What is the nature of usability discussions in OSS projects?
Is it different from what might be expected from the textbooks on how to do interface design?
Is it different from commercial software design?
What are the patterns of discourse and process that emerge within and across projects?
The study in [1] is exploring the bug databases for terms such as ‘usability’, ‘human computer interaction’, ‘interface’ etc. Author assumed that "bug reports that contained those words and were determined to be indeed about usability" and is to be investigated.
The bug report mechanism such as Bugzilla studied in [1] has shown a number of drawbacks:
"sometimes even a screenshot is not sufficient to uniquely identify the problem."
"...bug reports can reveal information about the reporter and the reporter here has clearly gone to significant trouble to obscure the text whilst still reporting the bug."
Rewording text elements of interfaces occurs relatively frequently:
"-Many usability problems can be addressed quickly and cheaply by rewording."
"-In our experience of teaching usability, it is a context where relative usability novices can play a useful role, serving an apprenticeship before moving to more complex problems."
"-Talking about the wording of interface elements is much easier to do in a mostly text based interaction medium such as Bugzilla than talking about graphical elements or interaction processes."
While fixing bug according to the bug report, the developer must avoid “ripple effect”. Ripple effect of bug fixing is demonstrated with the example of “dialog box” problem:
“In fixing this bug, it creates or accentuates other bugs; dialog boxes whose information no longer fits within the pane. Resizable dialog boxes had been used as a workaround for this problem, although one that various commentators to bug A saw as rather clumsy. The consequence was that fixing one bug created the need to fix other bugs.”
“Subjective usability bugs may need a more provisional approval process, while more evidence is collected of the relative incidence and severity of the bug….a duplicate identification tool would be a valuable addition to OSS projects.”
Bug reporting and classification are important to the OSS development:
Reporting tools that automatically provide contextual metadata further reduce the effort required by bug reporters. Subjective usability bugs may need a more provisional approval process, while more evidence is collected of the relative incidence and severity of the bug. Use of suitable keywords could distinguish provisional subjective bugs from more objective established bugs. That would enable investigation to continue without adding undue complexity to the system, and avoid premature discarding of a partial bug report.
Our analysis supports this suggestion and we note that such a tool's effectiveness is partially based on bug metadata. Tools such as GNOME Bug Buddy and the Bugzilla Helper promote structured textual reports but the clarification dialogues shown in Figures 1 and 3, and in numerous bug discussions, show that metadata is more valuable. Bug metadata more directly supports querying and partitioning of the bug reports, which should help to reduce duplicates and parallel bug discussions. Classifying any bugs is a complex process
However, classifying bugs (both usability bugs and functionality bugs) is a complex process:
One approach to classifying usability bugs may be to use the structure of the user interface itself as a hierarchical classification system. That is, the menus, sub-menus and dialog boxes of the interface become nodes in the classification hierarchy of the bug repository, so that a preferences bug can be located directly from the system's interface.
Bug report management to improve usability is a difficult task. [1] has described the research usability problem with specific key words in the database, but people, especially the non-technical people, can use different wording to describe the same problem without mentioning ‘usability’, ‘human computer interaction’, and interface’.
[2] is titled "Silver bullet or fool's gold: supporting usability in open source software development". [2] is a abstract by Twidale, who is one of the author of [1]. This abstract is about Twidale’s talk on the problem of creating usable interfaces for non-technical end-users. In this talk. Twidale is suggesting a coordination of end-users and developers in order to improve the OSS usability.
REFERENCE
[1] http://dx.doi.org/10.1109/HICSS.2005.266
[2] http://doi.acm.org/10.1145/1062455.1062468
No comments:
Post a Comment