In the Tournament Organization Forum, someone asked when we will be able to submit Tournament Reports on-line. Nolan answred “see my posting in Tournament Direction.” But can’t find it. A good question needs an answer.
Jim
What you are looking for is located at uschess.org/ratings/computer … stemupdate.
The underlying problem is that Tournament Administrator is an ancient DOS program, written on a 386 when dinosaurs roamed the earth. A modern programmer, with greater resources available, would design it to assign individual long filenames to the output files, and perhaps put them in a zipfile. TA (and all the pairing programs that scabbed off it) generates a rating report of three files, always with the same names. So unless the files are on a disk in physical proximity to the crosstable, it is extremely likely that something will go wrong. Fixing this would require either a) writing a new rating-report program (which would require replacing the established base of pairing software), or b) creating an interface to accept the existing files, with adequate error-checking. It sounds like the USCF is trying for the second, but I’m not holding my breath.
Thanks for posting that Link, John. I’ve posted comments in a few topics here, but we should probably have one topic specifically on the update process, including suggestions and comments.
The current plan is to create a NEW reporting format, one that has more information in it than the current Tournament Administrator format. WinTD and Swis-Sys currently support that format; Tom and Thad (the authors) have indicated they will upgrade their programs to support a new format once we have it ready and a way to test it.
The MIS Committee will be involved in that process, I’m sure.
Will we continue to support Tournament Administrator? I would prefer not to, because it is an INADEQUATE and poorly designed format.
Among other things, the TA format does not deal well with multiple section events, especially ones with different start/finish dates and with multiple TD’s working those events. (The latter may not be important for the ratings process, but the tournament crosstables are also used to derive TD history for advancing to higher levels of TD certification.)
The new, expanded format will also include updated or new membership information, something that IS generally needed in order to rate an event.
The online ratings process will probably work something like this:
-
The TD and/or affiliate will have to provide validation information,
either similar to the Members Only login screen or something more
secure. -
There will be an upload feature for new format files prepared by
WinTD, Swis-Sys and any other 3rd party programs that may be
out there. -
There may also be a web-based entry form, though I suspect it will
only really be useful for small events, say those under 12 players. -
There will need to be some kind of payment mechanism for
memberships and ratings fees. I could see
there being three options:
A Payment by credit card
B Deposit accounts to be charged against
C Mailed in checks. (This option would probably delay processing
until the checks arrive in NY.) -
Memberships sent in (and paid for) would be processed immediately,
though under some circumstances there might need to be a review
of a new member’s record if it appears that it might be a duplicate
of a current or former member. This gets sticky, because federal
privacy laws may prohibit us from doing the logical thing, sending
the list of potential duplicates back to the TD and letting him
decide if it is a duplicate or not. -
Even if memberships are ‘pending’, either because of payment
issues or the duplicate ID issue, it should be possible to do a
preliminary validation run on the event. This would make sure
the crosstable balances (if Jones played Smith, then Smith has to
have played Jones, with consistent results entered for both players)
and would also check for potential ID problems. -
If there are problems or potential problems with the event, the TD
would be sent a validation report. It may be possible to generate this
report immediately, while the TD is still online, otherwise we would
e-mail it to the TD. (That’s why we’re asking TD’s and affiliates to
give us their e-mail addresses, with a form for that purpose in the
Members Only area.) -
I’m not sure yet how the TD would respond to the validation report.
One possiblity would be to upload a corrected file for the entire event,
or for specific sections. (That’s probably an area that will need to be
coordinated with the pairing program authors.) -
Once an event has passed the TD validation process, there may be
other validation steps, but hopefully for most events it will be possible
to proceed to the ratings stage without requiring manual intervention
by the office. Note that events would NOT be rated until the checks
arrive and until the funds balance with the dues and fees owed.
When there are money or data issues delaying an event, that will
be noted on the event record, which is accessible online. (There will
probably be an update to the ‘Tournaments Received’ list soon that
includes several additional pieces of information we are collecting on
events these days, including the affiliate ID, the chief TD, and whether
or not the event is being delayed because of a need to contact the
TD.) -
Once rated, the TD would (or could) be e-mailed a copy of the
completed report. Several people have suggested that we also
e-mail the players involved in an event, to let them know it has been
received or rated. I’m generally supportive of this, though I think
it would be best to wait until the event is rated.
That’s probably enough for this post. I’ll post another message on the timing aspect of rating (and re-rating) events, that one probably deserves its own topic.
Tom and Thad (the authors) have indicated they will upgrade their programs to support a new format once we have it ready and a way to test it.
Depends on whether you’re talking about a .1 upgrade (free), or a new version (for which the users, not the USCF, will have to pay). The latter would be grossly unfair to the members. This proposal strikes me as another example of the “do it now, think it through later” attitude that created the current problems.
You’re ignoring the primary problem, the current format is INADEQUATE and antiquated. (When was the last time you ran across a dBase III file?)
It does not contain any fields to help cross check the name against the ID, it does not contain all of the information needed for each section, it does not contain membership information, it does not contain the information needed to tie out the money to be remitted for an event (dues and fees). It doesn’t deal well with non-swiss events (such as a double round robin) and I’m told there are problems with reports from some of the team events, though I’m not sure what they’re referring to. It does not support events longer than 20 rounds, though those are rare.
Any reporting format that does not support FULLY AUTOMATED processing of a tournament, including memberships, does not achieve the goal of getting the office staff out of the loop (except for supervisory review), and thus does not achieve the cost savings through improved efficiency.
My guess is that we will have to continue to support the TA format, but I don’t think it would be appropriate to offer reductions in the ratings fee for the old format, since it will not produce the cost and time savings we need and that a new format will facilitate.
Since the format has not been developed yet, I don’t know how the pairing program authors will deal with it, they may be able to do a point upgrade for their current versions but I don’t know how many TD’s are on older versions. (I"m one of them, I still use Swis-Sys DOS version 8 for the one tournament I run each year.)
It might be possible to use a web-based input form to put a wrapper around the TA format to deal with some of its weaknesses, but that would probably not be the first option offered or implemented.
I’m not even sure yet what type of format it will be. I’ve had conflicting advice on this point, some favor an XML-type format, others favor a tab or comma delimited format, possibly something that could be produced by a spreadsheet program like Excel. It probably won’t be a dBase-type format.
The TA format was developed without any significant input from TD’s or from the MIS committee. Neither of those mistakes will be repeated this time, though the final result will almost certainly not make everyone completely happy.
We can agree to disagree on this. The current system works … adequately … as proven by the fact that is has done the job for ten years. Maybe it doesn’t work as well as we’d like, but the world isn’t perfect. The assertion that what you want to replace it with will work (better) is just that – an unproven assertion. This is the same sort of thinking that brought us the TA system and rating reports on disk, long before the available technology was up to the job. When is doubt, hesitate.
The technology to submit events via e-mail was probably available in 1991, but the USCF office didn’t even have e-mail access until several years later. I think other competing rating services have already demonstrated that web-based submission is quite feasible today. Our ratings system is techologically obsolete, and has been for years.
The ratings programs have never been updated to try to utilize direct e-mail submissions, either. Given their design and implementation, that was probably not very practical.
Getting the membership and ratings updates transferred between the old systems and the new ones is a process that would have delighted Rube Goldberg. No wonder it breaks down every few days.
As a result of organizational gridlock, the USCF is still running on ratings software mostly written in 1991 (and not very expertly designed or written), software that is highly inefficient and very labor-intensive to use, difficult to maintain, and on hardware that was purchased at least five years ago and a networking architecture (Novell 3.1) that was obsolete even then.
The rewrite of the ratings system is YEARS overdue. Let’s not procrastinate it further just because chessplayers are loath to embrace change.
Undoubtedly the technology existed, but, as you pointed out yourself, it was not widely enough available to be of practical use. We’re talking about a time when you might be able to run Windows 3 (though why anyone would want to escapes me) if you could afford three or four grand for a 486.
The USCF decided that submission of rating reports on disk would be a neat idea. The DOS-based pairing programs of the day did a dreadful job. TA was enormously clumsy to use, and there is no reason to think that TDs did a better job entering results than USCF staff. (Note that Tournament Administrator was simply a front end for the old rating software, which goes back to the late 70’s.)
The hardware and software to do the job adequately became widely available in the late 90’s – let’s say around the shipping of Win 98, which coincided with the appearance of Pentium systems for under $1000. If TA had been written and distributed at that time, we would undoubtedly have a better system in place now.
If I understand what you’re aiming for, you want an on-line submission system which is fully automated – the rating report goes directly from the TD to the rating program without any staff intervention. I have strong doubts that this is possible. As I’m sure you will agree, the most common source or error in rating reports is operator error. If the error filter on your front-end is too low, it obviously won’t work. If you crank it up high enough to prevent errors, it will kick back many, many reports to the TD. You’re assuming that the users will figure it out. I’m assuming they won’t.
I agree that such a system could be made to work with a sufficiently computer-savvy user base. But you don’t have that and you’re not going to. (I speak from many frustrating telephone conversations trying to explain where the menu bar is.)
Since you are proposing a major change, it is incumbent on you to demonstrate that a) it is an objective improvement, b) it is practical to implement, and c) your customer base wants it.
“Note that Tournament Administrator was simply a front end for the old rating software, which goes back to the late 70’s.)”
I don’t believe that is a correct statement. The original ratings programs written back around 1978 were in COBOL, I’ve still got a printout of them. In 1990-91 George Wang rewrote the ratings system using a dBase III file structure and Clipper (a DOS programming language.)
The TA program was written by Jim Meyer in 1991, and the file structure it uses was designed as an input into the dBase/Clipper system. There was never a way to import data from it into the COBOL rating system.
BTW, as I understand it, George also got started on a rewrite of the membership and sales systems in Clipper, but for some reason the plug got pulled on that project. (I think George Fillipone did that during his time as the interim ED.)
Those are the programs that are still in use today. (The COBOL programs are also still in use, though these days with the new membership system and the demise of B&E I think the only thing they’re being used for is correspondence ratings.)
I was using e-mail to automate records management in 1989, that the USCF did not do so in 1991 or subsequently says more about their approach to using technology than it does the state of the technology itself.
Okay, are we talking about casual users who may run a tournament once a year or so, or those who do big and small tournaments several times a year? And as a counterpoint, does anyone have any idea of the number of TDs using TA?
My point is, most of the regular TDs are using something other than TA to do their pairings. And given what I’ve experienced with these programs, you have to be (or you become) handy with a keyboard and mouse, not to mention the OS and a program that has what seems like several hundred options in it. So what Mr. Nolan is proposing may mean people take awhile to adjust to a new way of doing things, but it’s a step in the right direction, instead of being stuck in one place just because some people have trouble learning.
Radishes
About a year ago I did an analysis of events submitted during the 2002 calendar year.
I was able to match up about 87% of the events on the excel spreadsheet being used to input tournaments received with an event in the crosstable history.
Around 80% of all ‘scholastic’ events were being submitted on diskette, around 65% of ‘non-scholastic’ events were being submitted on diskette.
Most of the events being submitted on paper were small, e.g., quads.
For events with 100 or more players in them, most (probably 95-98%) were being submitted on diskette.
For TD’s submitting more than 10 tournaments in a year, all but a few of them were submitting most or all of their events on diskette. (There was one TD who apparently submitted over 25 events in calendar 2002, all on paper.)
These trends appear to have remained relatively constant in the last six months. For events logged in using the new tournaments received log, implemented last October, over 2100 were on diskette, around 1300 were on paper.
My point is, most of the regular TDs are using something other than TA to do their pairings. And given what I’ve experienced with these programs, you have to be (or you become) handy with a keyboard and mouse, not to mention the OS and a program that has what seems like several hundred options in it.>>
A couple of months ago, I was chatting with one of our leading TDs (someone who directs dozens of tournaments every year). I asked why he was using the horrible DOS version of SwissSys (the one that produces wallcharts in 10-pt Courier), since I knew he had access to more recent ones. He replied that he just couldn’t get the hang of importing the rating supplements every two months – the files always got lost somewhere on his disk. The Prosecution rests.
I’m the first to admit that the current versions of the major pairing programs are very complex. I played around with one of them a few months ago and found it to be very non-intuitive, and I’m hardly an unsophisticated computer user.
If I ran a lot of large tournaments and could save a lot on my ratings fees by learning to use a program that uses the new format, I’d probably take the time to learn it.
I think there’s room in the marketplace for other pairing programs, too.