First of all i would like to thank the clear canvas team for the gr8 product they are building; For my organisation i am building a solution for storing and viewing dicom images of 3 different modalities, for the moment Clearcanvas imageserver and workstation are my premium option.

BUT! a real show stopper for us is the lack of a web viewer for clear canvas. Are there any options i should look at ? Are there alternatives ?

An other option would be that we could automate the workstation software to startup preloaded with examination defined in the startup parameters of the workstation; Are there possibilitys in this direction ? What we would like to do is : build a web page that displays all the examinations of a patient, when the user clicks on a hyperlink, the clearcanvas workstation viewer should start up with the particular examination. Is this possible ?

Thanx in advance

Hi, we do have a Silverlight web viewer that is in development right now. Can't give you any specifics on when it will be available publicly, but it does already work ... it's not vapourware :)

It involves some workarounds, but there are ways to automate the regular CC Workstation from within a web browser. The simplest way would be to define a custom "protocol" in your Windows registry that launches a command line "launcher" application, that then launches CC and sends the study info to the automation service.

I can't remember the specifics on how to define the custom protocol, but a quick web search should be helpful.

An example URI for your custom "protocol" might be CCIP://clearcanvasautomation/study=1.2.3.

Hope this helps,
Stewart

Stewart or Norman,

I was going to start on an Silverlight web viewer in September, but if you guys are already started can you guys open it up to the public so my team can help move the devolepment along?

Right now my team is working on extenting NHAPI to 2.6 and 3.0 and we are going to release it back to the public (clearcanvas).

We are also looking at how to improve the clearcanvas api that talk with the database to see if we can improve speed.

I would love to offer my team up to help with the Silverlight web viewer if you can open it up to us. Even if it is just Alpha code.

There are alot of good reasons to let us look at it because we are going to make it production ready and improve the speed for WAN.

Just let me know.

Gerald

Stewart or Norman,

Any feedback on the above comment? I would like to know so I don't try to duplicate the same work when I can have my team just improve on the silverlight viewer and give the code back. Also I am looking at trying to resolve the database locks. We are looking at using memcache.

Thanks,
Gerald

Hi Gerald,

We are in fact working on a Silverlight web viewer. We're not sure exactly when the code will be available yet, but it'll definitely be before September.

N.

Norman,

Cool. So I will wait. Are you guys going to look at speed over the WAN? If not then my team will look at it. If possible I would like to get a preview so I can have my team review it so that can already be thinking about how to improve the speed over WAN?

There is a patent on some of the ways to improve for speed over the WAN but we are going to write code around the patents (US and CA). Of course we are going to give the code back to the project.

What do you think about using memcache for all select statements and all others go directly to the database?

This would improve the speed and remove any locks issues.

I have done projects were I have use memcache and it has improve speed and performance over 100 percent.

Also if the database goes down for any reason you can still retieve just not store.

Thanks,
Gerald

Gerald,

I just wanted to confirm, but I assume your comments on memcache are related to the ImageServer?

On your past performance improvements, what type of system were you working with? Was it more of an IS system where objects were frequently loaded, updated, and then reloaded later? In general, the ImageServer doesn't really function this way, with somewhat the exception being the WorkQueue. The areas where we do access objects frequently (the Filesystem table, the ServerPartition table, and actually when we load locations of studies) we have implemented some caching mechanisms in the code already.

You're free to give it a shot if you want. One thing to note, it has to work in a clustered server environment. Our primary use case for the system is having multiple shredhostservices running on different servers accessing the same database and sharing filesystems, etc.

Steve