IOI Contest Day 1

Today was the start of the real competition at the IOI with the first of two five hour rounds with three problems each. From the GA meeting last night us team leaders saw the problems and recognised the last two as challenging, however, we still underestimated them as the scores have turned out lower than usual.

While the contestants wrote the competition the team leaders attended the 2nd IOI Conference, where some interesting papers were presented. Unfortunately I've been so zonked out that I had to skip the second half in favour of sleep.

As usual it was very stressful waiting for the contestants not knowing how they were coping. At least with the ACM ICPC we get to view the scores during the contest, whereas here we get absolutely no feedback. This was made worse when the spew of contestants came back from the contest arena into lunch and they kept on coming, and coming and we never saw our team...Peter eventually decided to look around for them and found them. The first signs weren't very good, but after hearing what the other teams had to say we felt a bit better.

One thing in past IOI's that has varied drastically is the time for results to be released. This year it was looking reasonable as they quickly got the results printed, but then they took their sweet time to release them and put them in the pigeon holes. They insisted that we could not collect them until they had placed everyone's results in the pigeon holes, which was a painful process to watch. Just before they eventually finished I was told they had enabled analysis mode so many teams saw their results on the contest computers. We stuck it out a couple more minutes and I was one of the first to collect the paper results.


The results were almost as expected from what the guys had thought of their submissions, minus a couple silly bugs. Our top score was 100, which given the partial unofficial results doesn't look so bad compared to what it might have in a previous IOI. A good score seems to be 140 -- 100 for printing and 40 for islands, although 100 is good enough to make a return in the next round of competition.

After the results were released we went back to the contest arena for analysis mode. We tried saving our solutions and the official test data to flash disk, but the judges for some reason didn't like that idea and were insistant that we stop immediately. I hope we weren't the only ones stopped as we weren't even allowed to print solutions. Anyway, we looked through the results and discovered a couple silly bugs such as the use of ints instead of long long in a single place for islands.


Once we were satisfied with things we went to catch the bus for the Nile cruise, which was supposed to leave it 17:00. Want to guess what happened? Why yes of course, we sat in the bus for about 2 hours before leaving with absolutely no reason given at all! When we finally did get going we were stuck in pretty heavy traffic, which is unusual for us at this time of day. It was about a 90 minute bus ride to the ship and we caught our first glimpse of pyramids along the way. I was so surprised to see them as I wasn't expecting them that when I did see them I shouted "Woah, pyramids!!" and my entire team abrubtly woke up!


And I just have to include this wonderful photo:


After waiting so patiently in the bus and taking the long bus ride, our patience was to be tested even further as we waited about another 45 minutes for the ship to arrive and be prepared for us. When we eventually got on we were offered a buffet dinner and we ran to be one of the first in the queue. It was pretty good food and a nice filler given the previously smallish meals we've been served. The dessert was especially delicious!


The dinner was accompanied by some humorous dancing:





The bus trip back included lots of sleeping:


Tomorrow we have a trip to the pyramids of Giza in the morning followed by activities and the GA meeting for the second day of tasks.

IOI Day 2: Opening Ceremony, Practice Session and Task Selection

Dear oh dear, the Egyptians really take African time way too seriously. The organisation to get into the venue for the opening ceremony this morning wasn't the greatest, but they definitely made up for it with the performances. The speeches were cut to a minimum, although we've been told they're holding them back for the closing ceremony where we'll have the president's son and expected future president give a speech amongst other top people in Egypt.




The speeches were followed by some Egyptian dancing, which was quite fascinating to watch especially with the dresses. We then had the world's most famous marimba player give a solo, which was pretty damn awesome! The crowd really cheered her on very nicely and it made the opening ceremony a memorable one.




Due to the delays they had to shift the practice session after lunch, which was made an hour early. Lunch was once again the big meal of the day (something I have yet to get used to). After lunch we hung around a bit more waiting to head over to the contest room for the practice contest.

The practice contest is always an interesting time as its the first time we get to see the contest environment. The handin and evaluation system has gone back to being a new system specifically developed for this event, something which I had hoped had died given the recent years. As expected with any new system there were problems and one major one was the networking issues they had. I'm not sure if related to the networking problems, but our team would get different feedback for different submissions of the same code! This was highly frustrating given that you couldn't trust the errors it spewed at you. There were a few other problems which I hope they'll sort out before the competition tomorrow.


I'm not sure where they bought them from, but look at what they're still using these:


The contest arena was separated into 13 rooms with about 8 teams per room, which is different to what I'm used to seeing in the past. We had the Spanish and Swedish team next to us and I had some interesting discussions with their leaders (two of whom I had previously met). The Swedish guys are hosting the ACM ICPC World Finals next year and I've been talking to them a bit about that. They claim that they're going to be putting on a real whopper of a World Finals, already have booked out three of their grande hotels! They're preparing to host the first World Finals with live commentary for spectators, streaming it over the Internet. Because of this they are also pushing to get automated judging, something which the judges are constantly resisting. Lets hope they can successfully outdo Shanghai, which would be really impressive.

After the practice session we had to rush over to our first GA meeting, leaving the contestants behind. The GA meetings can be rather amusing as they're highly democratic with voting on every minor point. They go as far as voting on the agenda of the meetings. This year started off no different to last year, with a couple issues raised. One was that the version of FPC used was unstable and the request was that they change the compiler flags from -O2 to -O1. However, as you can imagine this would cause countless problems in the uncertainty as they had tested the solutions using the exisitng -O2 flag.


After seeing the contestants for one last time we went to the second GA meeting where we were given the tasks for the first day of competition. Our job was to scrutinise the tasks and then translate them into our national languages. This is always a fun time as we get to discuss in detail with other team leaders the solutions to the tasks.



[This post was only published well into the start of the competition. The timestamp is the time of writing, not publishing.]

The first problem is really easy with only a small trick you have to notice, so I hope most of our team doesn't overlook this small trick and solves it successfully. The second problem is a tricky graph problem with lots of smaller parts that need to all work in linear time, which I haven't quite worked at a full solution to as I'm way too sleepy right now. The third problem is a really interesting task that stumped even the original author who submitted the problem with a quadratic solution in mind, but the scientific committee discovered an NlogN solution of which all I've heard is really icky. I'm sure I'll hear the solution soonish though. All-in-all a very nice problem set in my opinion.

So back to the task scrutinising. We had a couple minor issues with the task descriptions, but as we never considered them problematic enough for our team we never reported them. There was a total of ten minor and one major objection. The major objection was from the Croatian team that they had a similar problem to Islands in their national contest last year. The scientific committee said they were aware of this task and considered it significantly different enough to keep the task and everyone took their word for it without further question. Besides one minor objection, which I cannot remember, the others were all accepted. Then their was one guy who blurted out something minor about the Islands task that he wanted to be considered a major issue. Fortunately the voting process stepped in quickly which democratically calmed him down.

Once the tasks were accepted we moved onto translation. One advantage of coming from an English speaking country, however, is that we get to mingle with the other English speaking countries instead! It took a good while for them to bring in dinner and it arrived cold, as we have become so used to. I could have easily downed two of those meals, but many people had to wait for the second and even third batch to arrive so I never did.

Peter and I ended up spending most of the time, and yes it was a long time, chatting with the Brits and Troy. We kept wandering just when they'd let us leave, a major point of discussion in all years as we have to wait for the contestants to return from their activities and get back to their rooms so as not to meet with the leaders who have seen the tasks. We also have to wait for the tasks to be finalised, which thankfully has hasn't been a bottleneck so far this year.

After we got sick of debating on the Egyptian definition of "soon" I went to go have a talk with the Ghana and Nigerian leaders. We discussed some of the issues that they've been having in running their national contest to select their teams for the IOI. They sound very enthusiastic about getting involved and are even set on entering some teams into our ACM ICPC regional contest for Southern Africa with the eventual aim to break away and form their own West African regional. They're very eager and think they can get at least 6 countries from West Africa to participate and they definitely have a strong backing from the government. I talked to the about the African contest for university students I've been running and they were very excited to hear more about it.

Eventually at about 23:30 we were told we could leave, which was half an hour sooner than our original estimates so one point to the organisers! We all ran straight for our rooms after getting the word so the day was over.

Tomorrow is the first day of competition from 09:00-14:00 with an online contest scheduled to start at 17:00 (14:00 GMT) while us leaders attend day one of the IOI Conference. This is followed by a trip down the Nile river.

Waiting...

So the Egyptians like to keep us waiting by changing schedule several times. First they switched the opening ceremony and practice session around, then switched them back to their original slots and now they've delayed the opening ceremony by an hour. I suppose they have some important speaker that's got a very busy schedule or something. Oh, and the buses may leave after the ceremony starts, woopie!

More time to meet people though! Got to meet Ilham from Indonesia, a frequent reader and commenter on this blog, and someone who I cannot remember his name from Emirates.

Arrival in Cairo, Egypt

After about 10 hours of flying we arrived at Cairo airport at about 06:30 this morning. We were greeted by a couple of very helpful IOI organisers who helped us through customs and such and took us all the way to our bus.

We then had about an hour bus ride to the Mubarak City of Education where we'll be staying for the next week. The ride was slightly crazy, with drivers straddling often lanes between two other cars in the actual lanes. Not nearly as bad as my recent trip to China though. We passed north of Cairo (the place we're staying is about 20km outside of Cairo) and got to see some of the less upmarket apartment blocks. What was very strange was the number of satellite dishes they had, all of them looking as though they haven't worked for the past 20 years. Click the images below to see what I mean.





When we arrived at Mubarak City we were welcomed by our guide Mustafa, who has been very friendly and helpful as have many of the other guides. He quickly took us to register which was so fast I was dead sure we had to return later, but that was not to be. He then took us to our rooms. The one thing with the IOI is that you almost always stay in student dorms and this year's no different. The four contestants are staying together in a single room while I share a room with four beds with our team leader Peter Waker. There is a single bathroom shared between two rooms and we share ours with the Spanish team leaders.

After sleeping for a few hours we went down for lunch at 14:00. Here we met up with the Fieke Dekker, the British deputy leader, and Don Piele, the US team leader. The locals consider lunch the big meal of the day so we had some nice sausage, chicken and rice with soup as a starter and some really nice apple pie for dessert.

After lunch we went to pay the registration fees, which once again was painless and quick. We then wandered around and met some more people. I bumped into Mohamed Taha, member of the host scientific committee which I'm friends with from the ACM ICPC. I had a lengthy chat with Troy from Canada, Fieke and her husband Richard with Kim from the Netherlands and Rob Kolstad from the US bumping in a bit.

We then went to the IMAX to watch some Dinosaurs: Back to the Crutaceous or some other documentary in 3D. The 3D part was awesome as always, but the story line and acting was shocking. We also met up with the contestants at the IMAX (we split up after arriving) and went to the Internet lounge together where we met yet more people. There was also this random-looking Arab guy with this huge pot around his shoulders handing out this very strong berry-tasting juice.

After hanging around we went for supper at 20:00, where we met up with the Dutch delegation which we are very friendly with. Supper was, as we were warned, basically bread with a couple small pieces of chicken.

After dinner they had a party with some loud music. Not usually my scene, but the chance to meet new people is always well worth it. I met the Swedish delegation, who are hosting next years ACM ICPC World Finals. Along with other very interesting things I heard, they told me they've booked out the top three hotels in Sweeden for the event! We also met the Serbian delegation, one of the few with a female contestant and the team from Ghana, participating for their first time.


New Zealanders and Aussies, interesting bunch that lot. Well we've met some of the contestants from both sides earlier on and now again some more. Well lets just say they were the ones who got the whole party really booming. Here's the main culprit (blonde dude):


And this is how they partied:


It's an early morning tomorrow with the opening ceremony starting at 09:00 and the practice contest at 11:00. This is then followed by the first lot of GA meetings which are always most entertaining with their extreme level of democracy, while the contestants enjoy themselves at some Golf park.

Final IOI Preparations

The International Olympiad in Informatics (IOI) is the pinnacle of programming competitions for high school students. Getting the opportunity to lead your countries team is a great honour, one which I get for a second time at IOI 2008 starting in a weeks time. This year it's in Egypt, making it the second IOI on the African continent. Need I even say how excited I am to finally get the opportunity to visit this amazing country holding some of the world's most famous monuments, including the magnificent pyramids.


The South African team for this year was announced in May this year, but one of the members withdrew due to personal issues, otherwise we would have two brothers (a first for IOI?). The team I will be taking to Egypt, along with Peter Waker:

1. Mark Danoher: 1st in SACO 2007 and participated in IOI 2007
2. Schalk-Willem Krüger: Shown great success and is young enough to compete until IOI 2010
3. Saadiq Moolla: Narrowly missed the team in 2007 and participated in IMO 2006/7
4. Robert Ketteringham: Reserve called up last minute and is eligible to compete in IOI 2009

I will be presenting a paper at the second IOI Conference held in conjunction with the IOI. It is titled "Challenges Running a Computer Olympiad in South Africa", which outline the format of the various rounds of the SACO and identifies the challenges we face and how we have tackled them.

The team gathers together for one final training camp starting Tuesday. We depart on Egypt Air on Friday, arriving early Saturday morning. We hope to have a prosperous IOI, and I at least really look forward to enjoying what our fellow Africans have prepared for us (schedule here). I also look forward to once again meeting some of the top international talent, which is always worth the trip on its own. Keep your eyes on this blog for our experience in the ancient land.

Chinglish

Funny Chinese signs with broken English are all over the Internet. It's one thing seeing them on the Internet, but seeing them for yourself you realise how truly bad the situation really is. It's so difficult to believe that people can screw up so royally this often, but it couldn't be more true. I've put together a collection of some of the wacky signs and such that we saw in 9 days in China, without us actively searching for any! Hopefully this put things into perspective for some people.

Anyone want to try travel the hurl?

This was on the door of public toilets

Um...right!

Chinese Red Bull (check the bottom!)


Conjuring with grapes...um yes

What makes chips Natural & Warm?

A simple misspelling gone wrong! There's food behind there. ;)

Found outside the Oriental Pearl TV Tower; some of these are hilarious (click image to read)


What can I say...

Would love to know what they meant to say!

Not so much poor English, but funny nonetheless


Thank-you sooo much :)

And then there were some bad ones even at the RoboCup arena. The fact that this is an international event amazes me that these were there.

Looks innocent enough, right?

Look closer at the spelling of RoboC*u*p. That's the spelling used on the official t-shirts, but the certificates were worse -- Robocuop!

The Chinese seem to love their Happy Party


Um, yes...this is referring to RoboCup as was the one above.

Since when is RoboCup a pageant?!


Besides the funny description of RoboCup, notice the spelling :)

Cuil Fails to Impress

There's been a lot of buzz around Cuil, a new search engine produced by ex-Googlers. In my opinion it is mostly due to this very fact that they are ex-Googlers that there is such a buzz around it and if this is the case it is sure to be short-lived if they don't up their game fast.

I just recently got myself a new phone and I've been searching for what cool things you can do with Bluetooth on Ubuntu. I had been previously searching for this on Google and what would you know, the first two hits are very useful. Lets try the same search on Cuil. What are the results like? Well it seemed to only care about the "ubuntu" keyword and almost completely disregard the "bluetooth" keyword. I couldn't care less about Ubuntu 5.10 release notes!


Not a good first start. I'm sure many people have tried searching for themselves on Google just out of curiosity. I do so fairly regularly as it's one way I can track how Google changes results over time. Do the search on Google and you get my blog and website right on top, spot on! Do the same and Cuil and...oh no! Tiramisu recipe?! Ok so there's an Italian cook with that shares my name. What's the flight simulator screenshot doing next to a result on recipes?! Then there's the five results that look identical from looking at the results page, their snippets need some urgent attention. It seems they just increased the snippet length to get around that downfall of theirs, but nope it won't work for long. Two of these reults are in fact exact duplicates, so their duplicate detection needs working on if they have any at all. There's no point boasting about having the largest index without it being a useful one. I'd rather have 10 solid results than a million shitty ones with those 10 hidden amongst the mess.


Lets dig through a few more pages of results to dig deeper. Nothing interesting on the 2nd page but my Digg profile, how exciting. 3rd page, nothing. Dear these thumbnails next to the results are random, what's up with the guy weighing himself? 4th page, my Digg profile again (just with the www. in front this time!). And oh wow, finally my website...on the 4th page of results!! I tried looking for my blog but I gave after page 10.

In response to Tim Keller's more positive comments:

• Black background on the homepage: Might be nice, but not going to convert me
• Auto suggest: You get the same thing with the Firefox search bar or Google Suggest, which return ranked suggestions.
• Magazine style presentation: Nice at first glance, but is it productive?
• Simplicity:Personal taste, but the value far outweighs the clutter of the "Sign in" options and such. Google run exhaustive tests before changing their interface.
• Explore by category: Nice idea, already tested by Google while I was there, but from my experience the categories weren't often relevent.

In short, a few nice new touches to the interface isn't going to outwiegh quality results. While it may be promising, Cuil really needs lots of work before it gets anywhere in the highly competitive search engine market. And please stop the boasting in the about pages, not even when you do make it up there!