®| ® |
|
|
|---|
Last update: Friday June 3, 2005 9:44
As we've entered a new year (or for purists a new millennium), I thought it was time that I let you know where we were on Grading.
The more attentive of your have spotted an inconsistency in the closing date for submissions between my original communication and the subsequent one (dated 27/11 and 22/12 respectively on the BCF Website). I confirm that the deadline for submissions of grading results is 30th June 2001, but would be grateful if events that are finished earlier can be sent in as soon as they are ready, to cut down on the last minute rush. We will shortly issue a checker for grading data submissions (see below for details). I would ask those of you who are about to submit Grading data, and are able to use the submission checker to wait until you receive the checker.
Work is continuing on processing Grading results omitted from the published 2000 Grading List and correcting those anomalies that have been pointed out to us. It is expected that this work will be completed by end of January, when an updated master list, together with any additional software deemed appropriate, will be released to Graders.
Testing to confirm the feasibility of using the hybrid system for central Grading computation is expected to be completed by the end of January.
A manual process has been put in place to confirm that new players really are new. To date over 200 requests for new player codes have been processed, out of which about 30 have proved to be existing players. However, there is still a backlog of about 500 new player requests, which I expect to be cleared within the next two weeks. So if you are awaiting new player codes, please be patient. I would ask all Graders to send in their requests as soon as possible, since a late deluge will jeopardize getting the Grading done. In this respect, I am imposing a deadline of end of May (i.e. one month before the deadline for submission of grading results) for requests for new player codes.
Last year, not many people received error logs due to the processing time involved. We are still investigated how to streamline this process so that we can provide feedback to Graders on what we think they have submitted.
A data checker for submission files, which will be provided to all Graders, has been developed and is now under test. There are two versions, one that can check spreadsheets and requires Excel 97 to be installed on the machine, and one that just checks text files, which requires Windows 95 (or later). The hard disk requirements for the Checker are about 10Mb, plus 5Mb or so for the indexes on the .dbfs (which are optional).
We have processed all last year's files through the prototype checker. The failure rate was over 90%! Here is a list of some of the common errors and problems with last year's submission files:-
a) Last year, we allowed Graders to make up their own event code. Unfortunately, many of your independently came up with the same code, which could have lead to confusion of events. We believe we trapped them before they overwrote each other. Hence, to avoid a recurrence this season, event codes must be taken out from Roger Edwards.
b) There was a problem last year with generation of multiple references for the same new player. Therefore this season data will not be accepted unless local Graders use the issued new player references.
c) Many Graders have not included #FINISH at the end of their files, which causes the central Grading software to loop. There have been several other minor keying faults that have caused problems with the system. It is not sufficient for the Grading administrator to add the missing #FINISH centrally and process the file; because otherwise we have no check against files truncated in transmission. So the file must be referred back to the local Grader for a correct submission.
d) There have been several cases of missing #'s causing data overflow and the system to hang. Even experienced Graders have fallen into this trap. These, together with spelling mistakes in the field names, are the major cause of complete failure to process.
e) Several Graders seem to have a Y2K problem judging by the number of events that took place in 1900 & the number of players born in 2087. Correct submission requires four-digit years; failure to supply them is the most common error in the submission files.
f) After the four-digit year problem, the most common error is failing to properly specify the club. Either a CLUB CODE, or both a CLUB NAME and a CLUB COUNTY is expected.
g) Incorrect check digits on the BCF Grading reference.
h) Giving a result for a PIN that does not occur in the player list or duplicating a PIN number within the player list.
i) Giving the same value for PIN1 and PIN2 (which suggests that the player played himself) - some of these may have been misguided attempts to show defaults/byes/half point byes.
j) Failure to give details of results officer.
k) Messing up the time limit details - in any number of different ways.
m) Incorrectly filling in the name of the player. In particular, where separate fields are used for forename, initials and surname. A player called Alan Brian Charles Green has a forename of 'Alan' (not 'Alan BC') and initials of 'ABC' (not 'BC')
The data checker will detect all of the errors described in paragraphs. "c" to "m" and I would ask you all to use the data checker when you get it.
If anyone wants any further information, or has any questions please contact me.
C E Majer Acting Director of Grading 04/01/01