Capture Proposals Experiment

Keywords: COTS reuse time estimation

2005-12-21T16:49:00.000+09:00

001 L0020:/*+
@proposal COTS (Code Of The Shelf) LOC take about 2-6 times as
long to code as normal calculation statements.
@evidence - the time per LOC in the coding phase was double when code
contained several COTS modules that I was not familar with. Finding
and understanding how to use COTS modules increased the time needed
to add three or four method calls several fold.
@keywords COTS reuse time+estimation */

Keywords: inspection code parameters calculation

2005-12-21T16:20:00.006+09:00

001 L1047:/*+ @proposal check parameters of each function call when creating code and doing an inspection, check each parameter looks logicaly correct.
@evidence one of the parameters was wrong. The mistake could have been detected just by looking at each parameter to check if it was roughly correct
@keywords inspection code parameters calculation */

Keywords: design reuse modify

2005-12-21T16:20:00.005+09:00

001 L0677:/*+ @preferred design inserted add new proposal style tag to list of valid tags
@proposal read the design comment in the existing code before making changes to the design
@evidence forgot what some aspects of the original code did and missed changing a section of the design
@keywords design reuse modify */

Keywords: variable names design code

2005-12-21T16:20:00.004+09:00

001 L0081:/*+
@proposal name vars etc using a two stage process, first rough description then pick name
@evidence the name of the class was picked but did not work well, later a better name was identified, this suggests that having two stage process for nameing vars would produce less bad names and since the design needs doing anyway the overhead is small.
@keywords variable names design code */

Keywords: prototype interface

2005-12-21T16:20:00.003+09:00

001 L0047:/*+ @proposal add user interface design phase to ensure time spent working on interface
is consistent with what is needed.
@evidence design time and design review time increased by about 30% because I was
checking if the interface was workable, or if a different approach is needed. Splitting
this out for more complex user interfaces should allow a more appropriate allocation of time,
equaly it probably needs it's own check list.
@keywords prototype interface */

Keywords: COTS reuse time estimation

2005-12-21T16:20:00.002+09:00

Keywords: LOC time size scale estimation

2005-12-21T16:20:00.001+09:00

001 L0015:/*+ @proposal when LOC is about 500, the coding stage takes about 25% longer than when the LOC is about 250, bigger programs take more time per LOC
@evidence - the coding phase of the 250 LOC programs requireded less searching around to sort out integration of methods and class vars than with the 500 LOC programs.
@keywords LOC time size scale estimation
*/

Keywords: requirements design time estimate

2005-12-21T16:20:00.000+09:00

001 L0011:/*+ @proposal when the desired output is vague or untested the amount of design time is about 100% more than when the desired output was well defined.
@evidence - the design time for the untested and vague DefectHistroy.java and this program was about twice that of the fairly well defined PSP excercises
@keywords requirements design time estimate */

Keywords: size time estimate

2005-12-15T11:05:00.000+09:00

/*+
@proposal when LOC is about 500, the coding stage
takes about 25% longer that when the LOC is about 250,
bigger programs take more time per LOC

@evidance - the coding phase of the 250 LOC programs
requireded less searching around to sort out
integration of methods and class vars than with the
500 LOC programs.

@keywords size time estimate
*/

vague requirements time estimation prototype output

2005-12-15T10:21:00.000+09:00

/*+
@proposal when the desired output is vague or untested
the amount of design time is about 100% more than when
the desired output was well defined.

@evidance - the design time for the untested and vague
DefectHistroy.java and this program was about twice
that of the fairly well defined PSP excercises

@keywords vague requirements time estimation prototype
*/

___________________________________________________________
Yahoo! Messenger - NEW crystal clear PC to PC calling worldwide with voicemail http://uk.messenger.yahoo.com