This shows you the differences between two versions of the page.

Link to this comparison view

teaching:labs:xp:2004b:stories [2018/05/09 01:59] (current)
Line 1: Line 1:
 +^{{pic.underconstruction.gif}}|maybe this page is not needed anymore, as we describe our stories in [[extwiki>​|jira]]\\ otherwise this page should list the stories and link to the iterations where they where implemented \\  the brick-values may be interesting \\  clustering similar stories would give a pool of compareable stories| ​
 +===== This is in the moment a scratch page. For the real stories look at the [[iterations]] and in [[extwiki>​|Jira]]. Rather for customers private use. =====
 +=====  Meta =====
 +  * Clarify the notions and their relation: analyzer, detector, metric, range, threshold
 +  * is "​continous sniffing"​ a good name for what cultivate does
 +=====  Notes =====
 +  * Analyzers and Detectors as part of the project under consideration. Some as "​packages"​ - some as external libraries.
 +  * Scope of Detectors
 +    * Declare as prolog predicate (maybe completely hidden behind a user interface)
 +    * Per scope Detectors maybe enabled or disabled. For MetricAnalyzers there may be thresholds or ranges defined.
 +    * Check if a certain set of certain scopes is a partition of another destinct scope. ​
 +    * Attach to each scope (or to the combination of scope and detector) a PrologDoc which explains ​
 +  * Detect or check if an object is immutable.
 +  * tr: Detectors may be postponed if thay depend on a part of the source that contains errors.
 +  * Detector for SDP 
 +  * The abstractness metric should be improved. Maybe it has to be user defined. E.g. by providing a list of all packages in the order of abstractness.
 +  * Performance issues
 +    * tr: we shall use a query optimizer
 +    * dsp: for very hard metrics stochastic algorithms may be usefull. (Randomwalk on graphs shall find circles). Literature on that?
 +  * Far future
 +    * Refactoring to aspects and applying an aspect may be considered as invers to each other. You may decide based on a metric if you should refactor to an aspect or the other way round if you should apply the aspect and let him disapear in this way.
teaching/labs/xp/2004b/stories.txt · Last modified: 2018/05/09 01:59 (external edit)

SEWiki, © 2020