Re: Notes on the tab after the Cork IV, including a to-do list.

[Klaas van Schelven]

Hi all,

On a general note:
this is a lot.... especially given that we're running out of tournaments that we can do live tests on. Luckily most of the mentioned changes are unlikely to cause unforeseen disruptions. I'll do as much as spare time and a social life allow me... it would be truly great if Meir, Deepak and possibly others would jump in.

I have one more tournament coming up next weekend. This tournament will probably only run on a few of the below mentioned changes. Anyone else? Who is willing to use some political power here?

On Dec 3, 2007 12:48 AM, Ciarán Lawlor <ciaran....@gmail.com> wrote:

Hi guys,

I'd like to start by thanking Bob Nimmo for agreeing to run Tabbie 1.4 at the Cork IV. If we get this tab to a satisfactory position, it will be largely thanks to his generosity in terms of the time it takes to play with and familiarise oneself with a tab, and his commitment to finding a tab that can be perfect for Worlds for years to come.

Hear hear. I am extremely glad with this extremely well worked out feedback, truly the first in its kind.

[...]

STUFF BOB & I WOULD LIKE:

1) Can Team Position (1G, 1O, etc.) be included on the "Venue by Team" sheet from the print page.

Yes
 

2) Can the "Venue by Judge" sheet from the print page and the new "Venue by Team" sheet described at 1 be combined into a more compact sheet, perhaps featuring two columns, to be called a "floor manager sheet".

I think it's relatively clear why these are desirable. Any questions let me know.

Do you mean: same information and organisation of information ( i.e. 2 alphabetically sorted lists), but on one paper?
Can be done of course.

3) A display should be accessible that clearly indicates which ballots have not yet been entered, including the venue name and location, chair judge in the room, and the average points of the teams in that room.

[..]

this can be done.

4) On the judge input "start-page", can the judges be sorted by rank.

This would make it considerably easier to edit the judges rankings accurately and to make sure we have the right people on the right scores, both in absolute terms and relatively.

this can be done...

5) Can the Manual Draw page appear in order of average team points for the room.

I think this has been done correctly on the initial draw page, but when you enter the Manual draw page the rooms appear by venue which is very frustrating

can be done...

6) Re the hyperlinks on all the team names, Bob and I were wondering what these were supposed to point to.

At present they don't point to anything - may actually be an error page, can't remember - but we both agreed that this could have a benefit in terms of making some information about teams easily accessible.

they are supposed to point to something of this form:
http://www.smoothtournament.com/showcase/ubciv/team_overview?team_id=37

I would love to see screenshots if you experience problems with this screen.
 

7) When finalised, can the draw display be in keeping with the order in which the ballots are printed.

This would allow the paranoid amongst us to check to make sure we're not going to hand the ballots out to Location floor managers with one missing. They should continue to print by grouped by Location.

grouped by location it is....

8) Could there be multiple user access levels?

So there would be an admin level which would have all the privileges of the current system and  a data entry user, that would have access  only to an amended input section that would not allow them Finalise the Results or input motions so that one of the Admin team would have to do this. This would have obvious advantages in terms of ensuring a data entry person doesn't mess anythign up by finalising something they shouldn't, but  would also tighten security around sensitive information about who was top of the speaker tab, other standings, future motions, and motion bias data.

this is something that can not be done before Worlds. Maybe counter-intuitively for non-engineers, a feature like this actually enlarges the scope of the project enormously. (say by a factor 1.5). It ranks high on my personal list, but a lot of work is needed for this to be possible...

9) would it be possible to have the facility to link a particular team or judge to a particular venue permanently.

This would be great if we were to have any participants with any disabilities or mobility issues - which we did in Cork!

With the given structure this is quite hard.... I'll keep it in mind for later versions.

STUFF BOB & I FEEL IS ABSOLUTELY ESSENTIAL (EASY STUFF FIRST):

1) Review the Display creation, and make the text in the display much bigger.

At Cork clicking on the create display resulted in a 404 error. At Dublin it worked fine so presumably the problem arose from the addition of the auto scrolling functionality. At Dublin the display was very small for people to read.

I'd love to see screenshots.
Deepak - any ideas?

2) During speaker name entry, it must be possible to enter non-letter characters.

[WARNING: TAB NERD CONTENT!] During the registration process for both IVs in Dublin and Cork names were entered which included apostrophes, e.g. O'Neill, at the speaker name stage.  This resulted in the Tab generating an error message which indicated that such characters could not be entered. So far so good, except when you tried to re-enter the data for that team without the apostrophe the tab refused to let you because it said that team name already existed. Next, we attempted to find the team on the team list and edit the names. The team did not appear on the Team list and wasn't included in the displayed total number of teams at the top of the page. We found the record for that team by changing the URL for the previously entered team but it wouldn't let us edit the speaker names because it could not find the record. What happened is relatively obvious: The tab initially created the relevant entries in the databases for the team itself but when it next tried to create those for the speakers it couldn't and didn't. So when we tried to edit them there was nothing to edit. This complicated things in a number of ways - the tab didn't keep a record of the speaker scores for those speakers and we had to enter them to the tab later;  the names didn't appear on the ballots; and, most particularly, when a human error was made during data entry in that team's speaker score which caused a tie in the round, it wasn't possible to edit the speaker scores because they hadn't been stored anyway, and we had to change the scores of another team and record the change for inclusion on the speaker tab and team tabs later. As I type this, I have just come up with a nifty workaround that might have been better - make that team inactive after the initial problem when creating the team in the first place, but it would have meant that there would be no Limerick C on the tab, which is less than ideal. And it would be nicer if we could just include people's names properly in the tab!!

Agreed - this is a bug.
Also agreed on the "better workaround" - though I can see why you didn't think of that in the stressful circumstances of the tournament itself.

3) Is it possible to conflict teams specifically as well as an entire institution?

If not, can this be added?

Yes. Add a dot between university and team code (this has by now been documented). like so:

CAM.A,OXF,UTR.B

4) The editable judge allocation parameters didn't work

They just didn't do anything to judge allocation. After a bit of fiddling around trying to make it work we noticed that every time we returned to the settings page we discovered that the figures had been reset to the default. We also noted that this was the case regardless of whether we pressed the save button or not. In fact, pressing the save button seemed to cause an immediate return to the default settings. Can this be fixed fast so we can do some testing with it? Are there any live IVs left where we can test this aspect of the programme in real conditions?

I'll look into this.... the problem seems to be related to Windows / Tabbie All in One.
I think Deepak is a Linux-user, so it shouldn't occur on worlds.

5) THIS ISSUE IS HUGELY IMPORTANT: We want a complete review of the debate creation part of the draw algorithm - below i describe a situation which occurred with the draw for round three of Cork which appeared unacceptable to both Bob and me. As a result I would like a thorough investigation of what the principles underlying the algorithm are and what is the exact nature of any "randomness" within the tab.

[WARNING: TAB NERD CONTENT - BUT YOU NEED TO READ IT ANYWAY - I'LL TRY TO KEEP IT CLEAR]

So, I want to prefix my remarks by saying that the "silverline" part of the tab which allocates team position (1G, 1O, etc.) again worked perfectly, in providing perfect rotation for all teams over 5 rounds. This is great news!

So what happened: The competition had 80 teams, so after 2 completed rounds there was a "pool" of 5 teams on 6 points, 10 teams on 5 points, etc. In order to draw the next round 3 teams from the 5 point "pool" needed to be added to the 6 point "pool" at random - this is commonly referred to as "pulling up" - so that we would have group of teams divisble by four, commonly called a "bracket". Let us call the teams on 6 points 6A 6B 6C 6D and 6E and we call the teams on 5 points 5A, 5B, 5C, 5D, etc. where the letters are of NO significance beyond identifying the teams (6A is not necessarily the best team of those on 6 points). This situation arose in circumstances where we were drawing the third round repeatedly in an  attempt to assess how our adjustments to the adjudicator parameters would effect allocation of judges. and, then the draw did the following. The data for the tournament had been saved post round 2 result entry pre our fiddling.

We drew the round twice and, having got over the fact that the judge allocation wasn't changing, also noted that on both occasions the draw had been:

Room 1: 6A 6B 6C 6D

Room 2: 6E 5A 5B 5C

Initially, we checked to see if this was some freak occurrence by redrawing the round using the clean data for the tournament that had been saved post round 2 result entry and pre our fiddling.

The draw for the top two rooms was again exactly the same. We repeated this exercise a few times and found the same draw repeatedly. Next we checked to see if we could identify whether these teams were being pulled up on speaks - they were not.  So the first question is why did Tabbie always pull up the same 3 teams from the 5 point pool to create the top bracket?

This is very important.

Yes, it is.
 

[...]

ok, nerd warning still stands....

The draw *is* in fact random  - the only thing is that the 'randomness' is fixed (by design) into the program. (this is done by initializing our pseudo-random number generator with a fixed number). The reason for this is twofold:
* It is not desirable that the CA or whoever it is clicks the "redraw" button a number of times until he gets what he perceives to be a "fair draw". No re-throwing the dice.
* It is immensely useful in matters of debugging etc. that doing the draw again yields the same result.

Now, if you say this is undesirable, it can be changed (though I'm not sure I would agree that would be better).

You also (implicitly) raise another issue: the distribution of pull-ups among the debates in a bracket.

which is better:
[6A, 6B, 6C, 6D]
[6E, 5A, 5B, 5C]

or
[6A, 6B, 5A, 5B]
[6C, 6D, 5C, 5D]
?

In fact, as far as I know, WUDC rules don't say anything about this....

I've given it some thought before and then decided to 'just go with the rules', esp. since they specifically specify that it is strictly forbidden to take anything into account but randomness and a couple of explicitly specified elements for the powerpairing. The above "fairly distributed pull up" is not one of those elements.

Should it be?
In the "undistributed scenario" (the first) the teams 6A - D are "pullup neutral" - they don't get disadvantaged or advantaged based on how many points they already have.

6E gets an advantage of 3 weaker teams. 5A - C get a disadvantage of 1 "too strong" team: total unfairness: 6 (of which 3 is concentrated in one team)

in the "distributed scenario" (the second) 6A and 6B get an advantage of 2 each, 5A and 5B get disadvantaged by two teams. same for the other debate. Total unfairness: 16 - albeit that the unfairness itself is more fairly distributed (+/- 2 per team)

Tabbie prefers neither. Anyway.... for now it's not in the rules (as far as I know)

I'm open for arguments.

Klaas

[Steve McCarthy]

> 8) Could there be multiple user access levels?* ...

>
> this is something that can not be done before Worlds. Maybe
> counter-intuitively for non-engineers, a feature like this actually enlarges
> the scope of the project enormously. (say by a factor 1.5). It ranks high on
> my personal list, but a lot of work is needed for this to be possible...

Agreed as far as the program is concerned. A couple comments, though.

First from a tabsperson's perspective. At Vancouver Worlds, we had
three tabs people (one of which the tabs director); having more tabs
people for data entry wouldn't have made the tourney significantly
faster as far as I could tell. Even if you have a tournament of that
size, it's probably not a big deal to find a couple trustworthy
people...

But I like the general principle of multiple access levels, so from a
programmers perspective: putting this in the program itself would
indeed be difficult - as there is no protection in the code. Taking
the smoothtournament.com as an example, it uses the server's
authentification methods, correct? A server workaround could similarly
exist for a "data-entry" user account, who would only be able to
access one or two pages. I don't know exactly how smoothtournament is
set up, but this seems to me to be theoretically possible, and because
it wouldn't play around [too much] with the code, wouldn't expand the
magnitude of the project.

Of course, if you're running your own installation (as I imagine
Worlds will do), you would have to set all this up yourself, so it's
not much of a permanent solution as it is a workaround.

Just my two cents.

Stephen


On Dec 2, 3:52 pm, "Klaas van Schelven" <klaasvanschel...@gmail.com>
wrote:

> Hi all,
>
> On a general note:
> this is a lot.... especially given that we're running out of tournaments
> that we can do live tests on. Luckily most of the mentioned changes are
> unlikely to cause unforeseen disruptions. I'll do as much as spare time and
> a social life allow me... it would be truly great if Meir, Deepak and
> possibly others would jump in.
>
> I have one more tournament coming up next weekend. This tournament will
> probably only run on a few of the below mentioned changes. Anyone else? Who
> is willing to use some political power here?
>

> On Dec 3, 2007 12:48 AM, Ciarán Lawlor <ciaran.b.law...@gmail.com> wrote:
>
>
>
> > Hi guys,
>
> > I'd like to start by thanking Bob Nimmo for agreeing to run Tabbie 1.4 at
> > the Cork IV. If we get this tab to a satisfactory position, it will be
> > largely thanks to his generosity in terms of the time it takes to play with
> > and familiarise oneself with a tab, and his commitment to finding a tab that
> > can be perfect for Worlds for years to come.
>
> Hear hear. I am extremely glad with this extremely well worked out feedback,
> truly the first in its kind.
>
>
>
> > [...]

> > *STUFF BOB & I WOULD LIKE:*
>
> > *1) Can Team Position (1G, 1O, etc.) be included on the "Venue by Team"
> > sheet from the print page.*
>
> Yes
>
> > **

> > 2) Can the "Venue by Judge" sheet from the print page and the new "Venue
> > by Team" sheet described at 1 be combined into a more compact sheet, perhaps
> > featuring two columns, to be called a "floor manager sheet".

> > *
> > *

> > I think it's relatively clear why these are desirable. Any questions let
> > me know.
>

> Do you mean: same information and organisation of information (i.e. 2

> alphabetically sorted lists), but on one paper?
> Can be done of course.
>
>
>

> > *3) A display should be accessible that clearly indicates which ballots

> > have not yet been entered, including the venue name and location, chair

> > judge in the room, and the average points of the teams in that room.*

>
> > [..]
>
> this can be done.
>
>
>

> > *4) On the judge input "start-page", can the judges be sorted by rank.*

>
> > This would make it considerably easier to edit the judges rankings
> > accurately and to make sure we have the right people on the right scores,
> > both in absolute terms and relatively.
>
> this can be done...
>
>
>

> > *5) Can the Manual Draw page appear in order of average team points for
> > the room.*

>
> > I think this has been done correctly on the initial draw page, but when
> > you enter the Manual draw page the rooms appear by venue which is very
> > frustrating
>
> can be done...
>
>
>

> > *6) Re the hyperlinks on all the team names, Bob and I were wondering what
> > these were supposed to point to.*

>
> > At present they don't point to anything - may actually be an error page,
> > can't remember - but we both agreed that this could have a benefit in terms
> > of making some information about teams easily accessible.
>
> they are supposed to point to something of this form:http://www.smoothtournament.com/showcase/ubciv/team_overview?team_id=37
>
> I would love to see screenshots if you experience problems with this screen.
>
>
>

> > *7) When finalised, can the draw display be in keeping with the order in
> > which the ballots are printed.*

> > This would allow the paranoid amongst us to check to make sure we're not
> > going to hand the ballots out to Location floor managers with one missing.
> > They should continue to print by grouped by Location.
>
> grouped by location it is....
>
>
>

> > *8) Could there be multiple user access levels?*

>
> > So there would be an admin level which would have all the privileges of
> > the current system and a data entry user, that would have access only to
> > an amended input section that would not allow them Finalise the Results or
> > input motions so that one of the Admin team would have to do this. This
> > would have obvious advantages in terms of ensuring a data entry person
> > doesn't mess anythign up by finalising something they shouldn't, but would
> > also tighten security around sensitive information about who was top of the
> > speaker tab, other standings, future motions, and motion bias data.
>
> this is something that can not be done before Worlds. Maybe
> counter-intuitively for non-engineers, a feature like this actually enlarges
> the scope of the project enormously. (say by a factor 1.5). It ranks high on
> my personal list, but a lot of work is needed for this to be possible...
>

> > *9) would it be possible to have the facility to link a particular team or
> > judge to a particular venue permanently.*

>
> > This would be great if we were to have any participants with any
> > disabilities or mobility issues - which we did in Cork!
>
> With the given structure this is quite hard.... I'll keep it in mind for
> later versions.
>
>
>

> > *STUFF BOB & I FEEL IS ABSOLUTELY ESSENTIAL (EASY STUFF FIRST):*
>
> > *1) Review the Display creation, and make the text in the display much
> > bigger.*

>
> > At Cork clicking on the create display resulted in a 404 error. At Dublin
> > it worked fine so presumably the problem arose from the addition of the auto
> > scrolling functionality. At Dublin the display was very small for people to
> > read.
>
> I'd love to see screenshots.
> Deepak - any ideas?
>
>
>
>
>

> > *2) During speaker name entry, it must be possible to enter non-letter
> > characters.*

> > *3) Is it possible to conflict teams specifically as well as an entire
> > institution?*

>
> > If not, can this be added?
>
> Yes. Add a dot between university and team code (this has by now been
> documented). like so:
>
> CAM.A,OXF,UTR.B
>
>
>

> > *4) The editable judge allocation parameters didn't work*

>
> > They just didn't do anything to judge allocation. After a bit of fiddling
> > around trying to make it work we noticed that every time we returned to the
> > settings page we discovered that the figures had been reset to the default.
> > We also noted that this was the case regardless of whether we pressed the
> > save button or not. In fact, pressing the save button seemed to cause an
> > immediate return to the default settings. Can this be fixed fast so we can
> > do some testing with it? Are there any live IVs left where we can test this
> > aspect of the programme in real conditions?
>
> I'll look into this.... the problem seems to be related to Windows / Tabbie
> All in One.
> I think Deepak is a Linux-user, so it shouldn't occur on worlds.
>
>
>
>
>

> > *5) THIS ISSUE IS HUGELY IMPORTANT: We want a complete review of the

> > debate creation part of the draw algorithm - below i describe a situation
> > which occurred with the draw for round three of Cork which appeared
> > unacceptable to both Bob and me. As a result I would like a thorough
> > investigation of what the principles underlying the algorithm are and what

> > is the exact nature of any "randomness" within the tab.*

> > on speaks - they were not. *S**o the first question is why did Tabbie

> > always pull up the same 3 teams from the 5 point pool to create the top

> > bracket?*

[Klaas van Schelven]

good point

Smooth Tournament is actually pretty much of a "permanent workaround"
I've got some automated scripts that basically just toss out code (for the displayed tournaments, the input code is thrown out)
The rest works with http auth.

Klaas

[Klaas van Schelven]

On Dec 3, 2007 12:23 PM, Ciarán Lawlor <ciaran....@gmail.com> wrote:

Just to reply on a few things, and in particular on the very last, and most important, points: 

Re: the "randomness" for each team being hardwired:

Kaas can you clarify whether: 

a) each team is hard-assigned a new random number after each round which simply is not changed upon redrawing; or

b) each team is hard-assigned a random number at the start of the tournament which is their random number for the duration of the tournament?

If it is the latter, then this means that there will be a team with the highest number, who in any pull up situation, as against any other team will always be pulled up.

I'm opposed to this because as between two teams, before the tournament starts, decided at random, one team will always be pulled up and another always not. This is not fair, and why I think the randomness must be decided at each pull up stage.

If I can explain more clearly, I'm not happy if before round 1 starts we can put teams in order of who is most likely to be pulled up when facing one another, so that we can say Team X will always be pulled up because it has the highest random number. I accept that this is just the "luck of the draw" but I'm arguing that the luck should have a chance to balance out.

One problem that the designers have been trying to avoid is one where the someone keeps redrawing the round to get "a fairer draw". Obviously this should not happen but a solution might be to assign a new random number to each team at the end of the "Finalise Results" process for the preceeding round.

I agree.
For now I don't have the time to explain how this works - just that no team is advantaged or disadvantaged throughout the whole tournament.

Re: the allocation of teams to debates after pulling up - clumping v random:

I think I explicitly raise this issue and I think it's very important. Klaas you say that "In fact, as far as I know, WUDC rules don't say anything about this". I think you might be wrong here. The rules do explicitly say that teams should be pulled up. Now, if the Tab prefer to keep the teams on six together then effectively, you are not pulling up - you are pulling a team on six down to meet the fives and that would not be compliant with the Worlds rules. Does that make sense?

This is my source - correct me if I'm wrong... http://tabbie.wikidot.com/local--files/wudc-rules/Worlds_Tab_Rules_-_DRAFT.doc
It just speaks of "pools".... the fact that 3 teams of 5 (in our previous example) are pulled up does nothing to the fact that they are "pulled up" into the pool of teams on 6 points. The distribution of them after this is subject to the other rules...
That's how I read it.... anyone else having thoughts about this?

Are you sure that this preference to keep them together isn't in there somewhere and simply overridden elsewhere by positional allocation?

I think it is more likely to happen, yes... if not overridden by any other concern (i.e. positions in the debate) this is what's most likely to happen, because of the way the algorithm is built.

In terms of whether it might better to keep those teams on five together, I'm certainly open to persuasion but ultimately it's a decision for Worlds Council - what I'm more focused on is getting a tab that is compliant with the rules ... for now anyway!

I agree, see above.

That's all I have time for now - I think everything else you're saying either looks great or we will survive without it. 

Thanks to both yourself and Deepak for all your help with this,

Many thanks to you and Bob as well...

Klaas
 

Ciarán.

_______________

Ciarán Lawlor

Thorntown,

Kilsallaghan,

Co. Dublin.

E-mail: ciaran....@gmail.com

Phone: +353 87 1061101

Skype: ciaranlawlor

[Klaas van Schelven]

Hi all,

On Dec 3, 2007 5:38 PM, Ciarán Lawlor <ciaran....@gmail.com> wrote:

Article 3.D of those rules reads: "Once the pools have been adjusted in accordance with 3c) [pulling up teams at random] then the pools are divided into debates of four teams on a random basis... "

So after teams are pulled up they must be assigned to a debate on a random basis - the rules do not preference keeping any teams together on any basis - though they do allow swapping to achieve better position allocation - and in fact the requirement of randomness means that the rules prohibit them from being kept together as a rule within an adjusted "pool" (which I call a bracket :-).

Fair point.

I will adapt the algorithm in such a way that the distribution within the pools/brackets will be completely random (Deepak - please allow me a week or so to pick this one up)

I'm sorry if likening this to pulling down caused confusion. Keeping 3 teams on 5 together in the way seen at Cork creates a situation which would be achieved by pulling down: The top pool contains five teams, pull one down into the pool below and you get a room of teams on 6 points and, after dealing with the pool below, a room that must contain one team on 6 and three teams on 5 points.

Even if you don't agree with this analogy, it is undeniable that this tab's debate creation algorithim has repeatedly faced the criticism, albeit often from less well-informed or tab-savvy a group as this, that it pulls down and we may now have identified why those criticisms have been made. 

I have not heard those criticisms.... and I'm not sure if they were made about "NTU-Tabbie" or Tabbie as it has been developing in the past half year. Anyway I'm very glad you and Bob are now seriously working with us to improve the program.

To be honest, whether you agree with the analogy or not, it's clear that the tab should not prioritise keeping together teams that are pulled up. If you are correct in saying that this is what it is most likely to do (ASIDE: how do you mean "most likely" - surely it either does or it doesn't, right?) then we will have to change it before Worlds, and test it rigourously.

To answer the aside: the algorithm starts out with a "bunched together" layout. If it finds no problems with the position allocation, this is how it remains. However, from the moment it does start swapping positions around, it does not in any way try to maintain the "bunched togetherness". Therefore: "most likely".

Testing will be possible at the Israeli tournament I have coming up, as well as heavily by others (anyone out there?)

Klaas

[Klaas van Schelven]

Hi all,

The below point has now been implemented. Can be checked out in the trunk (though with current heavy development this may be a bit unstable for most of you)

Deepak - will you make sure it gets some testing?

Klaas

[Deepak Jois]

Hi,
I would like to go over some of the issues in the section titled - STUFF BOB & I FEEL IS ABSOLUTELY ESSENTIAL. Comments inline

On 12/3/07, Ciarán Lawlor < ciaran....@gmail.com> wrote:

STUFF BOB & I FEEL IS ABSOLUTELY ESSENTIAL (EASY STUFF FIRST):

1) Review the Display creation, and make the text in the display much bigger.

At Cork clicking on the create display resulted in a 404 error. At Dublin it worked fine so presumably the problem arose from the addition of the auto scrolling functionality. At Dublin the display was very small for people to read.

I am not sure why you got a 404 error. Correct me if I am wrong, but I assume you are talking about the scrolling page which is accessed by Accessing the 'Print' menu on top and the clicking on the link below 'Scrolling Display (for overhead projection)'.

I have made some minor adjustments to the display

Could you do the following things the next time you are testing on any computer.
1. Use Firefox Browser. We already have so many different screen resolutions and platforms that dealing with the (rather crappy if I may add) Internet Explorer 6 and 7 makes live even more hellish! It can be downloaded for free from http://getfirefox.com and is (arguably) much better.

2. View the scrolling display in Full screen mode. You can do that by going to View -> Full Screen. This works on Windows and Linux.

3. You can also adjust the default font size, because the display picks up a base font size from the browser defaults. I will post detailed instructions later.

Send me a screenshot of how it looks in your browser and I will look at further tuning the display.

 

2) During speaker name entry, it must be possible to enter non-letter characters.

[WARNING: TAB NERD CONTENT!] During the registration process for both IVs in Dublin and Cork names were entered which included apostrophes, e.g. O'Neill, at the speaker name stage.  This resulted in the Tab generating an error message which indicated that such characters could not be entered. So far so good, except when you tried to re-enter the data for that team without the apostrophe the tab refused to let you because it said that team name already existed. Next, we attempted to find the team on the team list and edit the names. The team did not appear on the Team list and wasn't included in the displayed total number of teams at the top of the page. We found the record for that team by changing the URL for the previously entered team but it wouldn't let us edit the speaker names because it could not find the record. What happened is relatively obvious: The tab initially created the relevant entries in the databases for the team itself but when it next tried to create those for the speakers it couldn't and didn't. So when we tried to edit them there was nothing to edit. This complicated things in a number of ways - the tab didn't keep a record of the speaker scores for those speakers and we had to enter them to the tab later;  the names didn't appear on the ballots; and, most particularly, when a human error was made during data entry in that team's speaker score which caused a tie in the round, it wasn't possible to edit the speaker scores because they hadn't been stored anyway, and we had to change the scores of another team and record the change for inclusion on the speaker tab and team tabs later. As I type this, I have just come up with a nifty workaround that might have been better - make that team inactive after the initial problem when creating the team in the first place, but it would have meant that there would be no Limerick C on the tab, which is less than ideal. And it would be nicer if we could just include people's names properly in the tab!!

Ok, the cause of this problem is now identified. This should be fixed. It was a problem specific to the Windows All in One version.

3) Is it possible to conflict teams specifically as well as an entire institution?

If not, can this be added?

As Klaas mentioned earlier it is possible already.
 

4) The editable judge allocation parameters didn't work

They just didn't do anything to judge allocation. After a bit of fiddling around trying to make it work we noticed that every time we returned to the settings page we discovered that the figures had been reset to the default. We also noted that this was the case regardless of whether we pressed the save button or not. In fact, pressing the save button seemed to cause an immediate return to the default settings. Can this be fixed fast so we can do some testing with it? Are there any live IVs left where we can test this aspect of the programme in real conditions?

This is fixed now

5) THIS ISSUE IS HUGELY IMPORTANT: We want a complete review of the debate creation part of the draw algorithm - below i describe a situation which occurred with the draw for round three of Cork which appeared unacceptable to both Bob and me. As a result I would like a thorough investigation of what the principles underlying the algorithm are and what is the exact nature of any "randomness" within the tab.

Klaas has made some changes after extensive discussion.

Klaas, may I request you to make another release just for testing the critical points that were raised? We can keep making the other improvements on trunk.

Let me know.

Deepak