Reputation vs. Posts
-
I've found it satisfying that my reputation is higher than my post count. But then I sampled a few other posters and found this is almost universal. It seems that having reputation more than twice the post count is normal.
So, @ben_lubar, care to run some stats on reputation to post count ratios? What are the:
- Median ratio
- My percentile
- Percentiles of ratios of 1 and 2 reputation-to-post-count
- The user with the highest ratio
Bonus points for ignoring the likes thread.
-
-
It figures I have to extract data the WTF way. Infiniscroll until I get to the users with zero values, copy from the inspector, paste into a file, and parse to a sane format.
My wife mentioned something about the Internet being full when I was infiniscrolling the sort-posts page.
Values have thousands separators on the initial page load, but not after "Load More".
-
Of users with at least one post:
Median ratio 0
I am at the 88th percentile
A ratio of 1 is at the 87th percentile
A ratio of 2 is at the 92nd percentile
@meglio is the user with the highest ratio (35 on 1 post in the Likes topic)
A ratio of 0 is at the 77th percentileOf users with at least one reputation:
Median ratio 1.3
I am at the 47th percentile
A ratio of 1 is at the 45th percentile
A ratio of 2 is at the 67th percentile
@Cassidy is the user with the lowest ratio (1 on 2940 posts)The Likes topic throws a lot of noise in the high end. Unfortunately, getting the data to filter it out would seem to involve an excessive amount of screen scraping.
-
You could always make a statistics plugin for nodebb.
-
@Greybeard said in Reputation vs. Posts:
The Likes topic throws a lot of noise in the high end.
The Likes topic is the best noise.
-
We never yet considered how to implement @shadowmod on NodeBB, did we?
-
@PleegWat i have not yet considered how to do so, no.....
suggestions welcome.
-
@LB_ said in Reputation vs. Posts:
You could always make a statistics plugin for nodebb.
Would require learning Javascript. Pffft.
-
@LB_ said in Reputation vs. Posts:
You could always make a statistics plugin for nodebb.
You could always just not give a shit.
-
@PleegWat said in Reputation vs. Posts:
We never yet considered how to implement @shadowmod on NodeBB, did we?
Recall that he worked with a backed up database instance. There's no way to get one via the admin interface of NodeBB. We'd also have to reinvent the queries in MongoDB.
-
It's probably less needed now people have no reason to check things like TL3 progress, attendance, 2^n posts etc.
-
@blakeyrat said in Reputation vs. Posts:
You could always just not give a shit.
these words of wisdom could be applied to absolutely any problem
-
-
@Adynathos said in Reputation vs. Posts:
@PleegWat said in Reputation vs. Posts:
implement @shadowmod
And what does it do?
on discourse you could request various statistics about your account and the forum.
how many posts you had, how many likes, the post curve for overall posts, the post curve for the top 25% monthly posters, stuff like that.
it was statsporn.
oh and you could get a report of your progress to TL3
-
@accalia said in Reputation vs. Posts:
various statistics about your account and the forum
Nice thing to add to the forum software, but I guess if it was widely used, the queries would bring Mongo down.
-
@Adynathos easy to solve, just create an automated process to backup the database and restore to mongo on another server. that's what we did with discourse because a single sneeze could bring discourse to its knees.
-
@accalia How many sneezes does it require to take NodeBB down?
-
@RaceProUK Only one if it's @ben_lubar or a dwarf
-
@accalia said in Reputation vs. Posts:
a single sneeze could bring discourse to its knees.
It could also bring the database backup process to its knees
-
@Jaloopa said in Reputation vs. Posts:
@accalia said in Reputation vs. Posts:
a single sneeze could bring discourse to its knees.
It could also bring the database backup process to its knees
oh that broke all on its own.... i don't think we had a proper automated database backup for months before we migrated away.
-
@Jaloopa said in Reputation vs. Posts:
It could also bring the database backup process to its knees
Towards the end, that didn't even need a sneeze; lack of space on a 60G drive put a stop to that.
-
@PJH said in Reputation vs. Posts:
@Jaloopa said in Reputation vs. Posts:
It could also bring the database backup process to its knees
Towards the end, that didn't even need a sneeze; lack of space on a 60G drive put a stop to that.
only because for whatever reason discourse required about 40GB of free space to successfully do a database backup.
-
@accalia said in Reputation vs. Posts:
@PJH said in Reputation vs. Posts:
@Jaloopa said in Reputation vs. Posts:
It could also bring the database backup process to its knees
Towards the end, that didn't even need a sneeze; lack of space on a 60G drive put a stop to that.
only because for whatever reason discourse required about 40GB of free space to successfully do a database backup.
root@hpdesktop:/home/sockbot# du restores/ -h 3.4G restores/2015-11-17 4.0G restores/2016-03-11 2.9G restores/2015-08-20 3.4G restores/2015-11-24 3.9G restores/2016-02-13 3.5G restores/2015-12-08 3.4G restores/2015-11-13 4.0G restores/2016-02-24 3.5G restores/2015-12-01 32G restores/ root@hpdesktop:/home/sockbot#
And that's just the SQL-alone backup. i.e. not including images.
-
@PJH said in Reputation vs. Posts:
And that's just the SQL-alone backup.
yes, that's the final backup size. for some reason the process used a lot more scratch space on disc than that (and yeah i'm sure 40GB is a slight exaggeration even so)
-
@accalia said in Reputation vs. Posts:
for some reason the process used a lot more scratch space on disc
First guess would be that they were generating the dump uncompressed and then compressing the file afterwards as a separate step instead of streaming into the compressor. Probably because it was simpler for someone to understand when they wrote the code, and they likely then leveraged the brain-damaged processing to do some sort of twiddly stuff on the intermediate files, making switching to doing it right annoying.
TCDCK: painting themselves into a corner with stupidity and bikeshedding.
-
@dkf said in Reputation vs. Posts:
generating the dump uncompressed
CORRECT!
@dkf said in Reputation vs. Posts:
compressing the file afterwards as a separate step instead of streaming into the compressor
ALSO CORRECT (and very stupid. use the pipes mario, use the pipes </ObiWan>)
@dkf said in Reputation vs. Posts:
then leveraged the brain-damaged processing to do some sort of twiddly stuff on the intermediate files
CORRECT!
and you know the best part? if ANY part of the process failed, the intermediate files weren't cleaned, and were just left sitting there on disc.
-
@accalia So easy to figure out what happened. :) I thought carefully about and assumed the most brain-damaged option in every case.
-
So what I'm reading is:
- Any status should include
WHERE thread_id <> likes_thread_id
- MongoDB is so fucking shitty that we can't do queries on it
MSSQL could run a query like this in no time, with the right indexes. Fuck, I'm sure MySQL could do it, too.
What the serious fuck is up with using database structures that can't be queried?
- Any status should include
-
@ben_lubar said in Reputation vs. Posts:
https://what.thedailywtf.com/users/sort-reputation
https://what.thedailywtf.com/users/sort-posts1305:911 rep to posts ratio. Wow, I am lower than I thought.
-
@Lorne-Kates said in Reputation vs. Posts:
MongoDB is so fucking shitty that we can't do queries on it
oh we can query it. we just can't query live without risking performance impacts to the forum itself.
we had the same issue in discourse.
-
@Lorne-Kates said in Reputation vs. Posts:
MongoDB is so fucking shitty that we can't do queries on it
They don't use it right. They're using MongoDB as a key/value store.
-
@accalia And we'd potentially have the same issue no matter the DBMS
-
@RaceProUK said in Reputation vs. Posts:
@accalia And we'd potentially have the same issue no matter the DBMS
absolutely.
that's why the only way i'm writing another shadowmod is if it has offline db access refreshed regularly. that way you can DOS shadowmod (by accident or malice) but you can't DOS the forum through that vector.
-
@Greybeard reputations are kinda skewed by the fact that a whole slew of upvotes got transferred over from , which had upvotes ("likes"), but not downvotes. It'll take a while for things to even out now.
-
@anotherusername said in Reputation vs. Posts:
which had upvotes
they had downvotes too. at least on meta.d they did.
they called them "banhammers"
-
@accalia well, true, but one of those gave you a rep of minus infinity or thereabouts. And they didn't trust average schmoes to wield them, just incompetent rageaholics like Jeff.
-
@blakeyrat said in Reputation vs. Posts:
You could always just not give a shit.
Listen, nobody here makes fun of your weird-ass hobbies like replacing functional water heaters, so why not do everyone else the same courtesy?
-
@anotherusername said in Reputation vs. Posts:
@accalia well, true, but one of those gave you a rep of minus infinity or thereabouts.
well yes. there was a bit of a balance issue, but the admins said it was only to foster a safe and accepting environment where all were welcome with.... -snerk-
BWA HA HA HA H HA HA HA HA!@
sorry, i couldn't keep a straight face!
-
@FrostCat Next week I'm gonna lasso-down a flag pole.
-
What the hey?
-
@FrostCat
08
is not a valid Octal number.
-
@Lorne-Kates said in Reputation vs. Posts:
What the serious fuck is up with using database structures that can't be queried?
Mongo isn't an RDBMS--there's no indexes (as far as I can tell from what people here have said.)
-
@blakeyrat said in Reputation vs. Posts:
They don't use it right. They're using MongoDB as a key/value store.
Yeah, the software is designed to be able to use mongo or redis interchangeably. For some reason.
-
@RaceProUK said in Reputation vs. Posts:
And we'd potentially have the same issue no matter the DBMS
LOL.
-
-
@blakeyrat said in Reputation vs. Posts:
Next week I'm gonna lasso-down a flag pole.
Pics or it didn't happen.
-
@FrostCat said in Reputation vs. Posts:
Mongo isn't an RDBMS
So it's a piece of shit that shouldn't be used to store relational data. Gotcha.
Why are we using it?
-
@boomzilla said in Reputation vs. Posts:
Yeah, the software is designed to be able to use mongo or redis interchangeably. For some reason.
I can't wait until some data object hit's Mongo's document size limit and the whole thing collapses like a house of cards.
-
@FrostCat said in Reputation vs. Posts:
Mongo isn't an RDBMS--there's no indexes (as far as I can tell from what people here have said.)
There are indexes. There are no fixed tables / columns over which to define and enforce relations.