[EpiData-list] new test release of EpiData Manager available (paste text as fields and read qes)

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Thu Jul 1 22:49:55 CDT 2010


I agree with this approach Jens.

I rarely use the EpiData editor as most others let you keep multiple files open without locking them. It might be good to have the ability to specify our favourite editor in preferences/options and have epidata launch that application when we want to edit or look at the output of codebook, etc. This could default to Notepad in Windows, something else in Linux and (maybe) Textedit in Mac. There seems to be no builtin plain text editor for the Mac - Textedit assumes .rtf unless you tell it otherwise.

Double entry is good, but I would not bother with the form of double entry where differences are pointed out as you do the entry. Just allow comparison of two files, as you suggest.

Agree that Manager should do the data verification after entry (will this include transformation of the data as is now possible?). Keep the Entry client to doing entry with checks.

Jamie

On 2010-07-01, Jens wrote:

> One of the aspects that I currently consider not needed is the development of an editor.
> Power users will each have their own preference for an editor and for the EpiData Manager we can optimise usage of the clipboard instead of developing a new editor.
> 
> Another issue is the need for documentation tools. My current understanding is that we do need:
> - double entry verification (comparing two files and showing differences)
> - count by id function (show a list of how many times a given id is contained in a number of files)
> - data verification according to controls built into a given data file.
>     e.g. whether all mustenter, range and value label definitions are met
> - verification of keys (unique) and listing of observations not meeting the defined key
> 
> And I also think these should be part of the Manager, whereas other functions such as:
>  codebook, listing and viewing data
> should be part of analysis.



More information about the EpiData-list mailing list