Asteroid Zoo Talk

General project brain dump

  • peterbees by peterbees

    Hi zoosteroiders
    As a newcomer, it has taken me a while to accustom myself to the interface here. However after a couple of hundred image sets I am comfortable enough to make some comments to the researchers. I know that a lot of the following will be repeats of what has been better said by others. Nevertheless here goes.
    First, a couple of gripes:
    • I took me a long time to realise that ‘DONE” does not mean ‘done’. Hence I could not figure out how to mark more than one asteroid/artefact! How about ‘OK’ or something less final than ‘done’ for a post-marking response button?
    • Around 20 percent of the sets are bad or corrupted to the point that finding asteroid signals is difficult or impossible. Could we have a ‘bad’ button as an alternative to ‘nothing’ so that these sets could be removed or repaired? Otherwise a ‘nothing’ response may wrongly imply that there is no ‘hit’.
    • While I’m on the subject, how hard is it to make the images register properly? Very few of the sets have fully stable registration, which makes our job more difficult. On the other hand the colour cycler tool (which I find very useful in most circumstances) is obviously able to do the job.

    In a more general sense, you have some decisions to make as to how to treat your cloud resource (ie. Us). On the one hand there are statistical and practical advantages to the ‘set and forget’ approach, where you give a minimal tool which is simple and equally operated by anyone who is willing. However, you clearly have a variably motivated band of members – some of whom could add significantly to the quality of your research if they felt more as collaborators than remote tools, even if it takes more time and effort from you (the primary researchers). Some suggestions, humbly offered:
    • Provide the best instruments possible to do the job, within the constraints of the medium. A proper image zoom and pan function, for instance, and a brightness/contrast control. I’m sure there are other analysis modes which you have access to yourself.
    • A bit more in the way of training or resources to help in identifying the various things on the images. The user-supplied examples and commentary are good, if you can find them, but I don’t see why these could not be gathered together in an expanded tutorial. (Remember, a lot of potential participants will be completely unfamiliar with astronomical images and may be discouraged by the strangeness they see – and the difficulty of asteroid identification in most cases).
    • Why not, also, a links page to a range of asteroid-related literature, from basic to current research?
    • There may be more that we can do with the images presented, even if it is a bit away from the core study. To mark anomalous things which are not asteroids simply as ‘artifacts’ seems a bit of a waste. Maybe there should be some way of distinguishing between different types of artifact, for instance (I’d love to know what the dodging telephones are…). Or just an extra button for “Hey. Look at this, this is weird”.
    • Personally I’m not in favour of a score board for hero asteroid hunters, but there seems to be a lot of sentiment in favour. Of course you can’t confirm genuine asteroids, let alone new discoveries, but maybe an acknowledgement could be made when a plausible candidate has been noted. Actually I’d like to know when (if) I saw something that few or no others had seen.

    OK that’s all – feeling much better – off to do some more hunting.
    cheers

    Posted

  • Dr.Asteroid by Dr.Asteroid scientist, admin

    Thank you for your detailed feedback. It's enough that it will take me a bit of time to respond completely.

    One point I will specifically address right now: The interface. We had a very detailed discussion about how complex to make the interface. Clearly, there are a significant number of expert users who would likely make the most out of additional capabilities. But in order to ensure that we can understand how the process works, we need to have a simpler interface so we can cross-correlate which objects are found by multiple users. I'm open to a reconfiguration and certainly the early users have noted the desire for a zoom option.

    But first - we need to make sure that what we've got is working as smoothly as possible and then can discuss how to improve the interfaces.

    Posted

  • peterbees by peterbees

    Thanks for that.
    I do understand the statistical advantages of keeping a simple and consistent interface so that all users can utilise more or less equal expertise levels - although I'm not sure this is the most productive approach for your project.

    I forgot to add a comment about the frustration of being told that you have missed a known asteroid but not being able to go back to review the images to see what exactly was missed and hence being able to improve picking skills.

    Sorry about the formatting of the original post, by the way. I cut and pasted from another application.

    cheers again

    Posted