To do
- Not all nodes are in a town. There are 947 but only 938 is in a town. Investigate why. |Priority=4|
- There are some nodes in %NODE_DETAILS% that do not have longitude or latitude parameters. |Priority=2|
- www.ucu.edu.ar, sunysb.edu, buitms.seecs.edu.pk, www.nren.nasa.gov, rzserv2.gsi.de, gluino.phys.unm.edu, and ouhep1.nhn.ou.edu
- Find a way to set up prefixes programmatically. |Priority=1|
- Verify security of Sesame Server. Do I need to set up User/Password? |Priority=2|
- Read this http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.161.8030&rep=rep1&type=pdf
|Priority=2| - Make a better plan to populate the repository. |Priority=5|
- A file keeping the last action taken to insert a triple could be used to resume the execution of the program if an error (bug, timeout, or connection problem) occurs. Make a more segmented task flow to increase the chance of populating everything. |Priority=3|
- Verify if makes sense to use only monitoring nodes in the "FROM" parameter. |Priority=4|
- Use the context in every insertion. Read about context indexes. |Priority=3|
- Study and test visualization tools. |Priority=5|
About Schools |Priority=3|
- Search about Lucene+Sparql to increase and enhance the results for querying over School Names. |Priority=3|
- Search data sets about schools other than DBPedia.
- There are some endowment and number of students from DBPedia that are not execatly numbers. For example, "app. 95,000". Right now I am discarding this kind of data. Maybe I should retrieve the number information from these data
.
Doing
- Working on storing PingER measurement data into the RDF storage. (June 21)
- Testing Open RDF Sesame as a triple storage.
- Testing the generated RDF data in visualization tools such as Google Explorer.
Done
- Geographic data about a sample of PingER monitored and monitoring nodes (the %NODE_DETAILS) is already inserted in the RDF storage (hence using our ontology) and is query-able using our own Sparql Endpoint. (June 20)
- Towns, Countries, Continents, and Schools are inserted. However, it is needed to investigate how the results can be enhanced, retrieving more and better data.
Could do
- Create a package under the package 'general' to write the ontology. Create constant classes for each prefix. Then, when programming, instead of writing strings for each statement, you would refer to the constants created. For example, create a class called GNONT with a constant called NAME. Then, when programming you would use GNONT.NAME. |Priority=2|
- The method pinger.rest.HttpGetter.getJsonGeonames can be optimized. There is a test that verifies if the geonames username has been overused hence blocked. If it is, try another username. There could be a priority scheduling policy; a Map with priorities could be used to implement. The username that was able to be used would receive higher priority than blocked usernames. |Priority=1|