We are rapidly approaching the first pre-public test version of the new analysis, but before getting to that stage.
From today v0.9.2.0 is released at the bottom of : Http://www.epidata.dk/testing.php
Changes from last experimental version are:
* a unified and simplified option naming convention expected to be the final choice (many options are one or two letter only, not a long word, which may be misspelled) * You may set "mark for deletion" and "verified" at observation (record level) (and remove again) * Reorder ændrer ny på visuel struktur * reorder command will also change the sequence in a saved file * unification of the shortcuts and function. e.g. exit editor and exit browser * New reports: count by (variables, eg id visitdate) , validate double entry files) * The browser now shows observations marked for deletion or verification * new functions to summarize numerical data: sum(..) and strings: concat(..) * frequency command can show confidence intervals * common naming of result variables after running variables (try: list r) * renamed set options and added some for default browser, colours etc. (to see these, issue: set - or open editor and add set from edit menu) * commands to alter the structure of projects. e.g. "un-relate" a related project, append or merge whole projects or single datasets within a given project. * fix many minor bugs and error situations, which would give a strange error. * all errors are now shown regardless of setting of: set "echo" := "on/off"". They are only hidden if you issue the set "show error" := "off";
Notice that the commands.html and the "how to adapt your pgm" references are found via the help menu in either the main analysis form or the editor. But also that several browsers may be open .
Please test and report to this list or in case of severe problems send an email to info at epidata.dk
In particular test data management features and the ability of the software to read all files created (rec, dta, epx, csv) . Save them again after some data-management, e.g. summation or reordering, and see in Manager or EntryClient that it actually worked.
Please do remember to work on copies of data. Real data should only be used in production release versions for actual data-analysis or data management
Work now continues with the table command and other issues, e.g. ability to shift from txt to html output or simple encryption and decryption of projects.
Notice that without the efforts from assisting other members of the list we could not have made the software to the current level. In particular thanks to Hans Rieder and Jamie Hockin for a lot of structured testing and feed-back.
best wishes
Jens Lauritsen , Torsten Christiansen EpiData Association Denmark