A Canadian researcher has drafted a paper about five areas where open source software often falls flat on its face - user-interface design, documentation, feature-centric development, programming for the self and religious blindness.
Writing in First Monday, a peer-reviewed journal on the internet which is devoted to the net, Michell Levesque, a researcher for the Citizen Lab at the Munk Centre for International Studies in Canada, said open source software would continue to remain an unknown quantity to most computer users until these problems were addressed.
Levesque said she had been offered a number of possible reasons for the lack of intuition in user interfaces: the fact that geeks gave more weightage to integrity rather than beauty, the gender gap within open source communities, the fact that the interfaces were intuitive enough for the programmers themselves, the belief that user interface design wasn't real work, and a belief that a user interface could be developed once the actual program itself was done.
"What the user will see - and what they'll judge the project based on - is the user interface. If it's inadequate, no one outside of other geeks will touch the program," she wrote.
Levesque, a student at Toronto University's Computer Science department, whose job includes designing and implementing programs to enumerate and circumvent state-imposed net censorship, said open source projects had a big problem with documentation - "if they provide any documentation at all."
"Because they don't have a contractual responsibility to provide this documentation, it's usually intended to be a general guide rather than a complete manual that you could hand to a novice. Imagine what the following sentence looks like to someone who knows little about Unix and is installing it for the first time: 'You will need a list of MD5 checksums for the binary files. If you have the md5sum program, you can ensure that your files are not corrupt by running md5sum -v -c md5sum.txt'," she wrote.
Feature creep was another area in which open source software tended to stumble. Levesque said the fundamental rule of Unix - that a program should do one job well and interface cleanly with other tools - was often forgotten by open source programmers who tended to add features to programs and forget fundamental aspects.
Another area of weakness was that open source programmers tended to avoid simplifying things which they themselves found intuitive. "A very common problem among software developers (and not just ones working on open source projects) is the fallacy that intuitiveness is problem-specific rather than audience-specific: what is easy to them will naturally be easy to everyone else," Levesque wrote.
And finally open source programmers tended to be dismissive of anything and everything developed in the world of proprietary software, forgetting that some things in proprietary software were well worth emulating.
"There are still many things which software available for the Windows operating system does better than any present Unix-based system. Rather than admit that Windows is ahead in some areas, the tendency is to just ignore those particular areas.
"Likewise, since Apple has had so much success with Mac OS X, the open source community should be investing significant amounts of time and effort into cloning Apple's good work, rather than insisting that Gnome and KDE are just as useful," she wrote.
Levesque concluded that resolving some of these issues would require a change of mindset; resolving others would require more work. However, the net result would be quality software.