in reply to Re: Terminal decline? in thread Terminal decline?
they visit regularly for a while, then drift away, depending on the demands of their life. However, new members join, with some old and some new questions.
I agree it was always so.
But my impression has been that for the last few years; new people come, stay for a shorter time and don't some back.
In addition; the threads are shorter; there is less open debate; less to and fro and interchange of ideas. Basically, just less.
And when I went looking; the stats; every which way I viewed them, backed that up.
I'd love a couple of hours read-only access to the DB; then I could really extract some useful -- though probably depressing -- numbers.
But, those I've already looked at tell enough of the story to allow me to be confident in my conclusions.
With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
Re^3: Terminal decline?
by jdporter (Paladin) on Jun 12, 2015 at 17:11 UTC
|
One class of statistic which, I think, readily paints the picture is the reputation of undisputedly popular posts. The Best Nodes over the past year have rep in the 50s to as high as 90. In the old days, breaking 100 was not unusual. In contrast, today's 50 Selected Best Nodes all have rep over 99; 5 are above 190, the highest being 260. Not one of those 50 is later than 2008. Of course, nodes which have been around since 'the old days' have had more time to collect upvotes; but I believe that in general the vast majority of votes are cast on a node within the first few weeks of its existence.
Another way to see something similar might be to look at the reputation of your own "best nodes". My 49 best nodes are all 2010 or earlier; and all of my nodes which are better than 90 rep (there are seven of them) are 2007 or earlier. No post of mine has hit 40 rep since 2012.
I am not a prolific poster, though; someone like BrowserUK might have a more meaningful trend line.
(Update: s/XP/rep/g for nodes.)
I reckon we are the only monastery ever to have a dungeon stuffed with 16 ,000 zombies.
| [reply] |
|
180320 219 2002-07-08 20:44
262941 208 2003-06-04 10:31
409922 143 2004-11-23 15:57
433614 134 2005-02-23 08:26
614216 127 2007-05-08 18:44
532956 125 2006-02-27 03:57
678668 97 2008-04-06 18:45
876153 83 2010-12-09 02:50
735910 72 2009-01-13 10:25
995694 71 2012-09-26 04:28
939629 61 2011-11-23 10:09
1128661 61 2015-06-01 22:09
1060710 55 2013-11-01 05:20
1090675 55 2014-06-20 19:28
Which is even more depressing if we can assume that at my best; I've gotten better year on year.
With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'
Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
| [reply] [d/l] |
|
FWIW, here are my highest reputation nodes of the year (2002-2015) (update: 2nd rep is as at 1-1-2018, as are entries for 2016 and 2017; 3rd rep is as at 15-05-2021):
- 2004 (337, 344, 349): Saturn
- 2005 (228, 230, 232): The Lighter Side of Perl Culture (Part IV): Golf
- 2006 (193, 198, 200): On Interfaces and APIs
- 2002 (179, 180, 182): Somersaulting camel
- 2007 (151, 154, 155): Dueling Flamingos: The Story of the Fonality Christmas Golf Challenge
- 2008 (146, 147, 147): Unix shell versus Perl
- 2009 (127, 128, 134): On Coding Standards and Code Reviews
- 2014 (90, 97, 100): The First Ten Perl Monks
- 2010 (87, 88, 90): Nobody Expects the Agile Imposition (Part I): Meta Process
- 2011 (86, 87, 88): Common Software Development Mistakes
- 2012 (77, 77, 78): The History of Acme::Bleach and Acme::EyeDrops
- 2015 (55, 58, 61): The Boy Scout Rule
- 2003 (56, 56, 56): Turning a script into a module
- 2017 ( 54, 54): High Performance Game of Life
- 2013 (50, 53, 53): Re: Old Monks go gentle into that good night.
- 2018 ( 46): Re: Splitting program into modules
- 2016 ( 45, 45): Re^2: Do subroutine variables get destroyed? (Deterministic Destructor References)
- 2019 ( 45): Re: Criteria for when to use a cpan module (Buy vs Build)
- 2021 ( 40): Re: Perl Contempt in My Workplace
- 2020 ( 31): Re^3: looping efficiency (Benchmark Example)
Though I have 26 centurion nodes, I've been
unable to break the 100 barrier since 19-Feb-2009 (though The First Ten Perl Monks has a chance to get there eventually ... update: and did get there on Nov 20 2019).
I speculate without proof that we lost quite a few users
(e.g. brian d foy) after the "hacked catastrophe of 2009".
See also: Perl Monks in numbers?
| [reply] [d/l] [select] |
|
| [reply] |
|
The Best Nodes over the past year have XP in the 50s to as high as 90. In the old days, breaking 100 was not unusual.
Indeed.
I've kept historical data on best nodes
since 2007. As you can see, prior to 2009, even the twentieth best node of the year was 100+ while today the best node never reaches 100. The last centurion was in 2011.
That said, I don't see any dramatic decline in best node
scores over the past three years. And I find it
encouraging that the recent Re: why does location of function matter?
quickly shot up to 61 in only a week or two.
In the good old days, if I had a node in the top nodes of the year, it used to pick up several points per month for pretty much the whole year. Nowadays, it might pick up a point or two for the few months, then tends to stagnate for the rest of the year; for example,
The First Ten Perl Monks is now stalled on 90 points
and will probably stay there indefinitely
(unless the act of mentioning it here "un-stalls" it :).
| [reply] |
|
|