What is your issue? Specifically what reports are skewed? What versions of Link and Dataview are you on? There were a lot of changes from 1.5.1.0 to 1.5.1.2. 1.5.1.0 had a number of issues. Everyone thinks they are on current software until they actually check and confirm. Dataview is not some mystical program. It takes live data from the SMDR of the switch and transforms it into a gui interface that is easy to read. All it is is a database of stored SMDR and other data from the switch that can be cross-referenced. If you think about this logically, either (1) the MCP is sending the wrong data, or (2) this data is not making it to the dataview machine due to a poor network, or (3) the dataview program itself is not referencing the proper tables within its database (this would clearly be a software issue).
(1) I could see a faulty MCP or a corrupt database doing this. That is very easily rectified.
(2) This is by far the most common issue. No data network is perfect. There are constant packets flowing for each disconnect and connected call. If even one is delayed too long or dropped on a poor network, you will have "stuck" calls. In an ideal world, you want dataview and link on a standalone PC on the same switch as the MCP. This is optimizing the possiblilities of lost data.
(3) Of course there are possible bugs as well, but this is easily proven too. You will need to compare the reports that you think are skewed to another report that proves it. Possible sources of these other reports are UCD reports straight from the MCP, SMDR reports, traffic reports, or even other reports within dataview. If you show a station port statistic report that can prove another report wrong, then there is certainly an issue.
Sorry, i tend to write books when i have a strong opinion. Long story short(too late for that), there is a scientific method that will resolve EVERY issue out there. The only unexplained is corruption, which does happen, but that can be cornered as well when other things are ruled out.
*sam sung* gets off his soap box*
