Asteroid Zoo Talk

The "Azoo Updates" page. (June 09, 2016)

  • MvGulik by MvGulik

    General Azoo Updates updates post.
    (For AZ users, by a AZ stalker.)

    Ps: Changed the link above to go trough the "IsUpOrDown" check, less waiting time when its DOWN again. And if its UP, its just one additional click away. (I would bookmark it that way too btw)


    Update page versions:

    -- 07:10AM October 06, 2015 (age: 6 months.)
    -- 08:29AM April 30, 2016 (added new data & detection preview images, removed old mpc data.)
    -- 08:15AM May 04, 2016
    ++ Addition of more per-case-data on the main page. (no data or other changes detected.)
    Note: It seems save to read the current "Processing pending" record status as ... "Failed".
    -- 09:35AM May 08, 2016
    -- (Bug-fix update) ... see related post.

    -- 10:01AM June 09, 2016

    (not checked for changes)


    Various other stuff:

    -- Old original & locked down topic: Asteroid Zoo Updates: Reporting Site
    -- In case the update-page its down ... again: DownForeveryoneOrJustMe

    • [May 7 2016 6:45 AM] - DOWN and out for the count.
    • [May 7 2016 6:33 PM] - Back UP
    • [May N 2016 n:nn nn] - DOWN again. Right ... lets just presume its a feature.


    Some data on the 2016-04-30 & 2016-05-04 version:

    • It contains 3325 CSS targets,
    • for which 626 did not fail, and of those,
    • 298 have at least 1 new detection.
    • counting (906) new detection's in total.

    For the 2015-10-06 version those numbers where: 3233, 2536, 790, (20674).


    New feature images:

    ++ The new* included image view / feedback feature,

    ++ The new additional per-case-data on the main page. (as of 2016-05-04 version)

    -- The "Previous Detection's" - MPC "Submission Data" was removed.


  • hightower73 by hightower73

    looking through this page, working from the bottom to top, right up till thr failed red notice there are 20 submisions, would take a while to go through the top ones, but this looks very promising


  • MvGulik by MvGulik

    Good Vectors. Only those clusters which follow a linear, equidistant 4 point vector across the 4 frames.

    Mmm, looks like the linear part might be a bit iffy.

    01_12APR12_N07069 | 27


  • MvGulik by MvGulik


    version: 09:35AM May 08, 2016 (Bug-fix update)
    No significant general changes detected. (visual compare)

    • The two top new-detection cases where downgraded. (presuming linear bug-fix)

    • 01_12APR12_N07069 2714 51 24 27 (now 4)

    • 01_12APR19_N04088 3106 33 1 32 (now 1)

    • And two seemingly new CSS detection's where added.

    • 01_12JUN19_N04086 (0, +1)

    • 01_12JUN23_N04065 (0, +1)


  • MvGulik by MvGulik

    Duh ... Seems I finally, probably as last one, spotted that general main-page sort feature. Kinda wonder why such a feature would be implemented without any visual hints that its actually there. But than again ... there is no point in wondering about human behavior.

    (waiting with adding it as official feature when it sheds its ambiguous nature.)


  • MvGulik by MvGulik

    New Update-page date: 10:01AM June 09, 2016
    (partially checked for changes.)

    Some preliminary stuff: ("June 09, 2016" versus "May 08, 2016")

    • 103 new records. (now 3428 in total)
    • 2752 none-failed. (ergo: 676 failed)
    • ... (not sure yet)

    The Odd thing with this update is the overall increase of the Click counts.
    Compared to the previous version it increased with a overall 229%.
    That's more than 1M in clicks. (Yea, right.)

    Based on the fact that there are 997339 AZ-Sets. There should be 3451 CSS-MasterFrames.
    With a current count of 3428 MF's on the Update page, that's leaves 23 MF's missing/untouched.

    Mmm, Based on the total-vectors in the "May 08, 2016" version, and the previous-vectors in the "June 09, 2016" version. I either missed a public-version, or AZ just skipped one.

    Combined with the fact that old MPC-used data is removed, that mean a gap in my MPC data. Darn!
    O well ... was already missing the first data-version anyway.