[EpiData-list] Almost releasable new pre-public version of Analysis (please test)

EpiData development and support epidata-list at lists.umanitoba.ca
Thu Jun 11 13:06:32 CDT 2020


Dear list-readers

A new updated version of EpiData Analysis (v1.7.99.1) has been placed 
for your testing on http://epidata.dk/testing.php

Please find below a summary of the changes. In particular notice that 
the requirement of [_n] in pgm's has been removed, unless you refer to a 
given observation. I expect the release as public in the coming week or 
the week after. Please test and report to this list.

As part of my job as a clinical epidemiologist and professor at the 
Odense University Hospital here in Denmark I have extensively tested the 
software during the last three months. We now use the pre-public version 
on a daily basis.

Now we have a running system in new EpiData Analysis. It is run by 
secretaries, who have no experience in programming. They just click an 
icon to accomplish:
a.import a csv file saved daily from our main IT-patient registration 
system (EpiData Analysis)
b. merge with previously controlled observations to evaluate only new 
observations (EpiData Analysis)
c. quality securing coding in the patient journal. (EpiData EntryClient)
d. create SPC graphs in classic analysis (EpiData Analysis Classic)
e. create lists for audit of patients where the quality standards were 
not met.

To run the system in the revised new analysis we only lack possibility 
of creating graphs, which I expect to include in the coming months, if 
all goes well ready by the end of this year.

One other aspect we need to tackle is the increasing demand for safety 
certificates and documentation in relation to secure systems. If any of 
you are experienced in this area please write me an e-mail (jl at 
epidata.dk) .

Without the efforts of Torsten Bonde Christiansen (Software design and 
programming), Jamie Hockin (testing, programming, Mac adaptation) and 
Hans L. Rieder (testing, approving solutions) the software would not 
continue to develop.

Please look into the website for inspiration: 
https://tbrieder.org/epidata/epidata.html

Best wishes

Jens Lauritsen, coordinator and initiator.
EpiData Association, Denmark


-------------------------------------------------------------------------------------------------------------------------------- 


Changes since v1.6.0.0

Major improvement:
- [_n] is no longer needed, except when desired.
- editor: Completely new. Many aspects improved. Works better in MacOs
- new !output option with save command - allows saving output content as 
html
- new "set" option to control format of saved output html|txt
- new "set" option allows user definition of css styles in saved html 
output
- new "set" option defines internal css or external reference to css 
file in html output
- extended fast createdate function allows conversion of date strings 
with named months.
- extended createdate function allows use of national month naming 
defined in set

Other new aspects:
- mergevar added as the last in a merged dataset
- default dataset used after merge is the merged one
- keyboard shortcuts unified and simplified.
- main window opens immediately if started with a pgm
- many commands: unified option naming, e.g. !l is labels

Bug fixes: - all solved
- keep : if used in a "for loop" errors could occur
- Assignment: incorrect control of missing in string variables
- validate duplicate : error if a variable was missing
- many minor other bugs or inconsistencies

Bug fixes: - partially solved
- errors in calculations when dividing by zero.
     fixed in Linux and MacOS, but not in windows always.

Known issues at this point:
- Analysis does NOT handle header "variables" at all (except keeping 
defined ones).
- Sometimes windows 10 gives a "list - index out of bounds" error
    in particular when running a selected block from the editor or when 
running several pgm files within a pgm.



More information about the EpiData-list mailing list