE and she couldn’t find anything.

Even though she knew where everything was after 2 weeks of use, we got the same complaint every week. She had already made up her mind: it was worse.

KDE and Linus

A few years ago, Linus Torvalds publicly announced his Linux desktop preference on the GNOME message boards:

I personally just encourage people to switch to KDE.

This “users are idiots, and are confused by functionality” mentality of Gnome is a disease. If you think your users are idiots, only idiots will use it. I don’t use Gnome, because in striving to be simple, it has long since reached the point where it simply doesn’t do what I need it to do.

Please, just tell people to use KDE.

Linus

The different camps typed their representative emoticons — and D: — and then life continued as normal. However, Linus was eventually forced to switch when the next version of KDE changed too much:

Interviewer: Another open source project that underwent a big change was KDE with version 4.0. They released a lot of fundamental architectural changes with 4.0 and it received some negative reviews. As a KDE user how has this impacted you?

Linus: I used to be a KDE user. I thought KDE 4.0 was such a disaster I switched to GNOME. I hate the fact that my right button doesn’t do what I want it to do. But the whole “break everything” model is painful for users and they can choose to use something else.

I realise the reason for the 4.0 release, but I think they did it badly. They did so may[sic] changes it was a half-baked release. It may turn out to be the right decision in the end and I will re-try KDE, but I suspect I’m not the only person they lost.

I got the update through Fedora and there was a mismatch from KDE 3 to KDE 4.0. The desktop was not as functional and it was just a bad experience for me. I’ll revisit it when I reinstall the next machine which tends to be every six to eight months.

The GNOME people are talking about doing major surgery so it could also go the other way.

Linus realizes that it might be the right decision in the long run, but there’s no escaping the fact that even the big guns in the field can’t spend his time relearning the basics every day, even if it’s the right thing to do. He has a mental model of how the desktop is supposed to work, and when KDE broke it, he fixed the problem.

Adult Learning and Malcolm Knowles

To me, a lot of the principles of API and interface design fall under adult learning. To state the obvious, most of the indended clients for interfaces are adults. We’d like them to learn how to use our products with the minimal amount of fuss.

However, if your interface is exactly like every other interface, there would be no reason to use it. You’re trying to solve a problem for yourself, and you’re trying to solve a problem for your intended users. Your users will have to learn how to solve their problems with your interface, and it had better be easy.

Enter the research of Malcolm Knowles. He is one of the big-shots in the field of adult learning, having done a lot of the early theoretical work. This site has a good description of his life’s work with researching, for the interested.

In particular, Knowles attempted to find the theory of what caused adults to best learn. Believe it or not, there hadn’t been extensive research in the field before he started. The closest that he was able to come was his theory of informal learning, through group sessions or other kinds of informal meetings. He also developed his assumptions by which adults learn:

Malcolm Knowle’s Assumptions on Adult Learning

1. Self-concept: As a person matures his self concept moves from one of being a dependent personality toward one of being a self-directed human being.
2. Experience: As a person matures he accumulates a growing reservoir of experience that becomes an increasing resource for learning.
3. Readiness to learn. As a person matures his readiness to learn becomes oriented increasingly to the developmental tasks of his social roles.
4. Orientation to learning. As a person matures his time perspective changes from one of postponed application of knowledge to immediacy of application, and accordingly his orientation toward learning shifts from one of subject-centeredness to one of problem centredness.
5. Motivation to learn: As a person matures the motivation to learn is internal.

The salient bullet point is #2: experience. It’s directly listed as a basis for adult learning. I’m not saying that this is a direct proof of my mental model of knowledge, because quite honestly it’s just not a proper conclusion. However, designers of APIs and GUIs should take this as an indication that experience is something to build on, not disregard.

To use an aforementioned example, I really like the ribbon interface in Office 2007! I was proficient after only a few hours, and I enjoyed using fewer keystrokes to achieve my goals. I thought that it mapped pretty well to the old menu-based methods, and it was pretty obvious when it didn’t. If I could use only keyboard shortcuts to navigate, it would be the ideal control.

However, it’s probably a bad idea releasing it in the same year as Vista, which already scrambled some menus, changed some positions, altered some interfaces, renamed some popular folders, added UAC, etc. It all added up to too much of a change: when people sat down at it for the first time, nobody couldn’t use the darned thing.

Live and learn!

Footnote

I remember a minor fuss when Linux 2.6.0 came out because the device driver interface changed. Linus himself descended from the cloud to state that he makes the right design choices for Linux without regard to compatibility issues. It was going to tie in really nicely with my point, and my Google Fu has failed me. up. The device driver interface changed from 2.4 to 2.6, and that there was a of device driver complaints. At this point, I just want to know if I was making this up.

Popularity: 20% [?]