Software for offsite Tracks

Competitors in off-site Tracks obtain sensor readings and submit position estimates through the EvaalAPI web service. The purpose for offsite-online Tracks is to emulate a causal process similar to the data flow of an on-site Track: input data can be read only once and only in sequence, and the timestamp of each submitted position estimate must be the same as the latest sensor reading. For offsite-offline Tracks the aim is to emulate a batch process with constrained computation timing.

Competitors find documentation and source code at the EvaalAPI starting page, so they can start implementing the EvaalAPI interface in their system and report request or problems on the contest(at)evaal.aaloa.org mailing list.

Subscribing to the mailing list is strongly recommended to all prospective competitors. The list is used by organisers for latest news and by prospective competitors and anyone interested for discussions and for asking clarifications. Private inquiries can be directed to info(at)evaal.aaloa.org or to the relevant Track chair.

Competitors can run the EvaalAPI code on their own systems for quick turnaround using the test data files for their relevant Track together with the test evaalapi configuration file.

Applicants can write to info(at)evaal.aaloa.org and obtain testing (reloadable) trial names for running the above mentioned data files on the EvaalAPI web service. This is useful for checking one's system on the official server and are able to evaluate delays and possible network problems. In order to maintain the interface as simple as possible and avoid using cookies, EvaalAPI maintains no notion of a web session, so no public testing (reloadable) trial names will be available, and each prospective competitor needs unique ones.

Competitors are expected to try out the EvaalAPI interface and report any problems or observations.  Suggestions for urgent changes (to be made immediately) or improvements (to be made for the next edition) are welcome.

Registered competitors can ask for scoring (non-reloadable) trial names, to be run one right after the other, with an interval between them of at most one hour, in any order. In other words, once a scoring test is finished, competitors can wait at most one hour before starting the next one, order does not matter. Timing will be checked on the log files.