Research Log of Web Science Students

We won’t be submitting our paper to PCSC.

For one, Google App Engine failed us.

Second, the tests we conducted were inconclusive. I also had doubts over it. So what if our system/integration testing was successful? That only proves that we coded the system well. But that doesn’t prove anything about the effectiveness of our learning system, which tries to posit a better alternative to learning.

I guess somewhere along the way I lost track of the motivations for this project. Well today I am setting out with a renewed focus.

While I’ve been talking about all this time how our app will be an alternative/supplement to formal learning, now the project will assume a futuristic setting: Informal learning has now been accepted, and the debate now is whether using Hardwire is better than using a scattered set of online software to learn. This is where the Hardwire becomes a mashup. It brings together all your online learning tools under one hub.

All the efforts now of our testing will go to proving that learning using Hardwire’s features is more effective than say for example using Diigo, Google Docs separately as part of conceptual learning framework. The test we thought of this afternoon is simple: Simulate the learning process in Hardwire and see the results. Specifically, a teacher makes a course ( a set of modules with prerequisites among them) or a module. Student learns that module and uploads his/her work products. Student has/her projects assessed via Projectrix. Also happening in parallel is another student learning a topic via his preferred method using software on the web. After this the students will be gaged as to who learner more using what method.

To be able to measure this, Sir Rom said that a set of parameters or some metrics should be used to assess learning using online tools. My first reaction was, “Where would I get those metrics?” to which he replied that I had to look for a paper that actually does that.

That’s that for the testing and paper writing.

Another goal of mine is to have Hardwire somewhat ready for production. Here are some of the features/improvements I’d like to work on immediately:

1. Refactor code smells
2. Pulling resources from social bookmarking sites and via Drag and Drop functionality make a course/module.
3. New layout.

As of the time on my computer (see screenshot) I am guessing we might not make it in time for Sunday. Ma’m Shelli and I will try to come up with a different play.

Today at around 6-8 pm the datastore was back up and we were able to test it with 5 people but while we were having dinner….

Paperwise I can probably say that this is a limitation of Google App Engine.

Indexes!

I remembered that I was able to solve this problem before simply by wiping the database clean. This was the easy way out. And today when I read the documention, I found out why:

Even if you don’t have many entities of the corresponding kind(s), the time indexes take to build or delete can vary widely depending on the total amount of data in your datastore, indexes currently building for other apps, and datastore load due to user requests, among other factors. In some cases, index jobs can take hours or even days to complete.

But it doesn’t stop there,

Even so, we can sometimes help with indexes that seem to be stuck. If you think this has happened to your indexes, feel free to ask us on the group.

I wonder what this means for complex business applications running on appengine. I think with slow index building, their queries wouldn’t work and their business specs compromised. But this goes for Java, I know the Python GAE is already very stable.

Is a move to Python in the future now in order? hahaha.

For now I think I must work on other bugs in Hard-Wire. I encountered a bug over the weekend with my Indexes and I already did the suggestions but apparently it takes time for Indexes to serve and as of now, they’re still not serving.

Uncaught exception from servlet
com.google.appengine.api.datastore.DatastoreNeedIndexException: no matching index found..      
<datastore-index kind="LearningResource" ancestor="true" source="manual">
<property name="resources_INTEGER_IDX" direction="asc"/>
</datastore-index>
...