I am about to estimate how much work it will be to get a running prototype of
the openehr-kernel, and I need to be able to run a few archetypes. I already
downloaded the source-code from ACode, which is a website in Swedish. I hope
I did understood it well and got the right thing, but it looks like it.
So I downloaded the source-code of the ehr-java-kernel, which is (I'm not
surprised), quite a lot.
Is there anyone who can tell me if this code should run smooth, this means, it
has no major breaks in it?
I am not a typical java-developer,although, "if you can scratch in COBOL, you
can scratch in every language" as my teacher told me long before the world
ever heard about Java as a programming-language. But I can read Java-code,
most of the time it is very easy to read, and learned the basics, I guess I
will find my way.
I also have need for some more/other technical details, like database-schemas.
I understood, it can take some time, before they will be published.
Can someone give me a hint about when that will be?
And my last question, is there already some work done for a
dotnet-implementation (which allows more programming-languages and therefore
easier to handle for many people)?
I am about to estimate how much work it will be to get a running prototype of the openehr-kernel, and I need to be able to run a few archetypes. I already downloaded the source-code from ACode, which is a website in Swedish. I hope I did understood it well and got the right thing, but it looks like it.
So I downloaded the source-code of the ehr-java-kernel, which is (I'm not surprised), quite a lot.
Is there anyone who can tell me if this code should run smooth, this means, it has no major breaks in it?
Are there not a comprehensive set of unit tests and regression tests as part of the source code? I don't know, I haven't looked, but for something as fundamental as an information storage and rrtreival kernel, the number of lines of test code should should probably outnumber the lines of actual programme code, or at least be very substantial - *every* single function should be tested, including all the corner cases. In the absence of such test suites, it is difficult to treat such a kernel as anything but a proof-of-concept. Now, writing a comprehensive set of tests is a lot of work, and since the kernel is open source, I am not suggesting that ACode should be responsible for all that work - but as the original authors of the kernel, one would hope that ACode at least set up a framework for unit tests as part of the project.
Can anyone tell us if such tests exists for the ACode kernel (sorry, don't have time to look myself right now)?
Are there not a comprehensive set of unit tests and regression tests as part of the source code? I don't know, I haven't looked, but for something as fundamental as an information storage and rrtreival kernel, the number of lines of test code should should probably outnumber the lines of actual programme code, or at least be very substantial - *every* single function should be tested, including all the corner cases. In the absence of such test suites, it is difficult to treat such a kernel as anything but a proof-of-concept. Now, writing a comprehensive set of tests is a lot of work, and since the kernel is open source, I am not suggesting that ACode should be responsible for all that work - but as the original authors of the kernel, one would hope that ACode at least set up a framework for unit tests as part of the project.
Thanks for your reply!!
There are a lot of tests, but I don't know if they test everything and in how far. I agree that it is good practice to write test-routines
I did not want to hold anyone responsible for something, I do not know much about the settings in which the code is developed.
I just want to know what the state at this moment is.
I guess that developers read this list, and they can tell in how far the code is stable and how close it is to a useable product.
- If they tell me that I better wait some time, then I can save myself the trouble of spending a lot of effort trying to understand something which will probably change.
- But it is also possible that they tell me that there are large parts which are very stable and useable, but there and there is something which will change. If this is the case, I can start a project of building an implementation of the good parts, if other people financial will allow me to do that, which again depends on the current state.
This is the kind of information which I am hoping to get.with my question
open source, I am not suggesting that ACode should be responsible for all that work - but as the original authors of the kernel, one would hope that ACode at least set up a framework for unit tests as part of the project.
With luck someone from Acode can give a definitive answer but I spent a pleasant evening with Rong Chen, their senior systems developer, after last year's Eurorec conference and he seemed very committed to the concept of exhaustive unit-testing. I would be very surprised if the appropriate frameworks did not exist.
I am about to estimate how much work it will be to get a running prototype of the openehr-kernel, and I need to be able to run a few archetypes. I already downloaded the source-code from ACode, which is a website in Swedish. I hope I did understood it well and got the right thing, but it looks like it.
So I downloaded the source-code of the ehr-java-kernel, which is (I'm not surprised), quite a lot.
Is there anyone who can tell me if this code should run smooth, this means, it has no major breaks in it?
Rong will be able to give you the best feedback ( I believe he is in Shanghai at the moment, so it mat be a while before he responds). But the CHIME group is working with this code in London, so we should also be able to provide some answers soon as well.
I am not a typical java-developer,although, "if you can scratch in COBOL, you can scratch in every language" as my teacher told me long before the world ever heard about Java as a programming-language. But I can read Java-code, most of the time it is very easy to read, and learned the basics, I guess I will find my way.
I also have need for some more/other technical details, like database-schemas. I understood, it can take some time, before they will be published.
actually, we spent a few days on specifically the approach for database connection just before my curent visit to Australia. Some design work in this area will be published soon.
Can someone give me a hint about when that will be?
And my last question, is there already some work done for a dotnet-implementation (which allows more programming-languages and therefore easier to handle for many people)?
there is a dotnet kernel alreayd working in Australia. It is not yet released open-source, but Ocean Informatics is considering doing that in the very near future.
Rong will be able to give you the best feedback ( I believe he is in
Shanghai at the moment, so it mat be a while before he responds). But
the CHIME group is working with this code in London, so we should also
be able to provide some answers soon as well.
>I am not a typical java-developer,although, "if you can scratch in COBOL,
> you can scratch in every language" as my teacher told me long before the
> world ever heard about Java as a programming-language. But I can read
> Java-code, most of the time it is very easy to read, and learned the
> basics, I guess I will find my way.
>
>I also have need for some more/other technical details, like
> database-schemas. I understood, it can take some time, before they will
> be published.
actually, we spent a few days on specifically the approach for database
connection just before my curent visit to Australia. Some design work in
this area will be published soon.
I am happy things are coming together now. I am not in a hurry,
I will wait for this information. Databaseschemas make things often much
clearer for me, everything falls in its place then.
>Can someone give me a hint about when that will be?
>
>And my last question, is there already some work done for a
>dotnet-implementation (which allows more programming-languages and
> therefore easier to handle for many people)?
there is a dotnet kernel alreayd working in Australia. It is not yet
released open-source, but Ocean Informatics is considering doing that in
the very near future.
Does it perform well?
Does it run from Mono?
I do not want to start a dotnet <-> java war. Both have their advantages, but
dotnet, for me, is more accessable
there is a dotnet kernel alreayd working in Australia. It is not yet
released open-source, but Ocean Informatics is considering doing that in
the very near future.
Does it perform well?
so far yes, no problems. We will open source it sometime before the end of the year...
Does it run from Mono?
haven't tried that yet, but it probably should - it is not GUI dependent.
Hi, me and my master thesis partner Joakim Fredriksson would really benefit from a working database or database-schema for openEHR. Do you have a more specific date when this document will be published so we can plan on when to work on this.
Regards
Jonas Andersson, laran813@student.liu.se
Linköping University Sweden
Thomas Beale wrote:
actually, we spent a few days on specifically the approach for database connection just before my curent visit to Australia. Some design work in this area will be published soon.
>Tim Churches wrote:
> open source, I am not suggesting that ACode should be responsible for
> all that work - but as the original authors of the kernel, one would
> hope that ACode at least set up a framework for unit tests as part of
> the project.
>
With luck someone from Acode can give a definitive answer but I spent a
pleasant evening with Rong Chen, their senior systems developer, after
last year's Eurorec conference and he seemed very committed to the
concept of exhaustive unit-testing. I would be very surprised if the
appropriate frameworks did not exist.
Ian
Thanks Ian!
Yes, at Acode we practise test driven development for all projects. It is particularly mentioned in the development guidelines published at the kernel project home page. If you take a look at the source code you will see unit test code for nearly all production code.
We will use unit test code as the safty net to drive the kernel forward and welcome the community to contribute to both the production code and the unit test code.