Meeting Notes for 8 October 2008 LIGO multiple channel plots Recording these meetings LIGO Evaluation News and Status Screencast demo

Meeting Notes for 8 October 2008

Present: Mihael, Tom L., Eric
Absent: Liz, Tom J., Bob

LIGO multiple channel plots

Eric reported that he's near to releasing a new version of Bluestone which can plot more than one channel on a single plot. He pointed us to examples he posted in a thread in the Gladstone room and talked us through some of what the plots show. You can, for example, see the difference between weekdays and weekend days in the anthropogenic band (yellow). Tom was pleased. He reported that when his students studied an earthquake they plotted each frequency band separately and posted them on the wall for comparison. Eric pointed out that such engagement is not a bad thing. But being able to see all bands together like this will let the students do the comparison more quickly.

Eric was concerned about whether earthquakes are showing up properly. Tom noted that his students found that the frequency band in which the earthquake showed up depended on how far away the event was. That could be the basis for an interesting investigation.

Tom asked about the status of integration of Bluestone with the e-Lab pages. Eric reports that it works, with some limitations. The new version of Bluestone has a "Save As" button and a field to give the file a name. When the student pushes this then Bluestone connects to the e-Lab site and uploads the image. It also tries to upload LIGO-specific metadata which would allow someone to reproduce the analysis, but getting these metadata to upload is not yet working. It will be. Also, tekoa does not have the software needed to connect to the e-Lab site this way, so this won't work on tekoa. Right now, with v0.61 on tekoa, when you try to save a plot it asks for reserarch group and password, but then the process fails with a warning (and the student is stuck). In the next release the "Save As" isn't even shown if it won't work. It does work on Bluestone running from www13. Since we plan to move Bluestone production use to the ANL cluster, and tekoa is going away, it's not worth the effort to install the extra software to make "Save As" work from tekoa.

Tom asked about the possibility of adding a link to the display of the plot in the e-Lab which would go to Bluestone and reproduce the analysis. This exists for cosmics. Eric says it's a feature which can be added in the future, but it does not exist now. Tom thinks it's an important part of the whole e-Lab/Grid experience.

There was some discussion of technical aspects of the interface between Bluestone (written in PHP, running on Apache) and the e-Lab pages (written in JSP, running on Tomcat). Eric thinks it's an elegant and general mechanism. To be able to re-run an analysis from a single button or link in the e-Lab would require implementing the interface in the opposite direction, in JSP. It may not be difficult to do, but Eric is not a JSP wiz (yet :-), so this might be a project for the new hire. There may be other things with higher priority, but don't forget about it.

Eric also suggested several ways that Bluestone and the e-Lab pages could be tied more closely together for a more seamless user experience after Bluestone is running on the ANL cluster. One, which he's already started implementing, is to have the PHP side remember the student's research group and password in the database (or as presented to the user, "on a keychain") and then use that info when needed instead of asking the student each session. (Right now it asks that info each session.) Eventually it may be possible for JSP and PHP to get each other's session cookies from the browser, and use that information to coordinate actions behind the scenes in our cluster, so that the user never sees evidence that the site is composed of multiple tools. That only works when both servers are in the same domain, so not when Bluestone runs from tekoa. We are not there yet, but Eric thinks we can get there.

Recording these meetings

Tom suggested that we should be recording our telecons so that we can play back details (like technical stuff described above). Unless it is horribly expensive, it would be money well spent.

Action item: Eric will find out how much it would cost us to record our calls. (DONE. Looks like $9/month)
Action Item: Tom will see if we can get money for this.

Eric suggested that we could also use EVO for our meetings, and that it might give us a recording capability. Except that he thinks the video requires too much bandwidth for him to join (and the video is not the useful part). If the audio is not as good as the phone call then keep the phone service, but use the chat and desktop sharing parts of EVO. Some of the QuarkNet staff have been using EVO, so it might be good to get feedback or suggestions from them (or via Bob?).

In any case, Eric has scheduled an EVO room for this time-slot through January so that we can try it out.

Eric also pointed out that the client is a Java applet, so if you don't already have Java installed and your browser configured to use it you'll have to go through the exercise. And this is an opportunity to see what teachers might have to go through if we want to use Java applets, such as an enhanced GUI for OGRE. (Eric is not suggesting that we use EVO for e-Labs, just for developer meetings. It's the need of installing Java that is the same for both, and might be instructive for the developers. )

Action Item: anybody who wants to should get an account on EVO (http://evo.caltech.edu), and try it out. If your machine does not have Java, or the proper version of Java, then note the process you have to go through to get the EVO client running. Teachers may well have to do the same to run Java applets we might wan to use.

LIGO Evaluation

Tom and Eric discussed need to get together with Dale to plan evaluation steps. Tom: should Dale be invited to these telecons? Eric: probably better to have separate call(s) with him.

Also discussed how to get teachers up to speed on e-Lab structure by example. The examples of Tom L and John K at the LIGO workshop were very helpful, because they could see the end result of the whole process. Eric suggests that before the teachers try the e-Lab on students, they should go through the e-Lab milestones themselve, playing the role of students, with Tom and/or Dale acting as teachers. Tom says he should go through the e-Lab completely himself, since he's not yet done this. When he does, this should be a talk-aloud execise. Dale should be encouraged to do the same. And these could be recorded using Screencast (see below).

Action Item: Tom and Eric need to get on a phone call with Dale about these topics, separate from the weekly Wednesday telecons.

News and Status

Eric had proposed talking about status display and newsbox today, but it seemed best to table such discussion until Bob could participate.

Screencast demo

Along the same lines of recording activities, Tom brought to our attention a tool which he found which can easily be used to record screen activity and microphone input into a video file. He guided us through a demonstration at http://www.screencast.com/t/CxGcN0xZVL2 Tom says it's very easy to set up recording, and the service will then convert the audio/video stream into any of a variety of video file formats, including Flash. The video file can be downloaded from them and hosted on our server, or hosted on their server, or uploaded to YouTube (but we noted the possibility of high schools blocking YouTube). This is a lot simpler that most YouTube videos because there is no camera -- it just records screen activity and microphone.

There are many potential uses for this:

1. Record a tutorial for how to use an analysis tool like Bluestone or OGRE or cosmics. We noted that it's better to do many short topics than one long video. For example, one brief video demo of how to use Bluestone for complete beginners. Then another on how to use intermediate, perhaps with several videos on separate topics.
2. Record other kinds of instructions, with screenshots and audio directions.
3. Record our talk-aloud evaluations (we didn't think of this until after the meeting, but it may be the most important use!).
4. Record EVO screen activity and audio from our weekly meetings (as discussed above)
5. Students might be able to use this to make "video posters" (again, an idea after the meeting - to be discussed further)
6. anything else?

Tom says Notre Dame has a subscription to this service and a server to host the videos, and as a result he has access to this and we can use it as we want, at no cost.

Action Item: Tom will provide Eric with the video from the example above, and Eric will incorporate it into an e-Lab page in the Library of the LIGO e-Lab on www13, as demonstration and proof of concept.

Action Item: Tom will use this tool in the next meeting to record EVO screen activity and the audio for our next meeting (or a future meeting).

-- Main.EricMyers - 08 Oct 2008