Walter

>>Are you saying that this way is easier or less troublesome than with Winsock ? I'd like to know the most solid and fast to handle the HL7 protocol.<<

I guess it depends on the need. If you just want to interface to one device, winsock might be a good idea. If you're wanting something for 10 customers to pull data from their different lab softwares, the file-in-folder option might save you a lot of comms development grief. If you are planning something to be interfaced to lots of different software, devices, whatever, at lots of customers, comms will be a nightmare if you try to do it all, better to look at www.seebeyond.com.

I see the interfacing as sort of like using Views in VFP- even if you are sure your customers will always use local tables, might as well use views so when the dreaded day comes you don't have an immediate rewrite on your hands...

>>Different variants of HL7 or different levels of communicating ?<<

Different variants of HL7. It varies country to country, but as healthcare internationalises you should plan that different institutions like labs may use different naming conventions. E.g. for "Mean Cell volume" you might get

OBS_RES|ST|MCV|Mean Cell Volume|L|||100|fL|80-98|H|||F

from one lab, and

OBS_RES|ST|8005|M.C.V.|L|||88.5|fl|(80.0 - 96.0)||||F

from another. This isn't the best example, but if you need to identify MCV, "8005" and "MCV" both need to be caught. FWIW, these segments are from HL7 messages that do not follow any of the published standards. That's what happens when a dominant provider sets a (wrong) standard and others follow.

>>I assume that you'll get an error string returned when you trying to send an invalid string. Are there any more gotcha's to take care of ?<<

I'm not sure of your need... retrieving information certainly requires a great deal of care- I've seen data dropped into the wrong patient because of a missing/malformed PAT_IDF ....

Regards

JR