locked
Unreliability of Microsoft Academic Search / Better Error Messages RRS feed

  • Question

  • First, I want to say that I am immensely impressed with what has been done with Microsoft Academic Search.  It is great to finally have someone working on a better academic research search tool.  The other one (that is at the G company) appears to be completely stagnant and they don't even advertise it as one of their products.  Competition in this space is not only good for both products but I believe that investment in these academic tools actually will directly advance and benefit humanity.  Props to Microsoft for taking the time, effort, and financial investment in such a project.

    With that said, it is frustrating how unreliable MSAS is at this time.  I was trying to show it off to fellow classmates and a professor when suddenly it appeared that everything just stopped working.  Links that previous I had used just stopped returning anything except an error:

    File Not Found

    Return to Homepage in 1 second.

    Most likely causes:

    • The requested item (paper, author, conference, journal, or organization) does not exist.
    • The link might be out of date.

    While I realize that the product is in Beta and not yet released this behavior and error messages lead users to believe that they are doing something wrong.  Maybe adding another bullet reminding people that this is a product under development (Beta) and not fully functional would be helpful?  Something saying it could be a system error and to try again later might be less confusing than telling people a link that worked 5 minutes ago now doesn't exist or is out of date.

    As another example, I just tried another search from the home page and it didn't even work.

    http://academic.research.microsoft.com/Search?query=skinner&SearchDomain=7

    We did not find any result related to "skinner".

    Search Tips:
    • Make sure words are spelled correctly.
    • Try rephrasing keywords or using synonyms. E.g. use "face detection", instead of "face identification or search".
    • Try less specific keywords. E.g. use "decision tree", instead of "arc-4x adaboost decision tree".
    • Use concise queries. E.g. use "neural network", instead of "recent papers about neural network".

    Friday, May 1, 2015 7:57 PM

Answers