Space Warps Talk

WHEN are the bugs going to be fixed?

  • N5bz by N5bz

    I still can't use dashboard (black image), I still get repeated images and need to log out and back in, I still get told I didn't mark a sim that I marked. I still see misplaced sim pointers.

    Is anyone working on fixing these problems?

    Is there anything we can do to help????

    Posted

  • anupreeta by anupreeta scientist, admin

    We understand your frustration. We have reported all these bugs and the tech team is working on fixing them. Some are partly fixed and some are difficult to trace. You can certainly help in many ways (read below).

    Could you ever use Quick Dashboard? if yes, please collect the images in Talk for which you get the black image in QD. If no, then please switch to Chrome browser and let us know.

    Some images are expected to be repeated (e.g. Training images ) if you have inspected a LOT of the images already. But other images should not appear over and over. Please collect these images in Talk as well when reporting this bug.

    General note: Please don't keep a session open for too long this may cause some apparent bugs. After a few hundred images, restart your browser and come back to classification.

    Sim mismatch issues are still being looked into. Please put markers on the lensed images not on the lensing galaxies. Again, please collect those images which tell that you didn't put marker correctly along with a snapshot of where the marker was placed.

    Thanks for your understanding!

    Posted

  • N5bz by N5bz

    I have never been able to use dashboard.

    I AM using Chrome browser and have been since I first signed in. Currently using 17.0.963.83

    I notice repeated images, sometimes separated by one or two others before the repeats. When I notice things repeating, I usually log out and back in. That seems to fix the problem. Often the 'repeated image' problem seems to go along with responses that do not correlate with the image I am seeing.

    It is as if Spacewarps thinks it sent one image while I am seeing a different image on my screen. One that was sent earlier?

    As a matter of fact, I suspect that when this happens, the 'image' that goes into my 'history' is NOT the one I was seeing. I suspect that the one in the history is the one I was SUPPOSED to be seeing. Does this make sense?

    I will see if I can nail things down any more closely. Is there someplace specific in the javascript console or other tools that I should be looking for clues?

    Perhaps some screen snapshots might be helpful? Someplace to which I can upload them?

    Posted

  • anupreeta by anupreeta scientist, admin

    Someone said that the QD started working on chrome after doing this --"... it needs "accept third party cookies" turned on. All is well now."
    Please see
    http://talk.spacewarps.org/#/boards/BSW0000001/discussions/DSW00001gr
    and
    http://talk.spacewarps.org/#/boards/BSW0000001/discussions/DSW00003c9
    to know what actions to take for any other problems that you are experiencing. Please add your reports under relevant existing bug reports. Thanks!

    Posted

  • N5bz by N5bz

    It would help if I knew which site(s) to put under the cookie block exception.
    Not sure I want to accept ALL third party cookies!
    (however, I tried it briefly and it didn't seem to fix the problem!).
    I read those threads a while back.
    Even posted a few comments about the double image problems a while back. 😃
    Thanks for your responses! Try not to let all the problems warp you! 😃

    Posted

  • anupreeta by anupreeta scientist, admin

    I assume you'd add exception just for - http://spacewarps.org/ since that's where the QD is not working ?!
    what machine / os are you using?

    Posted

  • N5bz by N5bz

    Using a dell Inspiron 640m laptop running windows 7, Enterprise, with service pack 1 installed.
    I tried adding that exception. Did not fix the problem with Dashboard.
    I still see only a black square when I invoke dashboard.

    Posted

  • akapadia by akapadia

    @N5bz Try adding an exception for http://spacewarps.org.s3.amazonaws.com

    Posted

  • N5bz by N5bz

    No joy. Anything specific I can look for in the Developer Tools?
    Here is some stuff I captured.
    https://filestogeaux.lsu.edu/public/download.php?FILE=bzinn/25383WyPda2
    There is a picture of the dashboard showing what I see (...50)
    There is a picture of a 'simulated lens' that shows nothing useful (...25)
    There is a picture of a sim-lens that I marked but seem to have missed (...32)
    There is a picture with TWO lenses, one of which I marked, but the 'credit' for marking is 'misplaced' (...46)
    There are several pictures that I think were 'duplicates' of previous pictures but not sure now why I captured them.
    There are several captures from the console that may, or may not be of interest to you.

    Posted

  • akapadia by akapadia

    @N5bz there are many server address that we utilize that would appear as third party to you browser (e.g. http://spacewarps.org.s3.amazonaws.com). Ensure that

    Are all permitted domains. Based on zooniverse lens problems\20130709 CONSOLE ALL classify, looks like our server had a momentary hiccup or the internet connection dropped, not a big item to worry about.

    There is one screen shot that contains two simulated lenses. Unfortunately the code only recognizes the first. The only item that we can work through at the moment is lack of Quick Dashboard. A colleague had a similar issue, I will ask tomorrow.

    Posted

  • akapadia by akapadia

    @N5bz Might you be using any browser extensions that forces HTTPS for all request (e.g. HTTPS Everywhere). If so, that will causes the Quick Dashboard not to work.

    Posted

  • N5bz by N5bz

    I don't think any of the extensions are forcing HTTPS for all requests.

    Here are the extensions I have loaded.

    Developer mode

    LiveMap UR Overlay1.2 Overlays UR markers onto Livemap Allow in incognito
    Papyrus: Street to river1.0 This script create a new river landmark in waze editor papyrus. It transforms the the geometry of a new unsaved street to a polygon.
    UROverview1.26 Allows UR details to be viewed by hovering over the UR marker
    Waze Livemap Navigation0.72 Display the navigation list in the Livemap on Waze.com
    WazeEdPlus1.2 This script helps you adjust the position of underlying bing maps
    WME Color Highlights1.6.2 Adds colours to road segments to show their status
    WME Junction Node Fixer0.0.7.6 Creates a new editor hotkey to lock turns, fix reverse connectivity, and restore original restricted turns.
    WME-Add-Ons0.44 Tools for Waze Map Editor

    Posted

  • akapadia by akapadia in response to N5bz's comment.

    @N5bz Difficult to tell as I don't know how those extensions work. You may want to disable all extensions and check again. If that works, then one extension is interfering when requesting data from our servers. We saw this happen with HTTPS Everywhere.

    Posted

  • N5bz by N5bz in response to akapadia's comment.

    No joy from turning off all extensions and taking out of the developer mode.
    As for the simulations, I suggest that they not be near the edge of the picture. Lenses near the edge or even on the edge should be picked up on other images (I am assuming that there IS a 10 or 20 percent edge overlap on images so that objects near the edge show up close to center on other images.

    I also suggest that simulations never be added to any image that has a real lense, so that the problem with 'a simulation lense being marked but flagged as wrong' because there is another simulation on the same image should not occur.

    Posted

  • anupreeta by anupreeta scientist, admin

    @N5bz There is indeed an overlap between images. Any object at the edge in one image will most likely be visible in the neighbouring one. This remains true both for real objects and sims. Sims are added to real galaxies, hence, there's no difference between the images with and without sims.
    On your second point, we don't know where the real lenses are to begin with and our main goal in SW is to find them. So, we can't avoid having sims+real lenses occurring in the same images. When there's more than one sim in an image, the feedback system can deal with notifying you about any one of them. So, I'd not worry about those cases. Also, note that multi-sim is a small fraction of total sims that you see. In the post-processing step, we do use all your markers from all sims, just that there's technical difficulty in providing feedback for multiple sims in the same image.

    Posted