Forgot your password?
Live Contest Blog
Hello everybody,
the results of the evaluation of the Plat_Forms 2007 contest have now been published.
Find a summary (recommended!), a short presentation, and the lengthy full report (recommended!) on page Results.
Feel free to discuss contest, report and results in the comments part of this blog entry. No registration is required.
 Lutz Prechelt
Alvar Freude, the Perl platform representative, who was present throughout the contest, found that most teams looked "typical" for what he would expect a team using that platform to look like.
To see whether this is really true (and to pass his idle time), he prepared a set of photos of each team (and even their flipchart) in the following way:
- Alvar has removed telltale platform signs such as logos and slogans (even the team numbers) from the photographs
- For each set of photos, you can choose via a checkbox what you think is this team's platform
- When submitting such a choice you will be shown what the other visitors believed
- When finished with all teams you can visit a page containing the true answers
A very nice game for all those of you who have not yet studied the team photos further down in the blog too closely (or have already forgotten what they saw).
Have a look at www.a-blast.org/plat_forms.
The live phase of the contest is now over.
All nine teams have delivered a solution in time.
Their servers are now down (because they were the teams' own hardware) and so the solutions can currently not be tried out any more. At some point in the future, we may make them available again publicly on an FU Berlin server, but when that will happen depends on how the evaluation progresses (which we do not know yet). In the not-too-distant future we will also put up for public perusal the source code distributions that the teams have put together.
In the meantime, the organizers thank all who have helped the teams by commenting on their public prototypes!
We expect to present the results of the evaluation of the solutions in May (and possibly a few pieces before that time).
[dami from team1]
Nice to be home, after a good night of sleep ... but my head is still partly in Nuremberg, thinking of what went well (quite a lot of things), and what could have been better (a couple of lessons learned). I'll post about those thoughts a bit later ... for the moment, I am trying to focus on the next thing to do, namely refreshing the voice and remembering all my texts for singing Rossini, Schubert and Spohr tonight. So I'm not yet over with stress !
Platforms was a very nice experience, and the given task was quite interesting. For a while we thought we could do it all (over-ambitious as usual!) ... but then a couple of unexpected problems rose and we lost too much time to finish properly : how frustrating !
Thanks again to the organisers for the excellent preparation.
So, we are now burning the CD...finished everything.
Went quite well, unfortunatelly we found some bugs at the end we could not fix anymore....but it is not so easy to concentrate after 30 hours.
Our statistics:
- Drunk Red Bull: 6 cans
- Slept (totally): 11 hours
- Went to toilette (totally): 22 times
- Lines of Code: No idea!
- Got "D"s (which means we where developing when the supervisers come): "A lot"
Cheers out there!
For all those of you who want to take a real close look at the precise task the participants are trying to solve, here are the three files we gave the participants:
- the requirements document platforms-task.pdf they received both on paper and on CD. It also explains the role of the other two files on the CD, namely
- ttt-questions.txt and
- PbT.wsdl
The requirements describe a task called PbT (People by Temperament), a simple community portal for contacting others based on personality type and other personality-related profile information.
we updated our preview server continuously the past hours. Now we announce an official public version.
- Status Page
- Displaying prviate information if I am in contact with the person or viewing my own status page
- Viewing TTT results and timestamp of public TTT
- With whom I am in contact with, who I sent RCDs and who sent RCDs to me and view their status pages
- Accepting or denying received RCDs (you will not be able to test this since there is no RCD_send method yet)
- Running the "Trivial Temperament Test"
- Results are displayed very minimalistic :)
- Viewing member list with 2-dimensional overview plot
- Searching contacts is NOT working!
Thank you all for your feedback ;)
Here is some information about when the teams as a whole decided to sleep, that is, when none of the team members were in the contest room. The data is a little misleading because some teams left and arrived all at once, while other teams send some members to go sleeping while others continued working. But even the team6 members all had some sleep.
Team absence during the night
| team | begin | end |
|---|---|---|
| 1 (EdG/Optaros, Perl) | 5:00 | 8:00 |
| 2 (plusW, Perl) | 1:00 | 7:00 |
| 3 (abaXX, Java) | 3:30 | 7:00 |
| 4 (Accenture, Java) | 1:00 | 8:30 |
| 5 (RevSys, Perl) | 1:00 | 6:00 |
| 6 (Oxid, PHP) | * | * |
| 7 (Globalpark, PHP) | 1:00 | 7:00 |
| 8 (Zend, PHP) | 2:30 | 8:30 |
| 9 (Innoopract, Java) | 1:00 | 9:30 |
By and large, everybody looked quite acceptable this morning and hardly any loud groans could be heard.
Two teams even report they had several beers together at a bar before they went to bed.
At about 10:00 I went around and asked each team to prepare a rough estimate of what fraction (in percent) of the functional requirements they expect to implement until the end of the contest (15:00), separately by requirements category.
| team | MUST | SHOULD | MAY |
|---|---|---|---|
| 1 (EdG/Optaros, Perl) | 70 | 80 | 65 |
| 2 (plusW, Perl) | 80 | 50 | 33 |
| 3 (abaXX, Java) | 100 | 70 | 50 |
| 4 (Accenture, Java) | 50 | 10 | 10 |
| 5 (RevSys, Perl) | 100 | 100 | 75 |
| 6 (Oxid, PHP) | 80 | 80 | 50 |
| 7 (Globalpark, PHP) | 100 | 20 | 5 |
| 8 (Zend, PHP) | 70 | 20 | 10 |
| 9 (Innoopract, Java) | 20 | 10 | 10 |
Fairly different amounts of confidence...
(Yes, the requirements specification document will be put online some time soon)
So we are back to work after about 2 hours sleep ... even such a short break was good to take.
We now will first work on a bugfix release, to be delivered early in the morning. Then we will spend the rest of the day adding as much functionality as we can ... knowing that the 100% target is unreachable until 15:00.