Timeline for Empirical evidence for choice of programming paradigm to address a problem
Current License: CC BY-SA 3.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 3, 2012 at 3:08 | comment | added | Steven A. Lowe | @Cris: in this case the answer is (b). See my comment on Graham Lee's answer. And don't overgeneralize this; there's still a lot of room for science in software engineering, but experiments involving people with specific skills learning to do things are extraordinarily difficult (i.e. impossible) to repeat objectively. If you'd like to continue this discussion, feel free to email me. | |
| Sep 1, 2012 at 2:09 | comment | added | OpenAIDealAccountDeleted | @Steven: I'm not sure if you saying (a) you've surveyed all current empirical research and found it wanting, or (b) you know, a priori and in principle, that SE isn't tractable to scientific techniques. | |
| Sep 1, 2012 at 0:52 | comment | added | Steven A. Lowe | @Cris: I have no issue with using the results of scientific research. I have an issue with calling unrepeatable empirical studies 'scientific research' - opinion surveys, anecdotal investigations, yes, but not definitive science. Use with caution. Trust your own judgement. | |
| Sep 1, 2012 at 0:14 | comment | added | OpenAIDealAccountDeleted | @Steven: OK, we just disagree on this. Using the results of scientific research is very precisely the opposite of arguing from authority! This argument's been done to death in EBP and other areas, where it's now very largely accepted that 'expert judgement' (see Philip Tetlock) is deeply suspect when unsupported by sustained and disciplined use of empirical evidence. | |
| Aug 31, 2012 at 19:44 | comment | added | Steven A. Lowe | @Cris: For the record, no, I do not believe that there can ever be a truly definitive study in this area; see addendum. The idea that a 'definitive' study can be used instead of expert judgement in order to make a critical architectural decision is 'appeal to authority', which is a form of logical fallacy :) In my experience - and I'm not making blanket accusations, just an observation - the quest for such things is most often either an attempt to avoid responsibility for a decision, or an attempt to support a decision that has already been made. | |
| Aug 31, 2012 at 19:40 | history | edited | Steven A. Lowe | CC BY-SA 3.0 |
addendum
|
| Aug 31, 2012 at 17:22 | comment | added | Philip Oakley | The quality of a software method is in how well it matches the local human need. In general, it is not constrained by the laws of physics (Scotty). It will be a long while before 'software' [discipline] manages to distil its immutable fundamental laws. For example see "Software Quality Metrics: Three Harmful Metrics and Two Helpful Metrics" in ppi-int.com/newsletter/SyEN-046.php#feature and ppi-int.com/newsletter/SyEN-047.php#feature | |
| Aug 29, 2012 at 21:27 | comment | added | OpenAIDealAccountDeleted | @Steven: the word for not trusting the results of a truly definitive study is 'superstition'. Perhaps what you really mean is you don't believe there can ever be definitive studies in SE (which statement would itself obviously require a large, well-supported body of evidence). | |
| Aug 29, 2012 at 13:24 | comment | added | jk. | If there was a definitive study with a reasonable experimental method the results of the study might be interesting; as it stands this answers seems to state that any study is worthless irregardless of method which is a little too non-scientific for my liking | |
| Aug 29, 2012 at 13:16 | comment | added | Steven A. Lowe | @GrahamLee: nothing is outside the purview of the architect when questing for solutions. Just don't let academic posturing in a vacuum override the architect's good sense! ;) | |
| Aug 29, 2012 at 13:15 | comment | added | user4051 | So is it outside the purview of an architect to look for prior work on which to build his or her opinion? Probably not - hence the question of whether and where such results can be found. | |
| Aug 29, 2012 at 13:14 | history | answered | Steven A. Lowe | CC BY-SA 3.0 |