Discosearch's and/or @tar's failings



  • Continuing the discussion from Local Dischorse CSS:

    @tar said:

    I'm going to :fa_bookmark: this thread in the naïve hope that I'll be able to find it again, after noting that a discosearch for motherfucking CSS wasn't able to...

    :cow2:


  • Banned



  • But you know what, fine, be all passive-aggressive about it. It's only your software's reputation on the line.




  • sockdevs


    Y'know, bug reports work best when they're reports of bugs :stuck_out_tongue:



  • :hanzo:



  • So you guys get different search results to me, then

    Dicsourcistency.



  • Did you click "Show More"? The size of your scrollbar suggests not.



  • Well, no, i just assumed that seeing as the topic was made today, it'd be in the first three or four results. I can hardly be blamed for assuming that search doesn't work, it's been pretty much useless as long as I can remember...

    How often do you go onto the second page of Google search results?


  • sockdevs

    @tar said:

    How often do you go onto the second page of Google search results?

    Every time the first page doesn't have what I'm looking for?

    OK, not every time, but often enough.



  • Alright, let me rephrase my question, then. How often does Google not have what you're looking for on the first page?

    I'm fairly certain that 90% of the time what I'm looking for is in the first 2 or 3 items of the first page...


  • Winner of the 2016 Presidential Election

    You clearly don't search for the same things me and @tar do. If I cared I could probably count the number of second-page-on-Google-search instances I had last month on my fingers.



  • @tar said:

    How often do you go onto the second page of Google search results?

    Dischorse shows the top 5 results initially. If Google only showed the top 5 on the front page, I'd go to the second page lots.
    I have no idea how Dischorse prioritises the topics, whether it's by likes or replies or what (@sam?) but I'd not give up if something wasn't in the top 5 results.



  • Google shows 10 results per page. This is in the top 10 Discoresults when searching CSS.



  • @loopback0 said:

    Dischorse shows the top 5 results initially. If Google only showed the top 5 on the front page, I'd go to the second page lots.
    I have no idea how Dischorse prioritises the topics, whether it's by likes or replies or what (@sam?) but I'd not give up if something wasn't in the top 5 results.

    Alright, fine, anyway, as I pointed out, I have this topic bookmarked now, so any inadequacies of Discosearch and/or myself are largely moo at this point... :new_moon_with_face:



  • @loopback0 said:

    Google shows 10 results per page. This is in the top 10 Discoresults when searching CSS.

    I would still expect some prioritization of recent topics, I don't expect to have to dig through pages and pages of month-old stuff to find a topic which is what, 2 or 3 hours old...



  • #Consider replying to several posts at once
    ­
    #Rather than many sequential replies to a topic, please consider a single reply that includes quotes from previous posts or @name references.
    ­
    #You can edit your previous reply to add a quote by highlighting text and selecting the quote reply button that appears.
    ­
    #It's easier for everyone to read topics that have fewer in-depth replies versus lots of small, individual replies.

    and again with this fucking shit....



  • Now that I'll agree is fucking annoying.


  • :belt_onion:

    why do the toasters love that @name guy so much anyway.
    :fa_facebook: him.



  • The point remains that some of us have such low expectations of Discourse at this point that we can hardly be blamed for assuming that something which was apparently recently fixed was still broken...


  • Winner of the 2016 Presidential Election

    Seriously! It should be @mention. Sheesh



  • yay, lets make it all about @me!


  • Winner of the 2016 Presidential Election

    @tar said:

    yay, lets make it all about @me!

    Should I register @me and make him respond back with his summoner's username? Yes, @summoner, it's all about you

    Filed Under: I wanna do that now


  • sockdevs

    @tar said:

    Alright, let me rephrase my question, then. How often does Google not have what you're looking for on the first page?

    Often enough :stuck_out_tongue:
    @sloosecannon said:
    You clearly don't search for the same things me and @tar do.

    Well, that's a given :laughing:



  • And also, I get that it's frustrating to hear people saying your software is broken. (Even if it is mostly broken in a lot of different ways).
    But you're an ambassador for the product, so being anything less than polite is going to end up reflecting on you poorly. It doesn't cost anything to say something like:

    Here's our new search feature, perhaps you're not aware of it? blah

    Manners don't cost anything.



  • I wanted to register @name but it turns out there's already a user called @name :laughing:



  • They must be so sick of notifications by now...


  • Winner of the 2016 Presidential Election

    Believe it or not, they've only been @mentioned 4 times. All done today.


  • mod


  • Winner of the 2016 Presidential Election

    @sloosecannon said:

    Should I register @me

    Don't think you can. 3 char minimum IIRC, unless @PJH changed it to less. If that's at all possible, that is.


  • Banned

    Probably going to add a bunch of extra weighting for "most recent" when searching, but you can always search for css order:latest then the result is #1, just like USA


  • Winner of the 2016 Presidential Election

    Unique, uncommon words will always produce the best results

    Ooooh!

    ... fucking nerds ...



  • Well, that's obviously the best result, isn't it?



  • @tar said:

    Manners don't cost anything.

    Can I bring you with me to the pharmacy when I pick up my next blood pressure medicine refill?



  • @loopback0 said:

    I wanted to register @name but it turns out there's already a user called @name :laughing:

    Imported user. Joined in 2013, so if he made any posts, they haven't been imported yet. I don't remember him, so could be an empty spam account.





  • Posted at 2007-06-02T22:08:25.28 Re: Java, diyGC:

    A few facts then: Space is limited (although rather large) and x,y,z are integers. Objects however are not limited (which is why I never even thought of trying to compare them directly, since that would scale terribly?). So yes, performance and especially scalability are issues :)

    Also this is not only used for collision detection, but also to display partial maps etc - having to loop through all objects to find out which of them are 'in scope' seems rather odd?

    Posted at 2007-06-03T12:30:49.42, same topic:

    @Nether said:

    @asuffield said:
    @TDC said:
    Assume a (large) 3D space with positions x,y,z and objects moving in it. Since I needed an easy way to answer the question "which objects are on a specific position?"

    Heh... this is a huge problem that comes up all the time (notably games and physics simulations) and none of the thousands of people who have worked on it have found any really good solutions yet. You're going to have fun with this one.

    If you know the set of objects in your world isn't going to be arbitrarily large, you can do direct comparisons between the position coordinates in your objects. If the resolution of your logical space is integer-valued, this can go pretty fast. Professional solutions seem to take the same basic approach but try a lot of clever tricks to speed it up. You can dynamically size your logical space into buckets along each dimension, or use static buckets along each dimension. You can group your objects into logical sets if not every object has to worry about colliding with every other object. You can keep ordered lists of your object positions along each dimension and update those lists on every tick of the universe.

    ...

    If anyone else knows some clever ways to approach the problem I'd love to hear them. My search for whitepapers back when I first worked on this yielded nothing fruitful.

    A popular approach is the use of an octree to
    logarithmically reduce the number of comparisons you need to find
    objects which are "near" another object. Ultimately, nobody's found any
    better method than picking a bunch of tricks like this and combining
    them in some way that makes sense for your problem - and every variation has to be hand-adjusted to suit the specific system you're currently trying to simulate. Many books have been written on the subject of how to approach this sort of thing, particularly for the specific variant used in collision detection. Lots of different approaches are known, but none of them are both generic and efficient - there aren't any good methods which will work on every problem of this type. It seems like one should exist, but it hasn't been discovered yet.

    Given the scope, significance, and relatively recent nature of the problem (in academic timescales, anything that wasn't a problem 50 years ago is too new for anybody to have a good understanding of it yet), it is a subject of current research, and more in the realm of academic papers than whitepapers. Here's a research project that has done a lot of significant work in the field; they have a list of papers linked from there, and you can find the rest of the field via citeseer. It's enough to spend a lifetime just studying this problem.


  • mod

    @Onyx said:

    fucking nerds

    Why are you looking for pron on the forum? Especially nerd porn. Awkward.



  • i am missing something obvious.

    but BELGIUM it.

    this guy never ever read a book on computer generated graphics and it's talking like he knows something.
    http://s2.n4g.com/media/11/news/890000/894648_0.jpg


  • area_deu

    You can always manipulate the search.
    See: lettuce post


  • Winner of the 2016 Presidential Election

    @abarker said:

    Why are you looking for pron on the forum? Especially nerd porn.

    I like big brains and I cannot lie.



  • @sloosecannon said:

    Seriously! It should be @mention. Sheesh

    Hey - don't pick on me!


  • Discourse touched me in a no-no place


  • Winner of the 2016 Presidential Election

    There is no damned way to get a single result, is there?


  • sockdevs

    @Onyx said:

    There is no damned way to get a single result, is there?

    You just have to know what to search for:
    [spoiler][/spoiler]
    …there have been some unusual discussions in the past…


  • Winner of the 2016 Presidential Election

    And if I got the context right... that's actually my post, including oneboxed Wikipedia articles...

    Why do I remember such shit and can't remember the name of that guy I went to highschool with that I met in the train the other day...

    Fucking brains, how do they work?


  • Winner of the 2016 Presidential Election

    @RaceProUK said:

    @sloosecannon said:
    You clearly don't search for the same things me and @tar do.

    Well, that's a given :laughing:

    Google deprioritises furry searches? Who'da thunk it?


  • sockdevs

    @Onyx said:

    that's actually my post

    Correct ;)
    @Onyx said:
    Why do I remember such shit and can't remember the name of that guy I went to highschool with that I met in the train the other day...

    The brain has a remarkable ability to retain the banal and uninteresting, while at the same time rejecting the useful and important :stuck_out_tongue:
    @Onyx said:
    Fucking brains, how do they work?

    "When a mommy brain and a daddy brain love each other very much…" :laughing:



  • I Jeffed 11 posts to a new topic: Jeffed from Discosearch topic


  • sockdevs

    @RaceProUK said:

    …there have been some unusual discussions in the past…

    yep..... the lounge is a weird and wonderful place. ;-)


Log in to reply
 

Looks like your connection to What the Daily WTF? was lost, please wait while we try to reconnect.