Databases Reference
In-Depth Information
decision-help.
Asking for an opinion, e.g.
“Should a business object know about
its corresponding contract object”
.
conceptual.
Questions that are abstract and do not have a concrete use case, e.g.
“Concept of xml sitemaps”
.
review.
Questions that are either implicitly or explicitly asking for a code re-
view, e.g.
“Simple file download via HTTP - is this sufficient?”
.
non-functional.
Questions about non-functional requirements such as performance
or memory usage, e.g.
“Mac - Max Texture Size for compatibility?”
.
novice.
Often explicitly states that the person asking the question is a novice,
e.g.
“Oracle PL/SQL performance tuning crash course”
.
noise.
Questions not related to programming, e.g.
“Apple Developer Pro-
gram”
.
Most questions in the random sample fit into one of these categories, but for some
of the questions (9.61%), two categories were assigned. The most frequent type
of question (39.22%) was
how-to
, followed by questions about
discrepancies
and
environment
. The first two columns of Table
15.4
show the detailed results:
Answered
No
Code
Sum
Accepted Not accepted
answer
how-to
151
67 (44%)
63 (42%)
21 (14%)
discrepancy
50
27 (54%)
11 (22%)
12 (24%)
environment
40
13 (33%)
17 (43%)
10 (25%)
error
36
19 (53%)
14 (39%)
3 ( 8%)
decision help
22
9 (41%)
10 (45%)
3 (14%)
conceptual
18
10 (56%)
7 (39%)
1 ( 6%)
how-to/novice
16
10 (63%)
3 (19%)
3 (19%)
review
13
12 (92%)
1 ( 8%)
0 ( 0%)
non-functional
10
6 (60%)
1 (10%)
3 (30%)
novice
5
2 (40%)
3 (60%)
0 ( 0%)
other
24
10 (42%)
11 (46%)
3 (13%)
sum
385
185 (48%)
141 (37%)
59 (15%)
Table 15.4: Question coding
15.4.2.2 Which Questions Are Answered and Which Are Not
Figure
15.2
shows the distribution of answers per question. The number of answers
per question is shown on the x-axis, and the number of questions with that number
of answers is shown on the y-axis using a log scale. 5,450 (14.19%) questions were
not answered. The remaining questions had at least one and up to 23 answers. Only
3,243 out of 68,467 answers (4.74%) were provided by the same person that had
asked the question.
Search WWH ::
Custom Search