Revamping 'Best Nodes of All Time' List

[nysus]

I'd like to propose that the 'Best Nodes of All Time' list be expanded and/or changed. Currently, there are only 10 listed. I think this causes some problems.

For one, I don't think we have a true Top 10 Node list. What we actually have is a Top 10 best oldest node list. That's because those on the list get voted higher and higher simply because they are on the list. It's like Buddy Holly and Chuck Berry become the greatest rock stars ever only because they were one of the first to win admiration. Jimi Hendrix, U-2, or $insert_your_favorite_rock_band_here never get a fighting chance. As it stands now, a post must collect 140 XP in one week just to get on the list. This is a pretty tall order.

But more important than the accuracy of the list, however, is the fact that we are losing a lot of Perl's of Wisdom. Take Ovid's post, for instance. It contains a lot of great information for advanced users, but unless users are extremely persistent with Super Search, it will probably never be viewed again after this Wednesday. I feel like we're burning the Library at Alexandria.

I think there are a couple of different things you can do. The simplest is to simply expand the list to say 50 or 100. But this is only a temporary solution. It will take longer, but eventually you'll run into the same problem we have now with 10. A better solution, perhaps, is to start a "Best 10 (or 20) Nodes of the Month" and keep those around in an archive. But I'll leave the details up vroom if he wants to change it.

But do others agree with me that our current "Best Node" system is lacking and should be improved?

$PM = "Perl Monk's";
$MCF = "Most Clueless Friar Abbot";
$nysus = $PM . $MCF;
Click here if you love Perl Monks

[mirod]

Actually a lot of those posts are pretty recent, due to the inflation in the number of votes available each day.

Let's see what we have there:

• 3 obfuscations (all from the same guy BTW, I guess he's pretty good at that ;--),
• 2 nice pieces of code, quite long, doing interesting stuff
• 1 annoncement from the Monastery Chief Mad Scientist,
• 1 shameless post specifically designed to get into the list (shame on you Dominus!),
• 3 posts about coding techniques, 1 being more of a tour-de-force and 2 giving practical tips and info on how Perl works and how we can use it.

Overall a nice selection of what goes on here.

So I think that list can stay as is, just the rawreputation of nodes.

What could be added, and it should be fun to do, is separate lists by specific topics. I guess for newcomers something with "the best generic posts on coding in Perl", including Death to Dot Star! for example, would be nice. For those interested in the sociological aspect of the Monastery a list of the most heated discussions (anything involving saving the ChatterBox or publishing the list of users would do I guess...), and of course the funniest posts (how about Life at the Monastery. Chapter 1 for example), plus a best PM how-to's (turnstep's home node already has something like this, etc....

Now how to compile those lists? I don't think relying on the post's reputation would work, it's too dependant on parameters such as when the node was posted. This more of an editor job(not in the sense of a janitor, but in the sense of a publisher).

I think some of the oldest monks could compile a first selection, then set up a voting process outside of the site (so Vroom does not have to get involved in that phase), generate the lists (maybe over a period of time, one per week/month, and maybe revisit them evry now and then with the addition of worthy new nodes) and send them to Vroom who then has just to add a link somewhere at the top of the page.

Any taker?

[nysus]

Actually a lot of those posts are pretty recent, due to the inflation in the number of votes available each day.

You raise a valid point but I think it's questionable at best to automatically assume that the inflation of votes balances out the artificial votes cast for nodes advertised on the "Best Nodes" list. It would take some number crunching to figure it out. My gut tells me that in a couple of months it will be near impossible for anyone to ever get on the "Best Nodes" list. We'll have to wait and see.

$PM = "Perl Monk's";
$MCF = "Most Clueless Friar Abbot";
$nysus = $PM . $MCF;
Click here if you love Perl Monks

[tilly]

My gut based on having been here about a year is that the inflation does balance. Here are the nodes of mine that have managed to get near or into Best Nodes. (I really don't remember which made it there, but they were relatively close.) Not one survived there (though one is temporarily back thanks to the list being extended):
1. The path to mastery
2. Threads vs Forking (Java vs Perl)
3. The nature of work (this had the longest stay)
4. Why I like functional programming
5. Random thoughts on programming (Did this get there?)
6. Re (tilly) 1: How does strict work? (Another I am unsure on)
7. Stay aware of security
8. Is PerlMonks economically viable?
Also note that a node from last week is (as of this post) within 7 points of getting onto the top 10. Even after the last week nodes go out, there is some upwards momentum. If a few people referred to it as a good node elsewhere, those 7 votes could be reached. And with some more inflation, it already would have been reached.

No matter how impossible the goals in Best Nodes seems, don't underestimate inflation. The best nodes of today would have been on the best nodes of all time when I started. There is a lot of turnover in best nodes. Not if you watch it. But note the list, go away for a couple of months, and look again...

[footpad]

Well, there is a big note at the top of the node: Also, don't vote for the nodes listed here just because others thought they were good. If you do the numbers behind these nodes will just increase but we won't really be showcasing the best nodes.

I'll admit that it would be nice to be able to locate former BNOAT's more easily, I find it kind of entertaining to periodically pick a monk at random, sort their nodes by Reputation, and then see what their most well-recevied write-ups were. (Granted, I'm easily amused at times.) With folks like Ovid, Chromatic, and others, it's a great way to locate some of the hidden treasures in the archives.

I've also found it useful to browse certain home nodes. Turnstep, for example, maintains "things you should know" and periodically refreshes it as new material is added. Most of the people on this list have lists of good nodes to know (myself excepted).

Thus, the information is out there and can be located with a little bit of personal initiative using existing tools, which sort of fits in with our preferences for "Search first, ask questions later." If changes are needed to Best Nodes (and I'm not really sure that's true), then perhaps adding a setting to control the size of the Best/Worst Nodes list would be best.

Alternatively, someone could simply post and maintain a list of "Older BNOATs" similar to Outside Links. That would be the cheapest solution (for the overworked, at any rate).

--f

[Masem]

Here's an idea; if there was a list similar to Best Users or Saints in our Book, but for nodes, sorted on reputation *but* not showing this reputation. The list would either show all nodes with a reputation over N (N being at least 80, probably more reasonable at 100). Nodes that word their way into the Best Nodes of All Time would most likely never disappear from such a list.

The other option when a similar concept was brought up, was to have a Hall of Fame, allowing monks to select a node a week for Hall of Fame inclusion, and nodes that got more than 20 or so such votes by other monks would be honored here.

---

Dr. Michael K. Neylon - [e-mail] || "You've left the lens cap of your mind on again, Pinky" - The Brain

[nysus]

Well, you are not supposed to vote for a node just because it's on the Top 10 list but it does make them much more visible. Therefore, they will collect more votes. Imagine I posted the exact same node in two different places. I choose to advertise only one of the nodes. Which node will get the most votes? Obviously the one that is visited more often---not because it's better, but because it attracts more visitors. And I think this makes the 'Best Nodes' list pretty lopsided.

$PM = "Perl Monk's";
$MCF = "Most Clueless Friar Abbot";
$nysus = $PM . $MCF;
Click here if you love Perl Monks

[OeufMayo]

I actually whipped up a script last week, that is scheduled to run once a day, which fetches the Best Nodes and Worst Nodes nodes in raw mode. I don't have any post-processing done for the HTML I'm getting back, but it shouldn't be too hard to make a searchable archives for previous Best/Worst nodes.

#!/usr/bin/perl -w
use strict;
use LWP::Simple ('getstore');

my @d    = localtime;
my $file = sprintf("%4i%02i%02i", $d[5]+1900, ++$d[4], $d[3]);
for ('best', 'worst'){
        my $url  = "http://perlmonks.org/index.pl?node=$_+nodes&displa
+ytype=raw";
        my $html = getstore($url, 'bestworst/' . $file . "$_.html") or
+ warn "Cannot get $url: $!\n";    
}
exit 0;

<kbd>--
my $OeufMayo = new PerlMonger::Paris({http => 'paris.mongueurs.net'});</kbd>

[arturo]

Hmm ... is it possible to $NORM-alize the list somewhat? I've not been curious enough to figure out whether there's a history of $NORM outside of jcwren's stats, but there's yer basis for a function; a node that got 50 back in the early days could possibly show up on the list, that way.

I could swear this discussion's been had before, but I was again too lazy (bad way) to do a search on it beforehand.

UPDATE props to kudra, who pointed out that previous discussion (one I even took part in), wherein she suggested normalizing the reps for the 'best of all time' list. No wonder I stole it and presented it as my own =)

perl -e 'print "How sweet does a rose smell? "; chomp ($n = <STDIN>); 
+$rose = "smells sweet to degree $n"; *other_name = *rose; print "$oth
+er_name\n"'

[nysus]

Holy crap! Check out best nodes now. It's now the top 20. The Gods have spoken!

$PM = "Perl Monk's";
$MCF = "Most Clueless Friar Abbot";
$nysus = $PM . $MCF;
Click here if you love Perl Monks