> I don't think there is an AQL engine open source yet, but in any case it only > makes sense when there is an open source openEHR EHR service, which there > currently is not.
I don't think it is possible to write an AQL engine right now, because it is not defined well yet. One can only anticipate on what one thinks it will be. So, my guess is it will be something between XPath and SQL. Instead of fieldnames, using paths replacing fields. I think the specification takes too long to arrive. It seems to me quite obvious what it will be. But writing an engine, from scratch, for it will cost, more then one year for a very experienced developer, and even then..... Companies like Oracle spend a substantial part of their developer-investment in good query-engines. Storing data is peanuts, queriyng them is the hard part. But there is a way around, use what others did, and that is, using path-based query-engines. There are quite some, also open source, good ones too. It is just a matter of storing your OpenEHR datasets path-based, and query them path-based. For example, using an XML-database. Maybe there are other possibilities too. And then you have a full featured AQL engine, as I think it will look like in the future, when the specs will be finally written. Maybe some syntax-translation is needed. What could the AQL-specification hold for promises, which are not already delivered by XPath/XQuery right now? I cannot think of anything. I think it would be wise for the OpenEHR community to look very well on what is already done by commercial companies and open source communities for years now, instead of reinventing the wheel, unless of course, when there are good reasons. It would make introduction of many OpenEHR implementations much more easy, and that would be good for worldwide success for the OpenEHR specifications. Ok, these are my two cents. I am very anxious to learn why the current XPath/XQuery-specifications are not good enough. Have a nice sunday. Bert. Verstuurd vanaf mijn iPad Op 20 apr. 2013 om 18:29 heeft Thomas Beale <thomas.beale at oceaninformatics.com> het volgende geschreven: > I don't think there is an AQL engine open source yet, but in any case it only > makes sense when there is an open source openEHR EHR service, which there > currently is not.
