Hello all, I am interested in the specifications for the RATINGREPORT.XML format which is to be supported this year. Does anyone on the forum have a file with the specifications? Alternatively, could you point me in the direction of the best person or persons to contact about acquiring the specifications? Thanks!
At this point I do not know when the XML specifications will be released or supported, because there are too many higher priority projects ahead of it.
Thanks Nolan, I appreciate the response.
I will be bringing this issue up at various workshops in Orlando, in the hopes of finding someone willing to do the grunt work on an XML standard for rating reports. I don’t really have the expertise and almost certainly won’t have time to work on it until 2015 at the earliest.
XML is a very verbose format, but there aren’t a lot of other non-proprietary choices out there.
For uploading a rating report, it’s probably not too bad. WinTD has an option to dump a tournament in XML in a spec which includes much more information than a rating report has, and the SN II copy of that is around 8Mb. A rating report might be 1/3 of that. By constrast, the TDEXPORT.DBF with the name, etc. information is 500K. Where XML would likely be a problem is as a download format for an expanded GDB. I could see that being 1Gb, maybe zipping to 100Mb.
What format/spec is generated for a FIDE report? I couldn’t find a spec hunting around the fide.com site and related sites (rules commission, arbiters commission, etc.).
The format for FIDE rating reports is the “Krause format.” It is a plain text format with fixed width fields. Details are available here.
As I recall, the tests I ran in early 2013 on a Gold Master Rating List XML format, with quite a few new fields in it, including FIDE ID and country codes and USCF norms-based titles, produced files that were around 150MB uncompressed, zipping down to around 18 MB, which is similar to the size of the latest Gold Master DBF file (14.2 MB for the regular/quick file, 13.6 MB for the regular/blitz file.)
I did post a sample of that format in early 2013, but I don’t know if anyone ever looked at it. The README file is still available at:
uschess.org/datapage/supplem … aft.README
If there is sufficient interest, I can create and post a full XML file later this week after the August Rating List is generated.
I had a long conversation with Christian Krause in Turin in 2006, he acknowledged that there are multiple weaknesses in the FIDE rating report format as it currently exists, but also recognized that changing it would introduce lots of complications, just as it would if/when USCF goes to a new supplement format.
There is an experimental version of an XML-formatted Gold Master file for August 2014 posted. (Look at the page where the Regular/Quick files are posted, below the links for the Gold Master files.)
In the XML file there’s a player with an ampersand in the name. This is a special character in XML and I think it should be escaped. The entry is located at line 940876 (PM me if you need more info). There are other occurrences of the ampersand in the file, but I guess this is the first one which the parser choked on.
SyntaxError: not well-formed (invalid token): line 940876, column 15
I’ll look into it, might take a few days.
The August file is no longer linked. Anyone have the current copy or an older copy for testing purposes?
The link has been repaired.
I see absolutely no advantage to XML. As you point out, it is extremely verbose (and named end tags are redundant, since overlapping is illegal).
You need to write code to handle the output from the XML parser anyway, so you might as well write code to parse a formatted test file.
If you want simplicity, use a NAMELIST. 
Interesting glitch: All the “national_master” tags have a year of 1991, even for players who were not born yet as of 1991. 
I’ll look into that after the US Open.
Proof that masters are born and not made.
There are too many masters that we did not maintain a record of when they first achieved their NM status, we currently show ‘1991’ for those, I may change it to “Prior to 1992” or something similar in the next trial version of the XML file.
But for those who didn’t make master until after late 1991, the XML database should show the year in which they made master.