Results 1 to 10 of 10

Thread: Minor fault - Probable rogue data - Simple enhancement suggestion

  1. #1
    Senior Member MarkRBright's Avatar
    Join Date
    Oct 2011
    Location
    West Yorkshire, UK
    Posts
    238

    Minor fault - Probable rogue data - Simple enhancement suggestion

    So yesterday I lost a Maxwell. No big deal really, it happens. But I had a bit of spare time so as it wasn't obvious where it came from I thought it was time to update my own stats. I maintain a spreadsheet which has all my own Boinc related data in plus some bits that aren't displayed in Free-DC. So I painstakingly went down each item in the Users Milestones page, clicking to find out what the number required for a Maxwell was, the number required for a TeraMiRB and I would have checked the number required for a MiRB, but as there is no way that I have found to get beyond the first 500 records so I couldn't. Luckily I have MiRB's for every MegaMilestone that I have so I am not really bothered by this, though others might be. So I update my table and find there is what I suspect a rogue entry in the SubClanks category (https://stats.free-dc.org/userbycpid...ubproj=sclanks) where the best entry before was around the 40 mark, there is a new entry of 76. Pretty excellent to maintain that level of SubClanking! But slightly suspicious as whoever is responsible has no name and no team, is literally just a score.
    So I get to thinking - not for the first time - how nice it would be to enhance the Users Milestones page (which you have said before needs some rework anyway) to include 3 new columns, showing the number of each MM Type needed to obtain a Maxwell, a TeraMiRB and a MiRB so that it looks like the attached picture.

    It is worth stressing that the data shown in the additional 3 columns would be the same for every person looking, and would only need to be collated once per day as you go along when doing daily roll-over. Presumably would want to continue to not show details for Mega Milestone categories where weren't being shown anyway. i.e. I have no 1B's, so I wouldn't get to see any details (new or current) for that line.
    It is also worth stressing that some people like badges, and some like MM's, Some like both but obviously everyone just likes crunching. Anything that both adds to the pleasure of crunching whilst making people more interested in your web site and what it offers is a double whammy - good for Free-DC and good for Boinc.

    So to summarise
    1) Minor fault - can't get past top 500 ranks when displaying on any entry in the Type column.
    2) Probable Rogue Data - first entry in the subClanks rank table looks very suspect.
    3) Simple enhancement suggestion - three new columns on the MM's table - calculate once per day - same for all users.
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	Boinc Suggestion re MMs Thumb.jpg 
Views:	160 
Size:	53.6 KB 
ID:	1392  

  2. #2
    Administrator Bok's Avatar
    Join Date
    Oct 2003
    Location
    Wake Forest, North Carolina, United States
    Posts
    24,450
    Blog Entries
    13
    Fixed the rogue entries by excluding data which has a null cpid (I may have to relook to see whether that's actually valid data in the first place).

    as for the 3 new columns that's not a simple task at all, I haven't looked at those numbers in some time though, so need to think about it.

  3. #3
    Senior Member MarkRBright's Avatar
    Join Date
    Oct 2011
    Location
    West Yorkshire, UK
    Posts
    238
    Hi Bok
    Thanks for sorting the rogue out - I got my Maxwell back
    It surprises me that the 3 new column suggestion is that tricky given that the numbers are currently readily accessible to anyone from the existing data with 10 minutes to spare. I simply look at the number held by whoever is ranked in 1st (and half it - rounding up), 100th and 1000th place for each Mega Milestone. Apart from the current issue of not being able to look at the 1000th at the moment, it is straight forward if a bit of a pain in the back side, and the sort of thing that computers do wonderfully well. I confess to being out of date with programming, but like I say, it surprises me that it is difficult. Perhaps I didn't explain myself very well ... feel free to let me know if that appears to be the case.
    Thanks again, you do a brilliant job with this site.

  4. #4
    Senior Member MarkRBright's Avatar
    Join Date
    Oct 2011
    Location
    West Yorkshire, UK
    Posts
    238
    Oh, I just spotted what is probably another rogue entry, this time in the subbok ranks (https://stats6.free-dc.org/mmakers/sboks) - albeit one with little impact other than it pushes everyone down one spot in rank - the first entry for 11 appears to have no name or team there either. I expect kashi deserves the number one spot more. :-)

  5. #5
    Senior Member MarkRBright's Avatar
    Join Date
    Oct 2011
    Location
    West Yorkshire, UK
    Posts
    238
    Quote Originally Posted by MarkRBright View Post
    So to summarise
    1) Minor fault - can't get past top 500 ranks when displaying on any entry in the Type column.
    2) Probable Rogue Data - first entry in the subClanks rank table looks very suspect.
    3) Simple enhancement suggestion - three new columns on the MM's table - calculate once per day - same for all users.
    Just in case you had forgotten about these, all three are still there albeit the Rogue Data one is now in subBoks. There may be more - I'll point them out if I see any next time I try to trawl through making up my manual set of item 3) data.

    To be honest, the first one becomes irrelevant (to me at least) because I probably wouldn't ever feel the need to go looking for position 1000 (and beyond) if item 3) was done. ;-) Though as you are offering to show all (or first 1000) I am sure there will be people out there who want to look. I suspect this is a rather trivial fix too as it works just fine for 25, 50, 100 and 500. i.e. it shows the first 25 of 500, the first 50 of 500 etc. It just looks a bit silly when it effectively ignores you when you choose 1000 or all.

  6. #6
    Senior Member Dirk Broer's Avatar
    Join Date
    Jun 2013
    Location
    Leiden, South Holland, Netherlands
    Posts
    368
    Citizen Science Grid now has at least the user_work.gz file back in their stats export, badge_user.gz is still missing:
    [ICO] Name Last modified Size Description
    [PARENTDIR] Parent Directory -
    [ ] db_dump.xml 2019-03-11 12:55 701
    [ ] host.gz 2019-03-11 12:55 9.2M
    [ ] tables.xml 2019-03-11 12:55 2.1K
    [ ] team.gz 2019-03-11 12:55 146K
    [ ] team_work.gz 2019-03-11 12:55 28K
    [ ] user.gz 2019-03-11 12:55 947K
    [ ] user_work.gz 2019-03-11 12:55 464K


  7. #7
    Senior Member MarkRBright's Avatar
    Join Date
    Oct 2011
    Location
    West Yorkshire, UK
    Posts
    238
    Quote Originally Posted by MarkRBright View Post

    So to summarise
    1) Minor fault - can't get past top 500 ranks when displaying on any entry in the Type column.
    2) Probable Rogue Data - first entry in the subClanks rank table looks very suspect.
    3) Simple enhancement suggestion - three new columns on the MM's table. The numbers of each MM required to get a Maxwell, a TeraMiRB and a MiRB

    To be honest, the first one becomes irrelevant (to me at least) because I probably wouldn't ever feel the need to go looking for position 1000 (and beyond) if item 3) was done. ;-) Though as you are offering to show all (or first 1000) I am sure there will be people out there who want to look. I suspect this is a rather trivial fix too as it works just fine for 25, 50, 100 and 500. i.e. it shows the first 25 of 500, the first 50 of 500 etc. It just looks a bit silly when it effectively ignores you when you choose 1000 or all.
    re 1) No change here. Still can't get past the first 500 in any MM, whichever way I have tried it. It's fine up to 500, whichever number of entries you choose. Even if you select 1000 it stops at 500.
    re 2) subClanks is fine thanks, but just spotted that there appears to be another rogue entry in strigggls. It apparently has 9, and the real second place person has 4, it means that no-one has a maxwell for strigggls
    re 3) Just to reiterate, the values in these 3 new columns could be calculated once per day - and the numbers would be exactly the same for all Boincers. Also by doing this I would quit harping on about number 1

  8. #8
    Administrator Bok's Avatar
    Join Date
    Oct 2003
    Location
    Wake Forest, North Carolina, United States
    Posts
    24,450
    Blog Entries
    13
    1. the sql was limited to 500 rows. I've upped that to 2000 right now, anything more I need to write a 'serverside' version of the code for that page which only gets the rows requested.

    2. Looks like this is due to old subproject data for accounts that have changed cpid. Deleted them out and re-ran the strigggls. I've added an 'IS NOT NULL' to the sql

    3. I still can't wrap my head around how I would dynamically or otherwise calculate these numbers the way it's written right now.
    Last edited by Bok; 05-23-2019 at 11:10 AM.

  9. #9
    Administrator Bok's Avatar
    Join Date
    Oct 2003
    Location
    Wake Forest, North Carolina, United States
    Posts
    24,450
    Blog Entries
    13
    I did some more thinking on the #needed for maxwell. I do have some intermediate tables I use in the calculation. Please take a look at your userbycpid page on stats6 and see if I'm on the right track ??

  10. #10
    Senior Member MarkRBright's Avatar
    Join Date
    Oct 2011
    Location
    West Yorkshire, UK
    Posts
    238
    I only needed to be able to get to 1000 really (and I suspect not many people would care too much about getting to beyond 2000), so I'm very happy with that. Though what you are doing re the third point almost makes it irrelevant for me. That is looking excellent, and you are definitely on the right track - as per my email.
    Also the strigggls are spot on now thanks. There will be quite a few people wondering where their extra Maxwell came from!

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •