The Countdown Database - round details?

[Graeme Cole]

Looking at that product of many hundreds of man-hours of work by people on this forum, the Countdown Database, I'm wondering what goes on behind it. I presume it's some sort of relational database like MySQL or similar, but how detailed is it? I'm guessing it has the potential to store everything, right down to the individual selections and offerings for each round, because the recap writer provides this information. But the pages that are there suggest it's only the contestant names, final score and maximum score of each game that get stored in the database, with the round-by-round details only going in the recap and on the wiki. Is this right or have I assumed wrongly?

The reason I ask is that it looks like there are more pages that could go on there if the round details were used as well, very much like what Statland does for apterous, like a page of the biggest comebacks, or most successful letterists/numberists/conundrumists, or most missed/offered/disallowed words, or, as Mike Brown mentioned on his site some time ago, things like how many times LEOTARD and TANGELO have been offered.

Are there plans to add further pages to cdb with this kind of analysis, or does the database not contain the raw data to that level of detail, or is it not practical for some other reason?

[Mike Brown]

I've often wondered the same thing myself (and came to much the same conclusions), but never got around to asking Charlie. Care to comment, Mr Reams?

[Charlie Reams]

My plan for the last few years, which it occurs to me I have never really articulated publicly before, has been:-

1) Get all the games recapped. This got completed way before I thought it would, which is why the later stages aren't quite ready yet.
2) Get all the details onto the wiki. This is pretty easy because of the magic the Recap Writer does, and is indeed mostly complete. The wiki is a useful intermediate step because on the one hand it's human-readable and human-editable, which lets us find and fix mistakes; on the other hand it's machine-readable because of the template system, so I can proceed to...
3) Write a script which reads and "understands" the wiki pages. This is mostly done already in order to provide the Classic Games for apterous, but there are some things that apterous doesn't care about (like original DC words/beaters, numbers methods) and these currently aren't captured by the script. Also there are some rare elements of the wiki markup that it can't handle, like words which were valid but scored nothing because of a length misdeclaration. All that would have to be done correctly in order to...
4) Spit all that data into the CDB database, which currently holds only summary-level details of the games. Then...
5) Write some sweet stats pages a la apterous (although this plan pre-dates apterous). And finally...
6) Have a lie down.

So we're currently somewhere in Phase 2, with a lot of infrastructure for later phases already done (for example, I own a bed). The remaining steps don't lend themselves to crowd-sourcing so easily, so it'll probably happen incrementally whenever I have time. But it will happen.

Hope that helps!

[Gavin Chipper]

Hope that helps!

HTH surely!

[Graeme Cole]

Hope that helps!

Very informative, thanks. Looking forward to when it's all shiny and complete.

[Mike Brown]

Hope that helps!

Very informative, thanks. Looking forward to when it's all shiny and complete.

Agreed. When it's all done, I will have to buy you a very large drink.