Controlling alignment and sorting are separate issues.

This is the current strategy for implementation:

1. Sorting a. Numerical variables: by numerical code: 1,2,3..etc b. String variables: alfa: a,b,c,d,e,f .....1234 ABCEFG....123...

But also that we should introduce options such as: /reverse : meaning reverse order /count: based on how many counts are in each category. if count and reverse are combined smallest counts first /limit: Only show categories with more records than the indicated value

The options I do not expect to have ready in first version, but the general sort order will be fixed.

WIth the suggestions of Jamie Hockin there is a good work-around of how users could have a string variable containing numerical codes could be sorted according to numerical value correctly.

2. Alignment Alignment is a completely different issue since this is not controlled by the individidual command, but rather by the formatting system of the output. All alignment and format is controlled by a "cascaded style sheet" contained in the file epiout.css. So for instance for the left most column this would be for right alignment: .firstcol {color: black ;font-size: 10pt;font-family: Fixedsys; font-weight: normal; text-align: right; padding-right: 10px; } and for left alignment: .firstcol {color: black ;font-size: 10pt;font-family: Fixedsys; font-weight: normal; text-align: left; padding-right: 10px; }

I suggest test persons try to work out different variants of the style sheet and we could provide these with the installation routine.

Currently we are working on build 18 of analysis. We have managed to incorporate the same source code for reading chk files as in epidata, such that we secure complete correspondance btw. the two programs. THe other major change has been to increase the precision of the internal representation of real numbers. Both changes require a lot of testing and going over functions, this is why there has been a while since release of last build.

Kind regards

Jens Lauritsen Initiator and Coordinator of EpiData.