This page last changed on Feb 20, 2008 by scytacki.
Importing and exporting from manufactures software. - design by May 1st? We need to know the parts of the object store before we can do this. We also need to look at each vendors software to figure out the issues. We have started analisizing.
We need to collect data from the manufactures hardware and display it in CCPortfolio. The plan for this has gone through some revision. Based on my initial estimate it will be 1 week for talking to manf to determine who is involved and what they can do. 2 weeks for design and 2 weeks for implementing the prep part for the manf. And then there will be support time while they are developing.
Required Sensor Summary
Initially we planned to integrate by giving the manufactures several options. And the curriculm would be written so it could support even the lowest leve of integration. This lowest level was for CCPortofolio to launch the vendors software. And then the student would cut and paste the data from this software back into CCPortfolio. After the curriculm writers looked at how this would restrict the activities they decide this wasn't acceptable. So now...
The integration needs to be good enough that it is easy for the student to collect data then go back to CCPortfolio and answer a question about it. Collect more data go back to CCPortfolio and answer a question about it. So...
We came up with 2 solutions for this:
- We define a driver spec and each manf. writes a driver to that spec. The driver could be a Java library, or a native library. It could also be a separate process that streams data back and forth to CCPortfolio. Whether we supply the transport layer for this separate process is undefined.
- We access their probeware software though inter application communication. This might be applescript on the make or COM on windows.
Carolyn took these 2 solutions to a meeting where all the manufactures were and they didn't seem turned off by this. So now we need to figure out the next steps. I doubt that we can support all of these manufactures. But if we had to then this is the approach I would take:
- determine which solution is more preferred: my guess is that would be the driver solution.
- this involves defining what we need from them a little better and then talking to technical people at each manufacture.
- probably a few iteration of doing this and having Carolyn talk to their CEO might be required.
- Lets assume they choose the driver solution
- We need re-examine the CCProbe code and see where the driver layer is. Then make sure this has all the functionality that we need and the support the manufacturers will need. Define this functionality and support in an abstract way not tied to waba or java.
- make the decision about waba and java.
- make a sample driver maybe with hardware so they can test it.
- send it off to them and give them some level of support as they are working on their drivers.
Unresolved issues with this approach
- how much do we do? how much do they do?
We have to draw this line so we don't scare all them off, but still we can accomplish the other things we need to get done.
- we need to deal with their development schedule. How much time are they going to need for this development?
- when does this need to be completed?
- deployment issues. If they do native drivers can we deploy them? Does the student need to have their software installed?
My opions for some of these things
- I would draw the line at the java driver. We will define a java/waba driver interface with an example. They need to do the rest. If they want to use this to talk to their software they can. If they want to use their native driver they can.
- it doesn't need to be completed until late fall perhaps October.
- we need to spend time developing the whole system and not just focus on this piece. Doing the prep work so the manufacturers can start will take a while perhaps 1 month of my 50% time. Lets say 2 weeks for design. And 2 weeks for the implimentation. So they wouldn't be able to start until mid-may at the earliest. But we if we do this other stuff first then that would push this back a few more months. So probably more like mid-july before we get them anything. Then supporting them will take time away from the rest of our development.