Wrong data in Workload

From SuperMemopedia
Jump to navigation Jump to search

Question

christian wrote:

I recently bought SuperMemo 15.2 and already have some data that seems to not work correctly. In the attached screenshot you will see that on my system the workload window displays repetitions done on the 22nd of March, which is currently 5 days in the future. This is not particularly bothersome to me at the moment, but it does make me doubt and worry about the integrity of my learning data.

You can download my collection at this URL to reproduce the issue: http://dl.dropbox.com/u/10190786/systems.zip

Hints

  • this situation could happen if this collection was used on any computer with a system date wrongly set
  • the collection does not indeed look like used past Mar 17, 2012, however, it contains a huge languages branch that looks like imported from another collection. Some wrong repetition data might have been included in that imported collection

Followup

This collection never was on any computer but my own and i never touched the system clock. As well, my IRC logs indicate it never changed at any point.

More importantly, please *do* take a look at the linked collection. Do note how in "ABC of SuperMemo 15\history\2012\03 (Mrz)\" there are no files for the 22nd of March, yet the Workload window actually shows repetitions on that date.

This looks like a genuine bug.

Reply

This is the first report of repetition record showing at future dates. If this was a bug, you would need to perform some unusual operations that other users never explore, or never notice the irregularity at the beginner stage, or never report.

Followup

The languages branch was indeed imported, however not from another collection, but from a number of qa-text files, which can be found here: http://dl.dropbox.com/u/10190786/qa-files.zip

If any repetition data came from there, then the data might have had some very interesting faults, though i find that unlikely, based on how simple it is.

A question i find myself asking: If your investigation does not show any signs of usage beyond March 17, which exact part of the collection data does Supermemo interpret as usage on the 22nd?

As for why this was not reported: I may be in the somewhat unique situation of being a software developer myself and having decided on Supermemo after already being familiar with multiple repetition software products. My first step was to get my target data into the software, observing its interaction with the data and paying attention to how my interaction speed would result in reactions from the software, so i would not forge too far ahead, creating repetition loads impossible to fulfill. I kind of doubt that you get many beginners doing that and consequently paying close attention to the workload window.

--Mithaldu 11:05, 19 March 2012 (PDT)

Imported data did not contribute

The Q&A files listed in the imported zip file do not include repetition data. A bug in the import process is more likely than a bug in a learning procedure (imports are rare as opposed to daily learning), however, it should still be very hard to reproduce and rare as it has never been noticed before.