http://www.perlmonks.org?node_id=498717

Hi all

I got an idea for PM this weekend.

Most of the replies to my question from fellow monks and saints have been something like ...

Please refer a link .. which already has the answer for the question you asked.

This gave rise to the idea of listing similar links when a person tries to create a node.

Well explaining it better...

1) When a person tries to create a node and preview it , there could be an option of what is the question on. For ex - .csv ( key word )and below the preview section Super search could work parallely to list out the matching links which has already been posted on the same.

2) I am curious to know the feasibility of such a feature going in PM since i feel it has the following advantages

a) Monks posting their questions can have an answer before they even post it.

b) Monks could go ahead creating a node if they are unable to find a match.

c) This could decrease the count of nodes and the load on the database.( by decreasing the duplicacy level )

Countering probable comment on this -

"When there is a super search , why should this be ?"

would be most of your thoughts/comments/approach.

Well i still feel super search is the strongest i have seen and all i am telling is just to link Super Search which runs on the background when anyone tries to create/preview a node.

I hope i would be alive to see the golden jubilee celbrations of PerlMonks.org but also i would like to see that there arent a million nodes out of which there is a contribution of different people asking the same question.

Again I could be wrong as I am tester.I would like to highlight a similar feature available in Gmail which displays only relevant ads when you recieve a mail or create a mail.

Again ! we arent copying them , we are better than them in linking a search to creation of node.

You are welcome to comment on this and I just request you that to be a learning for me if YOU DISLIKE/OPPOSE this proposal.

Regards

Prad

If our constitution allows us free speech, why
are there phone bills?
Update :: Could you all let me know what is wrong with this as it is getting severely downvoted..i expect replies to people who downvote since i can know where i was wrong , downvoting without an explnation is vague.
  • Comment on New node - Supersearch - Feature proposal !

Replies are listed 'Best First'.
Re: New node - Supersearch - Feature proposal !
by castaway (Parson) on Oct 10, 2005 at 09:29 UTC
    I quite like the idea of running a search while someone is creating a node, and showing the results while they are previewing it, I've seen it used to good effect in other forums.

    The small problem is that Super Search itself only searches in chunks, and thus would be difficult to integrate. The reason it does this is that it is quite database intensive. So we would need some better/quicker way of searching for results.

    The other small problem is, that users can already do a search before they post, and yet either they don't, or they fail to find what they are looking for, and still post. So requiring a search may not have a useful effect. People will have the wrong keywords, or not know what to look for.

    Having said that, I have wanted to have some sort of introductory text at the top of Seekers of Perl Wisdom incouraging people to search first for a while. Larger, flashier hints are needed than the small text under/near the textarea for posting IMO..

    C.

Re: New node - Supersearch - Feature proposal !
by Happy-the-monk (Canon) on Oct 10, 2005 at 09:32 UTC

    let me know what is wrong with this

    The general idea is that people should have been reading and searching for answers before they post questions.

    If they already did, they might be a little bit annoyed by some wise guy "artificial intelligence" telling them to look somewhere they already looked... but that's not the biggest problem.

    It would rather be challenging to devise the artificial intelligence to understand somemonk's request in the way it was meant. See, that's hard to do for people, it will take a lot of effort to make a machine that wise and insightful.

    Next thing, if the search takes too long, it might scare the poster away - he might find the process too tiresome.

    I find your post intriguing though, but I did not think it worth a vote - yet =)

    Maybe you might spend time to narrow in on how PM should be able to "answer questions in advance" ... right, I am getting carried away now.

    Cheers, Sören

      Howdy!

      If they already did, they might be a little bit annoyed by some wise guy "artificial intelligence" telling them to look somewhere they already looked... but that's not the biggest problem.

      Immediately brought to mind the horrifying thought articulated in the Sunday, 9 October 2005 User Friendly.

      yours,
      Michael
Re: New node - Supersearch - Feature proposal !
by woolfy (Chaplain) on Oct 10, 2005 at 09:26 UTC
    You propose a technical solution to what basically is a behavioural problem. Whenever somebody wants to ask a question, (s)he should first use SuperSearch before (s)he asks the question. That should be default behaviour. Trying to fix wrong behaviour with a technical solution like you proposed, would be the wrong approach. That's why I downvoted your node.

    The technical solution might become a nightmare, by the way. It might not be so much work to implement, but it would be a nuisance to the more experienced monks to use this. And running Super Search on the background all the time when a monk creates a new node, could be quite a strain on the system.

    There are quite some nodes on "how to ask a question". Newbie monks are pointed to them all the time. They should learn to use (and search) the Perl Monks website and they should behave the right way as soon as possible, and they should not be helped with a new "system" that they should not need anyway.

    (Update: changed spelling a bit.)

      I don't agree that a behavioral problem can't be, discouraged, for lack of a better term, with technology. Behavior and interface design, which is implemented with technology, are quite inter-related. Design an interface that sucks, and you can pretty well guarantee that users will not use it properly. Even after very carefully designing the system so that it is easy to use it appropriately isn't going to keep some from using it improperly. But, the second interface should provide self-reinforcing feedback (ie, if correct use feels natural, and is rewarding, most people will use it correctly), and thus partially "solve" the behavioral problem.

      So, am I bashing PM? No, I'm speaking in generalities. Is there room for PM improvement? Yes. Is the OP's idea something that should be implemented? I'm not qualified to provide a definitive answer.

      However I do think the question being asked, and resulting discussion is a good thing for PM. I do hope some sort of interface change can eventually be made such that new users are better encouraged to search first, and ask questions later.

      -Scott

        I disagree with you. Wrong behaviour can only be corrected in a limited way with technological changes. People who don't want to search and read first, but who want answers quick and easy, just are not going to search and read. They will ignore any search results that might be presented when the solution is offered that the OP has thought of.

        More experienced Perl Monks often answer questions with a link to a thread with a question and accompanying answer. By doing that, they educate the monks that ask those easy to answer questions. Search first, and if nothing is found, or if things are still not clear, only than ask the question. Most monks learn this lesson quick.

        Presenting the monk with a tool that offers search information while the question is being formulated, does not educate the monk, it is spoiling the monk, while putting an unnecessary strain on the system.

        An old saying applies here: the whole village is needed to educate a child.

      They should learn to use (and search) the Perl Monks website and they should behave the right way as soon as possible, and they should not be helped with a new "system" that they should not need anyway.

      Something about this bothers me. Isn't the whole idea of computing and automation to perform tasks that, yes, I could perform myself, but why should I if a computer can do it for me? Then I have more time to do the things that a computer cannot do, and which only I can.

      Your comment about the technical problems involved in providing this service is valid, because we don't have infinite resources, but I have to disagree with your philosophy that computers shouldn't be used to automate repetitive tasks because people "should behave the right way".

        I really do appreciate all those little and not so little innovations that make my life easier and enjoyable. Like my portable DVD player, which I use to see movies and tv series in bed.

        I certainly see the pros and cons of imbedding the super search technology into the page for writing a new node. It's not just making it easier to write a good node. It's not just putting more of a strain on the servers of Perl Monks. It's not just behaviour. It's all this and much more.

        If somebody doesn't want to read related nodes before posting a node, (s)he will not read the suggested related nodes.

        If somebody does read the suggested related nodes, and use them to adapt his/her new node, the new node might become extremely detailed, complex, long, unreadable, and difficult to answer. Even when the node is already an answer to itself.

        I think many monks would ignore this new proposed option.

Re: New node - Supersearch - Feature proposal !
by marto (Cardinal) on Oct 10, 2005 at 09:30 UTC
    prad_intel,

    In practice would this work?
    This would not be a problem if people used the site properly :)
    Your node Sample of Win32::GuiTest needed for this was titled "software upgrade for mobile phone" before you changed it, and the question several times, after people had replied to it.
    How would your proposed system deal with this?

    Matches you would have been given based on keywords of your title alone would not have helped you, and as you updated the question several times (mostly at the request of various monks, to provide a structure and explain what you were actually trying to do) I dont see how this feature would have helped you.

    I for one dont think that it is too much to ask people to Super Search before posting.
    But as I have attempted to illustrate, the site is dependant on people using it properly, creating nodes which are well structured and contain information relevant to what the poster is trying to achieve.
    As far as downvoting goes have, a read at Why downvote nodes without commenting on them?.

    Martin
Re: New node - Supersearch - Feature proposal !
by 5mi11er (Deacon) on Oct 11, 2005 at 14:46 UTC
    While most of this thread is hung up on how a forced search is not a good idea, I applaud Prad for suggesting ways to improve PM.

    I think Castaway's suggestion:

    Having said that, I have wanted to have some sort of introductory text at the top of Seekers of Perl Wisdom incouraging people to search first for a while. Larger, flashier hints are needed than the small text under/near the textarea for posting IMO...
    is definitely on the right track. Stretching the idea a bit further, perhaps the Monestary gates themselves need an update. Rather than displaying the most recent posts, should it instead maybe ask the user what they are searching for?

    Welcome to the Monestary. The Monestary has many things to offer; our greatest treasures are those that choose to give back to the Monestary. What is it you are seeking?

    • How do I :[text entry box]
    • My code is broken
    • I have this homework assignment...

    The first would be a way to submit a super search, the 2nd would be a link to a page explaining how to properly fill out a question form with the use of code tags and a form for doing so. The third would be a link to a page explaining that we won't do the homework for them, but will offer assistance to those who show they've done some work on their own but are stuck on something specific. Obviously there could be several more items in the list.

    Anyway, I'm throwing this out, hoping it has some rubber content such that it is able to bounce around and become an even better idea.

    -Scott

Re: New node - Supersearch - Feature proposal !
by ambrus (Abbot) on Oct 10, 2005 at 20:28 UTC
Re: New node - Supersearch - Feature proposal !
by phydeauxarff (Priest) on Oct 10, 2005 at 21:10 UTC
    a Super Search integrated into preview???

    I am sorry but....

      ECK!
    I can't even begin to imagine a way this could be implented without making the preview process take an interminable amount of time to execute.