Hello fellow project members,

here are the minutes of the last chat on the ESA project :

---

attending

present: Jose, Heikki

not present: Jeroen is still in South Africa, Erik couldn't join for personal reasons

---

summary

We discussed Jose's activities, what to do next in this project, and how to do it.

---

organizational

1. It has now happened with everyone of us four, that sometimes somebody is not able to join a meeting. Rather than postponing each time, it is proposed that we will have the "usual" weekly chat on Wednesday nights at 22:00, regardless of who may or may not be able to join. Starting next week (October 14th) when Jeroen is expected back in Europe !

2. After each meeting, minutes will be sent around to all four of us. This is the first example of such minutes. I volunteer to do this (but if I'm not in a meeting, obviously someone else must do it).

3. We must "slice" the work in parts that can be worked on concurrently. To do this we need a chat with all four project members present -- plus a real preparation from all four to discuss this in a meaningful way.

4. The next chat is scheduled for Friday October 10th at 15:00.

---

on the implementation

1. Erik, Jose and Heikki all agree that we shall use Axis2 to create our web services layer. Also, it has been decided that adding the web services layer is not a first priority; we will implement the CSW operations first, and when tested and done we will expose them as web services using Axis2.

2. In order to implement the CSW operations, we will model the CSW requests and responses in Java, and decorate this implementation with a JiBX binding. So we'll have, as an example, this following flow for a GetRecordById operation :

1. client issues GetRecordById request (an XML document) ;
2. this goes to a Java class with a method getRecordById(), accepting the XML document as input ;
3. the input XML is unmarshalled into a Java object graph using said JiBX binding ;
4. the requested record ID is retrieved from the Java representation of the request ;
5. the persistent storage is consulted (using Hibernate) to retrieve the record with that ID ;
6. this results in a graph of Java objects (ebRIM), and this is put into a Java representation of GetRecordById response ;
7. this GetRecordById response is marshalled to XML using the JiBX binding;
8. that's what is returned

alright, it's better to represent this with UML drawings and we will do so, this is just the minutes summary here :-)

3. Jose makes the point that we should use geotools for implementing spatial dimensions of queries; this project is only required to provide basic Bounding Box search parameters, but we'll implement that using geotools as it gives us a good starting point to support more complex queries later on ;

---

work items this week

1. Jose will download the code as it is now in the 3.0.x branch, look at it, and provide feedback and criticism this Friday in our chat .

2. Heikki is implementing the CSW operations' requests and responses in Java, using the CSW schemas to do it ;

3. Next meeting this Friday 15:00 !

---

This is the first time we actually have minutes of a meeting. We'll have them every time from now on. Please give me your opinion on how these minutes should be written if you feel they can be better structured, or any other suggestions. (and No, I'm not going to write them in Word !!)

Kind regards Heikki