Computer Frustrations

<< Ideas

Neal:

This is a pretty demanding complaint but it bugs me when I have carefully decided on a course of action and I am asked to confirm it (YES! and I was sure the first time too!), but at the same time it annoys the hell out of me that sometimes the "quit" button is right next to some useful button and I accidentally click the wrong one and everything closes. That's when I wish I did have that extra confirmation step. Usually I give my computer the bird and say "you know that's not what I meant to do, I just opened that program, why would I want to close it 2.5 seconds later!"

Basically I'm asking for accurate mistake detection without lots of confirmation boxes.

Dan Klein:

I present my perspective as a dude working in industry.

Any instrumentation that doesn't let you look at the raw data or analysis software that doesn't let you see the raw output sucks really badly. The developers make good guesses about what most of the users will need to do most of the time. But sometimes they're wrong and sometimes the folks ( i.e., me and colleagues) using the instrument or the machine want to do cool new things. Or just look at the data in a more familiar or powerful environment.

A lot of the time, support for getting at the data is mostly there but still effectively useless. The raw outputs might be there but timing values might be built into the analysis software. Or all the data may be there but in an ass-backwards format. The files might be binary or just really ugly and hard to parse. This leads to frustration and not getting use out of the tools that cost us tens or hundreds of thousands of dollars.

On another note, it isn't easy and seamless to track changes/version a scientific analysis. I've used a really cool literate programming setup (R/Sweave, which lets me stick R code into a LaTeX file) but having some way to fork the analysis or see what I was doing a week ago or work collaboratively with it would be awesome.

Lisa:

- Indecipherable error messages
- When programs don't respond
- Virus scans popping up in the middle of doing something else 
- Losing internet connection/server failures
- When people don't put a salutation in their e-mails but just start with whatever they wanted to say
- When people don't sign their e-mails