Discussion about the upcoming new version of KakaoStats

[Kakao]

I'm starting a major rewrite of KakaoStas with the main goal of making it update every hour. This is the best time for discussions about it. Although I will not include new features in the first versions, feel free to express you wildest dreams as I can consider adapting the (of utmost importance) database design to accommodate them.

[kajaco2]

One thing I'd like to see is a notice at the top of every page if the updates are not working or you've got the database down for maintenance or something. This should include an estimate of when things should return to normal (which you would keep updated as necessary). Right now the information is about a half day behind and there is no explanation or indication of when this will be corrected.

[Kakao]

I didn't know it was not working. Something crashed and I can only fix it in about 8 or so hours. I'm opening a bug topic.

[vtq]

The stability of Kakao Stats has to be improved. The stats are presented in a very unique and great way, but the update feature is halting to often. Some days you can see big holes in the point system and then all points come all together at a later update.

[Kakao]

That is sure very important. The holes are caused mainly by connection problems when the summary files can't be downloaded from Stanford.

I made a small improvement in this last crisis with a hardened download script. But I'm still stuck behind an unreliable ADSL connection and that won't change.

[masbuskado]

Well I really like the way kakaostats is at the moment, also I start a new topic with a few ideas and man if you need any additional help you can count on me... I am not here to tell you what to do but to help you in your quest... kakaostats is the best Folding Stats website out there but sometimes people overlooked the hard work you do!!!

If I can help you in any way just let me know! 

[AJMSmith]

One thought I'd like to throw into the pot is that when (rather than if) the stats fail to download for any reason when you get a new set average the numbers over the missing sets ... not absolute in it's accuracy ... but better than the present as will remove the large spikes (both up and down) that occur and produce misleading stats.