Analysis of Existing Results

[aarontabor]

In order to determine solutions for some of the existing bugs, we need to better understand characteristics of the set of existing results as a whole. Questions that need answered:

• Can I find a universal Header Identifier?
  • Place?
  • Bib?
  • Name?
  • = (assuming header will be all lines between two banners?)
  • - (assuming header will be all lines between two banners?)
• How many results pages use a character other than '=' for banner (used when separating fields)?
  • What are the other symbols used for these banners?
• Is there always a field separator banner? (how many result sets don't have them?
• How many results pages contain more than one set of results?
• Is there a specific format in regards to the race data at the top of the page?
  • i.e. an expected order of fields, ...

Answering these questions will help us determine whether we can manually address some of these abnormalities in the data (because they are super fringe cases), or whether the system will have to account for them.

[aarontabor]

I've got all the results downloaded, and I'm working at writing some scripts to answer these questions.