In article <1999Dec3.104237.1@eisner>, Kilgallen@eisner.decus.org.nospam wrote: > In article <828522$bho$1@clnews.edf.fr>, "Pascal Obry" writes: > > > > Ted Dennison a �crit dans le message : > > 8260kf$aan$1@nnrp1.deja.com... > > >> There's an Ada/Linux success story on that site that's worth a read ( > >> http://greenlime.com/Ada-Malaysia/success_story_1.html ). The project > >> was the complete reimplementation of a system, and the realtive > >> statistics for the two systems are quite striking. > >> > >> Of course one could argue that a complete redesign in *any* langauge > >> might have achived similar benifits. > > > > I would not have assumed that for the "Down Time". One could expect to > > see more problems during the first months for the new version. That's a > > real nice feedback ! > > That success story looked _so_ good (in general) as to be questionable. > Has anyone obtained independent corroboration, other than reading this > web site ? > > Larry Kilgallen > The graphs showed the averaged values. Take support costs for instance, both cost for mix language application (version 2) and version 3 indicates the averaged costs taken into account of man-hours involved to pinpoint a software bugs that include rectifying the problem. Version 2 showed at high of RM7500 for the first month of implementation compare to version 3 of the first month of implementation. It's clearly that to debug version 2, will involve more developers than version 3 becuase of the mixed-language modules in version2. But the readers have to take the version 3, which is built on top of the stabilized version 2 into account. Here, it will be difficult to separate the actual cost of debugging a stable codes and unstable codes. We take into accounts of the jobs done onto reported bugs. The number of reported bugs in version 3 is actually almost equaled to version 2. this statistic was not shown. But bear in mind that the bugs in V3 are a mix of human-causes and algorithms. On down time, integration between debugged modules and other existing modules are relatively easier. In v2, debugging a single modules with bug(s) involved other modules dued to language interfacing. That means shutdown of other modules as well. Hence resulting the increased downtime for v2 compared to v3. We don't guarantee the accuracy of the graphs (v2 values are based on total amounts billed to our client. For that time, we did not intend to collect this kind of statistics) but we had done the best to take all values into accounts. We don't have any independent corroborations on this. Lexical has done her best in compiling the statistics for the sake of comparing Ada to other languages. The statistics were intended for internal studies only but we decided to share this with other Ada advocates and those who have interests in Ada. It is up to individual to interpret the message we are trying to deliver. But if one needs further explanation, they can always contact those persons listed at the end of the document. We will most certainly give our best to explain. Hope this give a statisfying feedback everyone. -- Adrian BY, Hoe -------------- Sent via Deja.com http://www.deja.com/ Before you buy.