Hi,
Is there anyone who's done any work related to openEHR for the iPhone/iPad?
I guess one would need an implementation of the reference model etc in Objective-C since that's the only supported language on the platform.
If you are working on this or interested in the topic then I would be interested in knowing about it.
regards
Olof Torgersson
Hi Olof,
Since your query includes people interested in the topic, I end up being in the result set. I’ll write down a couple of points below.
I am really interested in this, but for me Mac is not an affordable platform at the moment. I simply can’t cover the cost of having a proper Mac and an iPad. This did not stop me from thinking about it though.
If you go with a native app, you’ll need to develop it in Objective-C, or that has been the case up until recently. With iOS 4.0, Apple introduced severe limitations to its SDK, the worst one being not being able to use any other language other than Objective-C and Apple development environment. (not so sure about the dev env)
This led to lots of criticism, and recently Apple eased it a little bit, which made monotouch ( http://monotouch.net/ ) an alternative solution to Objective C, or that is my understanding of the status. For companies with C# implementation, this is a serious advantage. Other companies would probably produce objective c back ends for their compilers if this practice gets established, but you’d still have to use Mac as a development platform to link to iOS functionality.
If you’re not doing C#, then your options are Objective-C and web applications. I won’t go into HTML 5 praising here, since I personally see it as a dangerous hype at this point, till the tooling side of things mature. For web applications, you can again leverage your existing investment with some improvements/customizations for Apple platform.
However, there is one critical aspect of native application deployment: the Apple review process. Apple has very strict rules for not allowing apps interpret code. How and why an application is considered to be interpreting code is completely up to judgement of Apple people, and if they think your app is hosting another app, then you simply won’t get your place on the app store. The reason this may be a problem is the nature of openEHR: various components may be plugged at runtime and Apple people may interpret some functionality as intepreted code. If you think about it, many runtime aspects of openEHR implementation read model descriptions and orchestrate things accordingly. This is the power and flexibility of openEHR, but to Apple, this means an app hosting other apps, thus losing control of their app store (the last bit is my view of their approach).
I can’t see any technical issues with Apple platforms. In case of Objective C it would be a big task, but that is all. Apple’s iron claws around deployment would be my concern. The web app is probably the safest when it comes to deployment.
Anyway, my 2 cents (is there a British form of this expression? just out of curiosity)
Kind regards
Seref
Olof Torgersson wrote:
Is there anyone who's done any work related to openEHR for the
iPhone/iPad?I guess one would need an implementation of the reference model etc
in Objective-C since that's the only supported language on the
platform.
The reference implementation in Eiffel could, in theory, be compiled
on the iPhone platform. The Eiffel compiler was experimentally ported
to the iPhone last year:
http://dev.eiffel.com/IPhone_Development
http://www.eiffelroom.org/blog/manus_eiffel/eiffel_and_iphone_ipod_touch
http://www.eiffelroom.org/blog/manus_eiffel/eiffel_and_iphone_ipod_touch_2
(The site seems to be down at the moment, so I'm posting these links
based on my Google-assisted memory.)
This would be by far the easiest route ... if it works, and if Apple
permits it.
- Peter Gummer
Hi Olof,
...
If you go with a native app, you'll need to develop it in Objective-C, or
that has been the case up until recently. With iOS 4.0, Apple introduced
severe limitations to its SDK, the worst one being not being able to use any
other language other than Objective-C and Apple development environment.
(not so sure about the dev env)
C, C++, and JavaScript as executed by the iPhone OS WebKit engine were
allowed as well.
This led to lots of criticism, and recently Apple eased it a little bit,
A quote from Apple that is just a few days old:
"In particular, we are relaxing all restrictions on the development
tools used to create iOS apps, as long as the resulting apps do not
download any code. This should give developers the flexibility they
want, while preserving the security we need."
http://www.apple.com/pr/library/2010/09/09statement.html
So even Flash is now possible again:
http://blogs.adobe.com/conversations/2010/09/great-news-for-developers.html
...
Anyway, my 2 cents (is there a British form of this expression? just out of
curiosity)
http://en.wikipedia.org/wiki/My_two_cents_(idiom)
Kind regards
Seref
Cheers,
Roger
Hello Olof and everyone
I think that the iPad / iPhone devices could be of some assistance as
far as the interface to an openEHR enabled application is concerned.
It could perhaps enable a more tactile manipulation of the entities
contained in the health record...(tactile and efficient, i wish, i.e not
a gimmick)
But the actual modification of EHR content and interaction with the
server does not need to occur on the device itself, this would certainly
complicate things a lot.
I vaguely remember Erik Sundvall talking about a REST approach to
handling openEHR content. In this case, all the iPad / iPhone /
iWhatever device would need to do is POST (or GET) data from specific
URLs just like many other web applications do now.
But until we hear any more news from that front, i suppose that you will
have to rely on a web application based on either the java reference
implementation or OSHIP. In this case, you could work a little bit with
the presentation layer through CSS (and the fact that the page is
accessed by a mobile device) to perhaps adjust it to the iPad / iPhone
screen size or other property.
I hope this helps.
All the best
Athanasios Anastasiou
Dear Everyone,
we here in Linköping are now running a student project were the EHR is to be viewed on a pad, most probably using HTML(5?) for presentation of data generated from our openEHR system (which is based on the Java reference implementation). This is a first run for us and a very small project but will hopefully give us a feel for what could be possible. The project is supposed to run during the autumn and has just recently started and, I should also add, the students doing the project are in the beginning of their training. Hopefully we can continue on from this project with master students or similar.
Regards,
Daniel
Hi,
That sounds very interesting. I guess the same backend/server-solution could be used for making a native user interface instead?
Regards
Olof
24 sep 2010 kl. 12.49 skrev Daniel Karlsson:
What needs to be considered is, for each particular app, what data are being captured and displayed? Usually it is not ‘everything’ - it will be a very narrow data set in each case. For each such data set, one or more openEHR ADL 1.5 templates could be created and converted to XSD and/or programming objects (dedicated classes for the particular template). These would greatly enable the construction of such apps and also the communication of the data back to a server. An app framework is needed form which any specific app created according to this programming model can be built. I don’t know that the language (Objective C, Java etc) is that important; the above can be done in any ‘normal’ modern language.
Some examples of ADL 1.5 templates are available - see the ADL Workbench beta, http://www.openehr.org/svn/ref_impl_eiffel/TRUNK/apps/adl_workbench/doc/web/index.html (see particularly the example archetypes at http://www.openehr.org/svn/knowledge2/TRUNK/archetypes/ you can use with this tool).
In the next few weeks, the XML OPT generation feature will appear in this tool. This is the basis for generation XSDs, programming objects, message schemas and so on.
This discussion brings another one to my mind. Should be templates
dependent of the target platform? (or in other words, is visualization
ruled by tamplates?)
Hello everyone
Further to Thomas' comment about generating the XSD from a template and through this, suitable data structures in "any" language and in relation to Olof's question, i think that if the focus is the GUI and the presentation of data and information, then the application can simply be a mock up.
It would be demonstrating the different novel ways the specific openEHR data structures can be manipulated through the device but it would not have to actually do any calls to an openEHR server (or, it could use local storage, again for limited demo purposes)
I hope this helps.
All the best
Athanasios Anastasiou
Hi, and thanks to everyone that has replied.
I agree that the best way would be to create a GUI on the iPhone/iPad and let the actual data be managed
on a server somewhere, thus one should not need to do so much core work in Objective C.
Personally I would like to create a native GUI and not something that is run in a browser since the interesting thing is really
to investigate the opportunities a native interface on a device such as the iPhone/iPad could provide.
Then the question becomes what are the existing solutions one could use for such a scenario? (some suggestions have been given)
I do have a reasonable knowledge of Objective C and the relevant frameworks since more than 10 years ( of course the iPhone specific parts are not that old, but conceptually it's the same as Cocoa (Mac OS X) development.
regards
Olof
24 sep 2010 kl. 12.11 skrev Athanasios Anastasiou:
Hello,
It is correct that we in Linköping currently work on an implementation of
openEHR using the REST approach. The implementation will be open source when
the implementation is ready. (In fact Erik and Marie are currently
discussing the latest commit to the project just outside my office door!)
Greetings,
Mikael
No offence but I prefer not to top-post. Specially considering this
thread is moving a bit OT.
Please comment below.
ons 2010-10-06 klockan 11:24 +0200 skrev Mikael Nyström:
Hello,
It is correct that we in Linköping currently work on an implementation of
openEHR using the REST approach. The implementation will be open source when
the implementation is ready. (In fact Erik and Marie are currently
discussing the latest commit to the project just outside my office door!)
I'm not familliar with what Mikael is involved in but I can tell that at
least the major integration project IFK2 is "heavily inspired" by the
REST philosophy.
The project adresses the possibility to integrate care documentation
applications (i.e. the journal) with quality-metrics systems. (In Sweden
there are several national wide registers dedicated to QA for certain
care processes)
For several reasons we developed a model where we separate the semantics
of the content from the model of the communication. Essentially we want
an interaction model that is dependent on the operations in the register
rather than the semantics of the domain objects.
Access is going to be governed via a SAML-ticket architecture against a
large user and service directory and services are accessed through a
MULE based service bus. The registers develop and publish changes to
their data requests on a regular basis. Hence it is nessecary to
decouple the services provided by the registers from the data requested
by the service.
For other reasons, we use a normal SOA architecture but, essentially
there are only operations defined for normal CRUD actions. One might
have the right to read but not write, create but not modify, et.c.
The semantics of the content is then packed to an (unpractically large)
CLOB wich is marshalled and validated on the end node.
This way, templates can be updated w/o the need to redefine the
services.
Project is being delivered for evaluation but there are some tecnical
conclusions being written for intended publication by the end of the
year.
Very REST, but a bit Off Topic in the dicussion on how I am to
professionally motivate my employer to purchase me an iPad... 
Regards
Tomas
/.../
Hello,
It is two separate projects. The project I describe is a research and
education project performed by Erik Sundvall, Martin Eneling, Marie
Sandström, Rong Chen, Håkan Örman, Daniel Karlsson and myself at Department
of Biomedical Engineering at Linköping University in Sweden. Erik will
probably send out some more information about the project soon.
Greetings,
Mikael
Hi,
If I understand correctly the things you mention, that is ADL 1.5 templates, opt generation and documentation of these do not yet exist,
Is there a timeplan for when they will be finished?
Regards
Olof Torgersson
24 sep 2010 kl. 12.56 skrev Thomas Beale: