ok..

let me get this straight. instead of downloading a max of a 15k xml file, I instead have to download a team.xml.gz file, and the user.xml.gz file, the sum of which was over 160 MEGS?!?!

You have GOT to be kidding me?!?! how is this going to save bandwidth?!?! I update once, once a day, and it was being run from a web-based PHP script that downloaded, parsed, and processed everything in under a minute. NOW I'm going to have to spend over an hour a shot, and download NOT just the data I need, but the data for EVERY other team and EVERY other user in the system?!?!? Whats wrong with this picture? Wanna cut back on bandwidth?

I say emphatically that bandwidth doesn't grow on trees, and this is going to KILL the setinews I write every night. Shoot, that news is the reason HALF my team crunches. Don't shoot us (or yourselves) in the foot. Come up with a better solution!

> let me get this straight. instead of downloading a max of a 15k xml file, I
> instead have to download a team.xml.gz file, and the user.xml.gz file, the sum
> of which was over 160 MEGS?!?!

The XML stats are really in a state of flux at the moment... their primary target is big sites like boinc.dk and the french site that provide databases about all users. I would suggest posting to the boinc-projects mailing list with a better idea of how to handle the XML stats for smaller sites that just need a few users' or teams' stats instead of ALL of them. They may have just forgotten you're out there... you need to remind them, and probably politely.

FWIW, I don't think many of the boinc devs are reading these boards right now. They're pretty darned busy with other stuff at the moment. :-)

Hope this helps....

Rob

and whats that boinc-projects mailing list addy?

I'll post anywhere, and I'm usually respectful, I just personally think this could be a project killer. ALLOT of people are in it for the competition, or to see their name every now and then. and those that aren't, are still encouraged by seeing those same results.

I've written the LittleWhiteDog SetiTeam News pretty much nightly for the past year at least (2 maybe?) and I know my php scripts would just laugh at me if I tried to download 160 megs.

So if you could pass on that mailing list addy, I'll add my 2 cents, hopefully bringing this issue to someones attention

The problem with this XML stats is, that the database is so large, that the files get this size or create timing problems. When this side come up there where files splitted in smaller pieces awailable, but this system crashes because it needs to much time and performance.

So the developers had to find a quick solution, which seems to be at first to support the large stats sites for having one working solution, better then having nothing. I didn't see a official statement, but I know this was in discussion and now it looks like it is done in this way.

I didn't believe the developers forget the smaller teams. A solution for this problem will be found. One way could be that the large stat sites distribute the information for the team stats scripts or a separate stats server is created in Berkeley and splitt the large lists in usable pieces.

And you can be sure that the developers didn't find much time at the moment for reading the boards, but there are still a lot of alpha and beta testers around, supporting the devs with informations about the problems and trends shown here in the boards.

If you want to tell Berkeley that there is a need for more supporting a special issue, this boards are the right place. This will force other users to talk about it and to support or deny and create a better picture as using the mailing list.

If you are developing software addons or i.e. stats sites and have a idea for a way to solve the performance problems when creating the XML files, or a other idea how to better support your creation of team stats, then the mailing list may be a good way. You find the list on the BOINC Community and resources site.

Greetings from Bremen/Germany

Jens Seidler

It Juste what I said on this thread :
http://setiboinc.ssl.berkeley.edu/sah/forum_thread.php?id=259

I try to do some Seti Stats using the Berkeley XML files.

The result is here, before da mighty reset ;)
http://sahstats.boinc.fr/


But, in wich format will the XML file be. It appears that the 'team_id' recently change : we don't have the url, country ... any more. Will they be reintegrate in this file ?

When will those file updated, and which format will the official seti crew choose ??

Thanks a lot.
http://sahstats.boinc.Fr