Reference Implementation Meeting 2011-03-01

From Direct Project
Jump to: navigation, search

Reference Implementation WG Call

Tuesday, March 01, 2011
12:00 PM EST

Beau -
Sign in on the RI workgroup page.

Agenda:
Status update on both teams. Looks like a lot of check-ins from both teams.

We had a good showing at HIMSS.

Umesh -
Glad to be back with you all.
#1 we've started the scrums back up again. Last week Ali and I were working on DNS specifically. What we did was harden up the TCP server quite a bit, especially with badly behaved clients. As we see other things, we'll keep checking-in those. The things that the system does do is it's a lot more aggressive about timeouts, the settings are neatly exposed, we also got rid of a few bugs. Ali and I have been extreme programming it. As we've been changing the stuff we've been writing a ton of those unit tests. Additionally, we've introduced the load balancing stuff, base classes, as well as folder writing, into both the core object model and into the gateway. Last but not least I've abstracted out some of the routing stuff so we can add a piece of that. If you wanted to you can write a route that will push to a web service.

Arien -
One thing I realized I forgot to do when we were talking about the DNS server was that I updated the spec a long while back and I added in the notion of the other kind of cert. record for X509, the URL cert. There are two codes that are documented for X509: 1 is where you place the cert in the zone file itself and the other is where the record is a URL. There is actually a well-defined spec for putting certs in the reference by URLs, so it would use the same coding conventions.

Umesh -
What we'll do once we're done with what we're doing, Ali and I will take a look at that. Ali is taking over some of the DNS stuff.

Greg -
I think some of that came from the VA and it is in the spec. It's just another type.

Beau -
Has Java already done that?

Greg -
We haven't added it in Java yet.

Umesh -
You and Ali can sync up then.

Last but not least for us, this week I'm going to focus on Docs, and also getting more partners up to speed with HealthVault. It's amazing the rate at which this is accelerating. Finally, the real momentum is going to shift over to XD.

Greg -
A quick update on the documents. I believe you guys are doing VC10. I believe it might even be of the same open source set. There is an equivalent for the Doxia pieces for VC10. I'll send you the info.

Umesh -
I'll take that down for now. I'm just going to focus on that. Yesterday I documented the DNS server settings. The first goal is to capture all of the settings. The next goal is that I'll start writing the English.

That's it from C#. We'll probably capture the bug fixes into a fresh release hopefully middle of March. That's my tentative goal.

Beau -
Vassil is attending the scrums again; hopefully we can get some traction on that.

We'll move onto Java.

I know Greg has checked in a large amount of code. We're up to 90% code coverage...


Greg -
Mainly what that was for… there is a gateway 1.2 that's out there now that has the auditing support and the extensibility where you can do more overwriting. Before I released that out I got code coverage kicked up to about 90% on the gateway. There are tons more unit tests. I did find some issues and fixed those along the way.

Most of my work is working on our custom HISP stuff. Getting 1.2 was the big thing. It went into release and it's out in Mavin Central as of last week. The Doxia book is update too. We ripped out some of that test setup stuff and now it points back over to the wiki and so forth.

Beau -
We've also had a new contributor to the team, Brian Hoffman. He's been a superstar with the UI pieces, making it a lot prettier and more functional.

Brian -
We're supporting DoD in our code and we wanted to make some changes to make it more functional. As we were having some issues trying to understand the interface. As far as going forward, it's probably good to open up the discussion as to where we want it to go - I don't want to railroad the work that's already being done.

Umesh -
C# has a UI that's been orphaned if you want to help out with that too...

Beau -
Anybody else have any other Java updates? [no] Vince has been working on the XD stuff and had a successful demo at HIMSS. There are a couple of bugs outstanding. That's really good news on the XD progress.

Greg -
I have a question on the XD part of what we're doing for RI: The XD WAR file, is that supposed to kind of be a mock endpoint where the routing pieces would go and send an XD message? [yes] Is there any possibility that that implementation could at least do something useful instead of throwing an exception? If not, does it even make sense for that WAR to be inside the RI?

Beau -
I don't know a whole lot about that right now, but that's something we can bring up on Thursday.

Greg -
I have one last piece. Umesh and I were talking about one of the pieces in the S&T spec and the way that we have done the implementation specifically has to do with required fields. We noticed that there is a required field ("To:" field). There is potentially a use case that we came up for that we're not sure if it's valid, but if someone with Outlook or Windows Live wanted to send messages to BCC only. According to the spec, that would break, but in practice and implementation it would actually be allowed because we're actually using SMTP RCP2 envelope headers to send the information across. Is that something that potentially a use case that we didn't think about? Is that a valid use case?

Arien -
There is language in the spec that says "there be dragons," particularly for Health Information Exchange. I don’t know enough to know whether we should be afraid or look the other way.

Arien -
There is definitely language in the spec that says prefer the SMTP commands to the RFP. I think the implementation is doing the right thing. The question is whether the sender/receiver should be doing that in any case, and whether the receiver can properly interpret the BCC. If there is some kind of implied security or privacy issue it's kind of an inappropriate thing to be doing.

Umesh -
It's possible that while it's legally okay, the receiving system may drop the message.

Arien -
I think it's a content/container issue, not a routing/gateway issue.

Greg -
That's fair. I was trying to think through that a little bit especially as we're getting to the XD bridge pieces. I'll have to delegate over to Beau and Vince on that one. I know at least the Java implementations could handle those use cases but if you're looking at it generically there could potentially be issues with that.

Arien -
I think the gateway is doing the right thing.

Umesh -
I concur. Good catch, though.

Beau -
Anything else?


Okay, we'll talk to everyone next week!