KDE Perceived Quality Survey (PESQUS)

Chad Robinson crj at lucubration.com
Tue Jan 18 13:37:28 GMT 2005

Human Aspects of Software Engineering wrote:
> We are conducting empirical research on the quality and interface aspects of 
> various open-source and proprietary software.
> ...

I also have a few comments about this survey. First, it's important that 
survey participants be party to the results, or at least get a quick view of 
what others are saying. You'll get a higher response rate if you promise to 
distribute your results on this mailing list, or provide an e-mail address box 
where a copy will be sent.

Your survey appears to focus on the differences between versions. However, the 
survey makes no attempt to investigate HOW the different versions were 
installed, for how LONG they were used, and to what PURPOSE.

Unless you're getting a single-digit response rate, I suggest that you add 
questions that will allow you to separate respondants by installation method, 
distribution, and usage level. It would also be a good idea to separate 
problems by type: caused by my own stupidity or lack of knowledge, caused by a 
legitimate bug, caused by my stupidity but it was really somebody else's fault 
because of poor documentation and/or user interface design, somehow it's 
Microsoft's fault, etc.

I installed KDE 3.0 with Mandrake, IIRC, and used it sporadically; I hadn't 
yet switched to Linux permanently. I had some problems, but they didn't affect 
my work. However, they could have been real problems if I had actually been 
using it full-time.

I installed KDE 3.1 manually from the latest CVS at the time. I had a few 
problems here too that DID affect my work, but I'm pretty sure they were my 
fault for missing a dependency somewhere. I did a lot of the build out of 
order, and that surely caused even more problems.

Then I went full-time in Linux, so I built 3.2 and 3.3 via the Gentoo ebuilds. 
I have had very few problems here. However, now that I'm full-time these have 
had much more opportunity to affect my work. Knotes had an annoying bug that 
caused loss of data until I softlinked its data file, and I lost some 
important notes from a customer meeting before I realized it was happening. I 
also had a few ARTSD problems during a presentation that caused a few pointed 
jokes. The overall effect on me was very high compared to earlier versions, 
but that's because I was actually USING it for something important at the time.

It's also important to differentiate between layers. The Knotes issue was with 
a KDE application, but not KDE itself. It's a nice little app, but I doubt it 
receives much code review by core KDE developers. On the other hand, ARTSD is 
a core layer, and I've had a LOT of problems with it. I hope I don't offend 
any ARTS developers but it's so irritating that I've set aside time next week 
to recompile KDE without ARTS support. I don't see how it adds any value - my 
applications work fine talking directly to ALSA, and I never get any stuck 
daemons consuming 99% of my CPU. Irritating? Yes. A real problem? No.

Finally, as others have pointed out, most of us have experience with other 
operating systems. Even if you don't plan to publish Windows or MacOS numbers, 
it would be a good idea, statistically speaking, to include those questions as 
well. Some users are more irritated by problems than others. Two users, one 
responding "low" and the other responding "high" could be seeing exactly the 
same issues with both platforms, but one is less tolerant of the problems. By 
including other operating systems you could normalize your data across the 
users' entire experience bases.


PS: What is the difference between "KDE froze during its operation" and "KDE 
caused a freeze"? These two questions seem redundant.
This message is from the kde mailing list.
Account management:  https://mail.kde.org/mailman/listinfo/kde.
Archives: http://lists.kde.org/.
More info: http://www.kde.org/faq.html.

More information about the kde mailing list