Beefy Boxes and Bandwidth Generously Provided by pair Networks
Perl: the Markov chain saw

Perlmonks Threaded Article Viewer

by jcwren (Prior)
on Dec 15, 2001 at 20:13 UTC ( #132218=monkdiscuss: print w/replies, xml ) Need Help??

I sent you this file in order to have your advice...

Last night I started playing around with an idea that I've wanted to implement for some time, an idea named the 'Perlmonks Threaded Article Viewer' (or PTAV).

My ultimate goal (assuming it doesn't turn into a GUP), is to have a mechanism where a user can 'subscribe' to a thread, and when a new node appears in the thread at or below the article subscribed to by the user, the user will receive some sort of notification (email, /msg, whatever).

In writing the bit of code that I did last night, I realized other than that basic functionality, I have no idea how I want to present the information that's available. So you can pick a year, a month, a day, see all the root level nodes for that day, select one of those, and see an indented list of replies. But so what? It's data, but it's not really information.

The data in the database is dates, authors, node types, and titles. There is no reputation, no article text. The program will always link back to Perlmonks for the actual article text, etc. This is useful to know when considering how this data might be put to use. And no, THERE IS NO REPUTATION INFORMATION.

Basically, I'm looking for some input on this idea. How can this data be manipulated so it becomes information? Sure, we could reimplement Super Search, but what good is that? Other than the subscription mechanism, can anything useful be done? I'm sure there's some statistical information that could be pulled out, but I'm not really looking for that kind of idea. Anything like that will go into the next update of the Stats Pages.


And yes, I did steal the look'n'feel (such as it is) from Google

e-mail jcwren

Replies are listed 'Best First'.
Re: Perlmonks Threaded Article Viewer
by blakem (Monsignor) on Dec 16, 2001 at 00:03 UTC
    Looks very nice! The "perlmonks archives" contain some hidden treasures that unfortunately don't get enough attention.... I do have a couple of ideas for your interface though:
    • Include the number of replies (either direct replys or nodes in the entire thread) in the top level view... My custom newest nodes interface lists that information. It helps me keep ontop of hot topics, and also lets me give some extra attention to unanswered questions. Plus its trivial to implement.
    • This ones a bit trickier... Sometimes interesting subthreads sprout out of relatively mundane original threads. If the title changes significantly during the course of the thread it would be nice to see that on the top level view... noting that "Fun with closures" is burried under "My CGI is broken" would be nice to see at a glance.
    • The third suggestion I have would be to parse Best Nodes a few times a day, and flag threads/nodes that made the list. I'm sure the historical data is long gone, but going forward, I think it would be an excellent way to turn your data into information. --- if you could add this before I leave for the holidays, I promise to put it to good use when I get back ;-P
    • Update Another thing I could see myself using after a perlmonks absence is a way to select groups of threads based on keywords. Logging in after the holidays and getting a list of the top level nodes that don't contaion 'CGI' 'Win32' or 'Tk' * and with "I made best nodes" tags on them would be quite keen.
    * no offense, just picked a few strings that might be towards the bottom of my to-read list when I get back.


Re: Perlmonks Threaded Article Viewer
by grinder (Bishop) on Dec 16, 2001 at 04:16 UTC
    grinder screams

    This is simply brilliant!

    This is a marvellous idea. I've long lamented that it's so hard to dig up ye olde nodes in the monasterie. And with this I've been noodling around looking at some ancient history, and it's really easy. Funny how weird and wacky the early meditations were.

    Some random thoughts. It would be nice if...

    • the pages had meaningful <title> tags;
    • you factored out the "top level nodes" on each line, into a <th> heading;
    • ...which would let you go to a layout with four years side by side with the month breakdown indented and underneath each year. (The fifth year would go below, and so on) Otherwise the top level page is too sparse;
    • one could jump straight to the Perl Monks thread, once you reached the page with the nodes of the day;
    • you displayed the daily root nodes grouped by category (SoPW, Meditation, Code, Snippet, PMD);
    • I had a page with all my root nodes. Admittedly for the more prolific monks this could be a problem;
    • you made an XML ticker-style interface to let me probe the backend database (date=20010319&cat=sopw), (author=grinder&cat=meditation);
    • the page with the nodes of each day in the month was laid out seven columns across, and resembled the calendar. I actually parsed the output of cal to do this once, but I got better. Now I would use Date::Calc or a close cousin;
    • if there was a best of. The day with the most nodes, most questions, most meditations and what not.

    I'm still reeling. I'll sleep on this and see what else comes to mind. Anyway, if you implement some of these ideas, I have the feeling you'll be on your way to extracting real information from the mass.

    Do you want patches? :)

    g r i n d e r
    just another bofh

    print@_{sort keys %_},$/if%_=split//,'= & *a?b:e\f/h^h!j+n,o@o;r$s-t%t#u';
(jeffa) Re: Perlmonks Threaded Article Viewer
by jeffa (Bishop) on Dec 15, 2001 at 20:36 UTC
    I followed the link and must say that what you have so far looks very nice. My suggestion is to provide a simple text box to allow someone who knows the node id of the parent node they wish to follow to quickly 'subscribe'. The reasoning is that this user will probably be perusing PerlMonks, stumble accross a topic, and not want to have to drill down to that node.

    And of course, this brings up maintaining the list of 'subscribed' nodes for a particular user. Does the subscription 'run out' after a period of time? Probably not, so the user will need an interface to manage their list.

    Looks good so far! ;)


    (the triplet paradiddle)
Re: Perlmonks Threaded Article Viewer
by jlongino (Parson) on Dec 16, 2001 at 10:58 UTC
    You're off to a great start. I've got a suggestion but I don't know whether or not it can be implemented:
    • I think it would be most useful if the subscription portion of the program could be implemented on PM in the same way as our "Personal Nodelet". Where there would be a checkbox for "add current node to subscription list" link or button. It would be extremely easy, intuitive and convenient.
    I really like the suggestions Blakem made about giving thread depth levels too. Sometimes I find that a top node may not appear to be of much interest (perhaps just from the initial subject or title), but top level nodes with greater depths tend to include interesting offshoots.

    I find it difficult to use the date oriented approach as I'm usually interested in tracking thread by topic/content. I don't mean that as a criticism as I'm sure some people would find it useful. Kudos to you! This is a very difficult project to tackle.


(jcwren) Re: Perlmonks Threaded Article Viewer
by jcwren (Prior) on Dec 23, 2001 at 09:38 UTC

    I've added some more functionality to the PTAV viewer, located here.

    The search box now supports searching by node number or a user name. More clickable navigation has been added, and threads now have the number of replies listed out to the side of them (note that the counts are off until Monday morning, when the I update the database. But the node counts for the last week should be pretty much on).

    One thing that is still broken is that the clicky-links to the right of the search box don't yet do anything.


    e-mail jcwren

Re: Perlmonks Threaded Article Viewer
by belg4mit (Prior) on Dec 15, 2001 at 23:06 UTC
    Doh1 You broke it! Looks like you diabled the allowance of index.cgi as a default document in the cgi-bin directory?

    perl -p -e "s/(?:\w);([st])/'\$1/mg"

      I'm trying to figure out what planet you're on. The link works for me, and no one else has reported a problem, and there have been dozens of hits. What exactly are you talking about?


      e-mail jcwren

        The link provided works but the links for the years on the resulting page do not (check you error_log should be more than a few from 18...)

        perl -p -e "s/(?:\w);([st])/'\$1/mg"

      actually it's, but yes, it's broken... if you edit the URL and insert somewhere, all seems to work fine...

      ... Quidquid perl dictum sit, altum viditur.
Re: Perlmonks Threaded Article Viewer (Trend)
by toolic (Bishop) on Jul 30, 2010 at 00:30 UTC
    UPDATE: (Jul 31, 2010 at 01:11 UTC)
    It looks like there may be a bug in how PTAV is now calculating the Root Node Count. As a sanity check, I picked a recent day at random: June 8, 2010. Super Search shows that there were 18 root nodes created that day, but PTAV only shows 9. That's a big discrepancy.

    I picked another day at random: June 8, 2003. Both Super Search and PTAV match for that day 7 years ago (23 root nodes).

    I conclude that the recent PTAV results are incorrect, and the following discussion is invalid.

    Here is my original post:

      What say you, monks?

      Asking those that are still here will probably not elicit good explanations.

      Is it possible to go straight to the horses mouth and ask former active members to say why they are no longer such? Anonymously or otherwise. Perhaps those here that are still in contact could email them to ask them to post a short explanation of their reasons for moving on.

      Perhaps it could be wrapped up in a poll? (It'd be nice to see a useful one for a change :).

      Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
      "Science is about questioning the status quo. Questioning authority".
      In the absence of evidence, opinion is indistinguishable from prejudice.
        Nevermind :(

        See my update: Re: Perlmonks Threaded Article Viewer (Trend)

        Asking those that are still here will probably not elicit good explanations.
        You are probably right. However, you might be the perfect monk to recount some event in the history of the Monastery, given your longevity as a monk and your exceptionally high frequency of participation herein. "I remember the hurricane back in '05 that wiped out the barn." I'm not saying there was such an event, but if there had been, you'd be someone in a good position to remember it.
        ask former active members to say why they are no longer such?
        That would be insightful, indeed.
        Perhaps it could be wrapped up in a poll?
        Good idea. I doubt I'll get to it soon; so if anyone else wants to do it, feel free.
      For what it's worth, I was quite active in these parts in the early naughts, after which I went on hiatus for a few years. Recently I've been poking around the halls again.

      In my case, there was no particular reason, just changes in life, work, interests. More recently I had to become proficient in p-p-p...p-p-p...Python! (don't tell anyone) Naturally I spent less time with perl during that.

      Even more recently, I started dabbling around with some XS extensions and perl, plus have begun experimenting with perl6. Messing about with perl again naturally led me to wonder what the monks have been up to, so here I am.

      Things do seem slower, though. Personally I think EV2 is showing its age in this 2.0 world.


        Things do seem slower, though. Personally I think EV2 is showing its age in this 2.0 world.
        Good point. Maybe the response time of the site started degrading in the 2005-2006 timeframe. Several people have voiced their impatience over the years. It can be aggravating -- but I still think it's worth the wait :)

        Thanks for sharing your experiences otherwise.

        EDIT: The problem has been remedied and the numbers PTAV reflects are accurate (again)

        Interesting. Although I haven't researched this, I would likely suggest it's a problem in the XML feed, from which that data is derived.

        If that's indeed the case, I may as well take it off line, as it's only wasting bandwidth.


        e-mail jcwren

      What say you, monks?

      Piss off :) There is too many possible reasons and overwhelming lack of information that the problem defies quantification

Log In?

What's my password?
Create A New User
Domain Nodelet?
Node Status?
node history
Node Type: monkdiscuss [id://132218]
Approved by root
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others making s'mores by the fire in the courtyard of the Monastery: (3)
As of 2023-10-02 21:09 GMT
Find Nodes?
    Voting Booth?

    No recent polls found