ResultSetAssociatedDataAndColumnAddedDecoratedCollection. Nuff said.
05.05.06 3:30 PM
By
dplass
TopCoder Member
Here's an update from round 2 of the CDDC; the design project was XSD2DDL:
| Handle | Score |
| Pops | 96.42 |
| adic | 96.12 |
| nicka81 | 94.27 |
For the developers it was the Testing Framework.
| Handle | Score |
| sindu | 98.49 |
| biotrail | 98.19 |
| visualage | 97.76 |
Some interesting line items:
cnettel got a perfect score of 100.00 from developer reviewer
and
sindu got a 99.00 from
cucu, also in development.
While I was backstage, traugust was leapfrogged by two competitors and dropped from third to
fifth place.
As I noticed in the
forums, calling this round the "Finals" is a bit of a misnomer. It's
actually the 3rd out of 3 appeals rounds for the TCO '06. For the designers, they were appealing
DataView Decorator for C# and Result Set Collection Decorator for Java. The developers were appealing
the Bread Crumb Trail Tag for Java and the Bread Crumb Trail Control for C#. Initially the starting scores
were
very close for both design and development:
| Design | Development |
| Handle | Score | Handle | Score |
| Pops | 88.85 | cnettel | 92.41 |
| adic | 88.45 | sindu | 92.26 |
| kyky | 88.07 | colau | 91.22 |
Less than a point separated the top 3 designers, and less than 1.5 points separated the top 3 developers.
It's anybody's game at this point!
I decided to follow the appeals of a single designer (since
yesterday I looked at only the development side of the CDDC Round 2), and settled on
adic because he was in second place and I wanted to see if he could leapfrog
Pops to repeat his previous design championship wins. I started off by only looking at his screen, but there were just so many appeals coming on so many different items for so many designers, that I will have to resort to more of a summary than a play-by-play.
adic looked over his three reviews and original submission for over 15 minutes before making his first appeal. While flipping through the screens I saw that he had about 5 projects (including the TCO projects) in review and one project for which he's a reviewer. Busy guy! By 9:15,
Pops and
kyky had both already submitted appeals, but he had not.
At the beginning, there were a lot of algorithm competitors also present and watching.
reid and both
po and
ploh were trying to follow the appeals on
AleaActaEst's and
aubergineanode's screens, but since they weren't familiar with the process or the tools, it was as mysterious to them as a 1k algorithms problem is to this blogger. (This made me feel better, also supporting one of my theories that "everyone has something they're good at.")
One extremely long class name caught
reid's eye named (deep breath)
ResultSetAssociatedDataAndColumnAddedDecoratedCollection. Whew!
The general strategy that
adic followed seemed to be to open each review and appeal
everything that was not a 4 out of 4, from top to bottom. A few times, a forum post (including one made by
fellow competitor
Pops) was used to support an appeal. It seemed like he referenced his UML
and use case diagrams in
Poseidon a lot more often than he referenced the component specification.
In more than one case, a reviewer seemed to misunderstand the intention of the design, which allowed arbitrary
"mapping" between columns using a mathematic equation. Multiple times, he explained how the end-user would be
able to define intermediate columns to achieve the desired behavior. In fact, this was one of the times that
he referenced the CS, and even I could see that it clearly described exactly what the reviewer was unclear about.
During the course of the morning, the scores and placements for both design and development changed wildly.
Here is but a sample of the way things changed (all times are
PDT):
- At 9:27, adic submitted his 3rd appeal. By this time, kyky was ahead of Pops, 89.22 vs 89.00 (adic's score
hadn't changed from his original 88.45)
- adic submitted two or three more appeals, whcih were accepted.
- At 9:41, adic was in first with 90.06 points, kyky second with 89.67 and Pops in third with 89.00. All but two designers have increased in score from appeals.
- At 9:49 kyky increased to 90.02, but still in second (by just 0.04 points!)
- by 9:51, kyky was up to 90.27, the other two unchanged.
- All this time, adic was submitting appeals about every 5 minutes
- At 10:02, some responses came back and seven out of the eight designers had increased points. adic was in first with 90.31, kyky in second with 90.27 and Pops still 3rd.
- A few more appeal responses later, and kyky took over first place with 90.74, adic unchanged, and Pops had bumped up to 89.25
- Then, a whole bunch of appeal responses came in for kyky, and he went from 90.99 to 91.49 to 92.14. adic and Pops were
a "distant" second and third with 90.31 and 90.08 points respectively. nicka81 was in fourth with 89.47 (which would've put him in first place at the beginning of the appeals phase!)
- adic just widened his lead on Pops by another few tenths, to 90.64.
- At the end of the appeal phase, here are still about 20 appeals oustanding, so even though the current
standings as of 11:15 am (PDT) are as follows, anything can still happen:
| Handle | Points |
| kyky | 92.14 |
| adic | 90.64 |
| Pops | 90.08 |
| nicka81 | 89.97 |
| aubergineanode | 88.12 |
| ThinMan | 83.31 |
| real_vg | 77.45 |
| AleaActaEst | 65.09 |
I don't know about you, but this was pretty exciting, watching adic and kyky swapping places (and -- this just in -- Pops moved up to second with 91.23 at 11:31 PDT) for almost two hours! At the very beginning of the phase, Rustyoldman stopped by and said "anything exciting"? I answered "Well, it's interesting at least", but as it turned out, watching the board change every few minutes was exciting, to me at least.
Stay tuned for the final results!
With just two blog entries more to go,
— dplass
PS. The development appeals phase was pretty exciting as well. sindu got a perfect score from reviewer
cucu, who also gave near-perfect scores to cnettel (99.62) and traugust
(99.37). He also got a near-perfect (99.06) from . As of 11:33 AM PDT, the development results were: (with still 11 outstanding appeals)
| Handle | Points |
| sindu | 95.39 |
| colau | 94.24 |
| cnettel | 93.45 |
| visualage | 93.05 |
| biotrail | 92.42 |
| traugust | 90.23 |
| oodinary | 83.18 |
| zjq | 54.54 (who had zero successful appeals :-( ) |
Reviewer cucu sent me this humorous appeal and response text:
Appeal Text:
I am so so sorry for submitting this unfinished component, actually,
when the other competitors were busying developing, I was in San
Atonio to pariticpate in the ACM/ICPC world final 2006. Finally, we
got the 13th place, which is the best place ever got by my University.
:) Ah, Wish you get less appeals today. :D
Appeal Response:
Congratulations about the 13th place!! I didn't know there was such a
place, I'd have thought that after the 12th there would be the 14th.
|
