This topic contains 2 replies, has 1 voice, and was last updated by  Eliot Muir 9 years ago.

Friday hacking time for Iguana!

  • It’s four of the projects that people trying out here in our offices on Friday for the masters class. Still have another 4 people that have yet to figure out what they would like to do.

    Well – that turned out to be a lot of fun! I think everybody that participated got a lot out of it. It pretty cool to host a workshop event every person was doing something new and different.

    It definitely took people out how they would normally think about interfaces and what the role of the interface engine can be.

    Some of highlights – we got a great idea to expose TCP/IP connection information on our LLP Listener components – that will make it easy to make a custom dashboard to show that information.

    There was a lot of good discussion about the ideal monitoring solution. The main concepts that stuck out for me are that:

    • It’s key for notification systems not to be noisy. If they give lots of false alarms they will be ignored.
    • The system should be good at only bubbling up issues that the system admin can actually do something about. If the problem is something that only the another person can fix on Monday there is no point losing sleep over it at 2am on Saturday
    • The real state of the art in notifications is where you can give good notification information to the other side. i.e. the people you would have to talk to on Monday.

    Few people were doing interesting projects that related to this goal – I’m using people’s first names since they might not appreciate being mentioned online.

    Casey built a notification script which queried the service log of Iguana for errors and filtered out spurious ones – namely login errors from staff mis-entering their passwords.

    Roberta was successful in modifying the channel manager to filter out a set of channels that were specific to the client that was logging in. In that way the lab’s clients could log into a portal and see the state of their interfaces.

    After she did that she went on and ran with the ideas in this document:

    She built out an example of showing the tracing of a lab order through the interface. The idea here is a show a trace of each lab order from the point of view of the non-technical user so that they can see all the steps just like a Fedex tracking system only for a lab order:

    1. When the lab order was requested
    2. When it was received by the lab IT system
    3. When it was performed and key details about that
    4. When the results were sent back via the interface

    All very powerful stuff and way ahead of the curve of what most labs offer in terms of visibility to their clients for this information. There is so much progress that could be made in the industry here. It’s a lot to do with limited IT resources – but if you Iguana and participate in with us in these applications then this quite possible to achieve.

    Ryan and Jonathan managed to implement the ability for the global monitor:

    So that it can actually stop and start channels on those remote Iguana instances. That’s pretty wild – the efficiencies in being able to give commands to remote Iguana instances rather than having to log in to awkward VPNs and remote desktop sessions is huge. That’s just the start of what you can do when you have the technology to remotely control an Iguana instance behind a firewall.

    Jay got his project working for an external Iguana app can could resubmit messages. We’re keen to take that to the next step of putting some of the HL7 data into a form with fields denoting the fields like SSN etc. Once again something that could really be just the thing for a lot of work flow situations.

    Bret and Sergei got to apply the regression testing app to use it as a tool to convert a old vmd converting HL7 into text files – over time I think we’ll build up a library of these testing scenarios.

    There was a lot of discussion around using the channel manager to stage interfaces from development, to test and then production. Seems like our clients are all over the map in this area – some organizations have very well defined processes in this area – others less so. I think there was a lot of benefit for everyone who attended getting a feel for how their peers in other organizations deal with the same problems.

    All in all a successful event. The plan is to really hone in and get good at these events so that come the user conference we’ll be able to have a track that can do this on large scale – a lot of the feedback that people gave us from our last user conference is that they really wanted more hands on sessions where they could learn specific technical skills.

You must be logged in to reply to this topic.