I wish this were the whole reason. It may be why people are bored with the stories on the frontpage; there doesn't seem much if any change in those. But comments do actually seem worse. Though I still have hopes of reversing that trend.
As far as the stories are concerned, I don't think they ever all resonated with me. To be honest, I probably got more out of Reddit in early 2006 than I ever did out of this forum. Part of that maybe been because of one time benefits of becoming an online junkie-- finding your essays, seeing some of Steve Pavlina's older pieces, learning about a wide variety of niche interests, etc...
Since getting to HN, I'd say that about 1/3 to 1/4 of the front page has been consistently interesting. Too much of the rest has been self-promotional blogging, valley gossip and current fads (diet, exercise, productivity, etc).
I still wouldn't say that I see the site in a bad light. Both the stories and the discussions are a valuable resource. I'm just diverting a bit more focus towards stack exchange sites and screencasts.
Depressingly, HN would rather front page an article about personal growth than admit that maybe, just maybe, all of this meta bitching is setting the tone. That's from the top down and I partially, and with respect, blame pg for carrying the flag of comments suck -- instead of taking an opportunity to turn common sentiment around.
Common sentiment is now comments are declining in quality. Wherever that came from, that's now parroted fairly often and is undoubtedly planted in the back of everybody's head as they go about their business. That cannibalizes some comments, because people now think "am I contributing to a decline?". That makes people more inclined to snark, to discourage others from adding to the decline. This is the exact sort of thing I watch happen all day long in the comments now.
Take one recent example. Someone quoted a highlight from a linked article because it was humorous, and you could have missed it[1]. They presented no original thought on it themselves, they simply quoted it verbatim. Some asshole came along and berated him for it, as if he'd committed some kind of grievous sin. The fact is, I enjoyed the quote in question, and I upvoted him because I'm glad somebody else noticed. Worse comments have happened, but now there's such a fixation with "quality" that people are starting to up the snark. Which adds to the very problem we're trying to fix. Look at that entire thread.
The rub is that rather than focus on opportunities for improving going forward, this community has an inane fascination with returning to the way things were. Your thread was too late, I'm afraid, as the negativity has accelerated any decline that exists.
You're dead-on; HN is reverting to the mean. This navelgazing lament about the times that were is setting the pace for the future evolution of HN, though. The high-karma users have spent so much time wondering how to get back to the past -- a now completely unattainable goal due to growth, I'm sorry -- that there has been immeasurable lost opportunity on focusing on where to go from here.
I'm not sure that technological improvements are going to fix it, either. Even if the negative trend is as pronounced as believed (and I don't think it is), acknowledging it and spending countless comments focusing on it has cost precious time in reversing it.
Rather than embracing the inevitable evolution, understanding it, and finding ways to make the evolution more positive, we've established blood-curdling fear at the mere prospect of evolution. Now, that glacial evolution is sliding along on a trail of fear instead of a trail of positive outlook.
I may well the guy you're talking about who was berated and then downvoted -- even though two others besides you had posted that they liked my comment. And when I replied to the post that berated me, I was once again downvoted. At least one comment has been deleted but it's still a pretty interesting snapshot of what the mood is on HN these days http://news.ycombinator.com/item?id=2437225
I am speaking from the perspective of improving HN, so I'm startled (should I be?) to find that the first reply is an admitting that the community will eventually split apart.
In the end you stay with all the complainers, but none of the non-complainers. This is a real risk for a community. (In the case of the game it was 'simple' to fix this: change the game. It is far harder for something like HN)
I think jed is right; the navel-gazing is part of the problem and the more meta-discussion occurs the lower overall quality gets. This belief is informed by my experiences on a variety of fora, and is not HN-specific. Tackling the problem then, should occur in a way that doesn't draw attention to the fact that there is one, but fixes it silently in the background. This is partly why I like hidden karma scores; they (in theory) reduce groupthink and handwringing about karma.
My suggestion is this: pick a few dozen people who you think are good contributors, and write a quick script to dump their voting patterns. Take a look at how they're voting, and choose some number whose votes you think best represent what you'd like HN to be. Then make those people a class of superusers, but do so invisibly. Neither they nor the rest of HN should know they've been tapped.
I envision the superusers as being vote accelerators. Their votes would not only confer the typical +/- 1 karma, but would tag the comment they voted on. An up-tagged comment would gain, say, 2 karma for every upvote, and a down-tagged one would lose 3 for every downvote. Those values are provisional and arbitrary, and should be changed. Tags could stack or cancel each other out, but that's something else that should be experimented with.
If the superuser scheme seems to be having a positive effect, consider making it viral. That is, if some receives, say, 200 positively tagged comments, they too become superusers. Again, that constant should be chosen carefully. The idea behind making superuserdom viral is that it scales automatically (given a wisely chosen threshold), and offers insurance against decreased contribution from the original set of superusers.
Karma can be a very effective way of cultivating the kind of interactions the community desires, but can also lead to gaming and groupthink. By increasing the value of trusted users' votes, but keeping that mechanism hidden, I believe you should be able to maximize karma's benefits while minimizing its drawbacks.
Anyway, I apologize if this post is off-topic or should be addressed elsewhere. I figured a meta-thread was as good a place as any to introduce it.
I also wanted to say in postscript that although HN may be suffering the growth pangs of any forum, it still offers far better discussion than most sites I've found, so the doomsday-ery is perhaps premature.
That actually reminds me of the metacritic scheme. IIRC they take reviews and then (a) translate them into a numeric score according to some method they don't tell you about and (b) weigh each critic by a different weight, of which they don't tell you about.
So your idea is not only interesting, it is probably workable.
My general hypothesis is that you can improve comment quality by giving users more metrics than just an upvote or downvote.
For example, consider the following:
Some people are saying there's a lot more negativity around here. I think that's true. Also, I've found that if someone is negative towards me in a way that seems personal, I get defensive and have to work hard to not respond in kind (and sometimes I respond in a mean way anyways). Sometimes these kinds of arguments get a lot of upvotes, because lots of people take one side or the other, and new people get drawn in.
One way to deal with this particular problem would be to treat comments that are "mean" differently from comments that just aren't interesting or insightful or correct.
For example, you could just cut off, or gray out, any comment thread that becomes "mean," thus diverting attention from it; people already involved would be able to continue if they want, knowing nobody else will probably look at it, and new people won't get drawn in. How you detect when a thread becomes "mean" is not easy, but maybe you could have a "mean" button/flag. A person clicks the "mean" flag to vote the thread as mean, but their vote's power is weighted against their karma, and then there is a threshhold that, if reached, actually designates the thread as "mean" and greys it out or whatnot.
Now, the point of this comment is not actually the specific "meanness" example, but the hypothesis I laid out at the beginning of the comment.
I mean, hell, you could try a bunch of different metrics. Just create a little "metrics bar" that allows users to vote on various things (like "meanness," etc.) and rotate a few different options through over time just to see which ones help or give interesting data. Keep the normal voting much more prominent, though, and make this metrics bar less prominent, so that it can be ignored and doesn't impact the UX much for those not interested.
Although maybe I shouldn't say "unfortunately," because that problem may be easy to cure. E.g. it might work simply to add a karma cost of for commenting, an idea I'm seriously considering trying out.
Shiiit Paul, hacker news has been screwed up since the beginning. Remember this video?
http://www.viddler.com/explore/rentzsch/videos/31/
That video would show up on the front page about once a year, and is by one of your most popular members and commenters. The video is full of slander and straight up factual errors and even gets almost all of cryptography and network protocols wrong. It's probably one of the worst videos on security out there, yet, what did HN do?
Eat. It. Up. Nobody flagged that video, told tptacek it was bullshit, or bothered saying one damn thing about it other than cheering it on and laughing. You definitely didn't boot him off the site, and in fact gave him a top spot in your system. This is not how you tell people you don't like 'meanness'.
Over the years here and before I started commenting people called me a cocksucker, cock, asshole, douchebag, and every mean ass thing you can imagine. They do it to other people too, and I sure as hell don't see anyone getting flagged or kicked off. If you pick someone unpopular to insult you're totally allowed to be a dick here.
Paul, HN is in decline because you don't actually enforce your own rules. If you don't want people to be "mean" or talk like idiots then don't let mean and idiotic crap on the front page. Half the time, when one of my many clever ruthless jokes shows up on the front page I giggle. My jokes demonstrate what you've been allowing for years but only seem to care about now that it's insulting your own ethos of the "uber hacker scientist ruling the world":
You guys love cruelty and insult, I'm proof.
I can say with a fair amount of confidence that Zed's fully aware of this effect, and is doing what he does intentionally. He has informative content, but also knows how to push precisely the right buttons. It's often entertaining, but the abasive style is also a brilliant and effective form of self-promotion. (Think Donald Trump or Rush Limbaugh.) He knows precisely what buttons to push to get a reaction, and (as far as I can tell) enjoys himself immensely. It's great to watch if you like to see flames written to people that take themselves too seriously.
So I don't think you're telling him anything new here. If you sit down at a table and don't know who the sucker is...
I know it's off topic here, but could you either elaborate on this or point us to more detailed rebuttal?
- no personal attacks
- no "x is an asshole because" inflammatory comments
- ostensibly duplicate content should be deduped
- no flame bait?
I agree that people making personal attacks should be punished. Perhaps a reset of karma and a limited time ban? Obviously there isn't going to be a silver bullet, but I agree with Zed here about enforcement.
Easy to monitor as well.
People are very creative in being mean to each other if they really want. Also it's not specific words that hurt but the context in which they are used.
Also consider that people on this site are hackers and so naturally we can all speek Leet [2] ;-) In these days of unicode it's extremely difficult to prevent people posting words that look like swear words, but which on inspection are just unusual characters.
I shortened my comment above but my full suggestion is to apply time-delayed negative karma to posts with Carlin swears rather than ban them. Who is going to spend 45 seconds finding the 𝛍 to avoid −2 karma?
Firstly, I gave unicode as a simple example of the lengths that people could go to in order to swear. But it's actually even easier than that in practice. I you want to ban the word X then I can replace it with a slight misspelling which carries the same intent - and when you ban that a naughty user will just pick another one.
(Bad words aren't the problem we should address, it's the intention behind words which is the real problem.)
Secondly, the Scunthorpe problem is more severe than you realise and can't be solved by considering only whole words. For example, some words have more than one meaning: "tit", "gay", "sod", "knob", "cock", "prick", etc. Some or all of those words would be filtered out by a naive filter.
There's also the additional problem that people may not necessarily agree on whether a word is a swear word or not (even if it does not have multiple meanings like the examples I listed below).
For example, in 1977 the Sex Pistols won a court case about the title of their album "Never Mind the Bollocks" in which it was ruled that "bollocks" is not obscene in the UK (but may of course still be considered obscene in other countries!): http://en.wikipedia.org/wiki/Bollocks#Obscenity_court_ruling
And finally, some words mean different things in different coutries. For example, the word "fanny" means one thing in the US and quite another in the UK.
Naive filtering of swear words is never going to work.
As I outlined above it's just about the incentive and setting the tone. If you have to write "siht" you'll think twice about whether it needs to be said.
I might do well to add that I am not suggesting this for prudish reasons. I took a creative writing class in college and the teacher banned curses and drugs. Intro to Creative Writing students otherwise substitute force for substance.
> Comments are getting worse.
>> What is worse?
> Comments are overwhelmingly mediocre.
So, I ask, what is mediocre?
Comments are the thoughts of human beings using your system, and there's no technological or psychological means to have any semblance of control over them. Their perceived quality is a direct reflection of the community you have. Any attempt to "fix" perceived quality is going to divide your community, and any disruption or improvement to the community is going to indirectly adjust perceived quality.
Just like none of us writes like Thomas Jefferson these days and a number of words have fallen out of favor, so has your community evolved and a new definition of "average comment" must be found. This is all a relative game, anyway; a comment that you call mediocre I might find to be a golden nugget of awesome, and why should my opinion matter any less?
As an aside, I wish I had a dollar for every suggested technological fix to the declining-comment problem. Based on how this thread is shaping up, I'd be doing better than my 401(k).
It doesn't really add anything to the discussion. Snarky comments really don't belong here on HN and I've noticed a trend of them getting far more up votes than they used to.
Why does the mediocrity apply to the comment you're replying to, but not your own? Your comment is dripping with snark and someone called you on it (thankfully). I actually think your comment is shittier than the one you're replying to, because you put yourself on a pedestal of telling the rest of us what to do.
Was that comment at the top of the discussion before you snarked it to oblivion? Probably not. It was most likely already in the deep hell of very light gray, and your content-free comment just made it worse. Calling the kettle black, in my opinion.
I think it was fine to say the guy's comment was "dripping with snark" (a specific criticism), but not that it was "shitty," and not that he was "telling the rest of us what to do."
I would call those bits "mean" because the amount of insult they add is disproportionate to the amount of content they have. In other words, it would have been worthwile (IMHO) to find a better way to say the same thing.
Now, the whole point is that who am I to say what's mean? And I agree, hence the proposal I made to allow people to "vote" on what's mean, and have a threshhold. I mean hell, different communities find different things to be unacceptable "mean," and that's fine.
"Mean" for a girl scout meeting and "mean" for the LKML are pretty different; I think HN should probably be somewhere in between.
And the point is, if HN becomes like the LKML, we're gonna enjoy it a lot less; that's why we want to limit meanness, not because we're big softies.
Can we please stop this silly fained-hypocrisy treasure hunt? There is certainly appeal to a game everyone can play, but it's a pointless waste of time. What exactly do you hope to achieve? A world where spammers and trolls say what ever they want and well meaning people just remain silent?
Sometimes the only way to stop immoral violence is with immoral violence.
I can distinctly remember several times the comments on an article were written by absolute experts in the material, full of insight and interesting content.
I could believe comments like those have become rarer.
Reading your second sentence strictly, we can infer that to you, quality means (a) expert-level familiarity with a linked article, (b) insightful -- difficult to generalize because insight is different for different people, or (c) full of interesting content.
Of the three of those points, only one of them isn't dependent on the perspective, knowledge level, or interests of the reader. What is insightful to me might be common knowledge to you. What is insightful to you might be tediously stupid to me. What is interesting to you might be boring to me. On, and on, and on.
I'm just expressing one vote that I agree with some peoples' consensus, that a particular type of quality of HN has decreased; and it's a type of quality that I value.
It seems the chief complaint is that comments are no longer increasing the happiness and confidence of the community, but rather raising anxiety and hostility. My startup is based on an algorithm for measuring the emotional impact of text and does in fact measure these four emotions.
I'd be pleased to help measure how comments' effects on readers have changed over time. I would do it myself, but I know you don't want me anonymously scraping every comment for the last 3+ years, and it'd be much faster to just get a data dump from the server.
Drop me a line if interested. :)
If that works, then requiring a proof of work (a simple-ish logical puzzle) attached to a snarky comment could bring a scalable barrier.
http://news.ycombinator.com/item?id=2429832
Fortunately the person I replied to took the comment in the sense it was intended, but I also know that people can be really defensive in this area and can perceive a comment as snarky even though it wasn't meant to be.
A quick visual reminder that is "below the fold" - ie below the reply button, so easily ignorable for frequent users, but very visible for new commenters that may not have read the guidelines...
Could have a list of the key guidelines, or show one of the key points at random, or even set it based on what you think is one of the main issues affecting the type of comment on the site.
I consider myself a bit of a Reddit refugee, and given the kind of comments that are popping up lately, I think there are many of us here (and if you look at Reddit now, it has been similarly declining, with an influx of Digg- and 4chan-like content and comments). I think a large part of the problem is that people are flocking to HN, and somehow "colonizing" it without trying to integrate into the "local culture". Hence the mean-spirited comments, etc.
Is karma the right thing for this? Maybe there should be a new thing called a "leave a comment" ticket. Every comment will consume a ticket. Every comment that gains strictly greater than 1 karma will gain you 2 "leave a comment" tickets but you can never pool more than 4 "leave a comment" tickets. I've left pure supply and demand here though and I've capped profit, maybe its more like supply and demand with government oversight. Every 3 days (tunable, like all prior #s) if you don't have any tickets you get a ticket(Socialism?). I think people will either quit being mediocre or will stop coming to the site, for better or worse. All this is under the assumption that other hackers don't like to read mediocre comments. I think all is lost if such an assumption is false anyways. BTW I have no idea how the HN comment system works now so someone correct me if this idea is perpendicular to the current system.
Perhaps some way of visualizing the way distribution of the karma bands for the people who voted on a comment (both positive and negative) would be good - I would actually go and check this occasionally to see if I was getting too far out of step.
e.g. Something that told me that your comment may have a total score of 30 but the people that have a karma level over 10000 who voted all downvoted etc.
(This one, it's mean and full of snark)
Since then, I've periodically left HN for periods of time because of the disappointing shift in focus from science and tech to "entrepeneurship", whatever the hell that is.
Now half the posts I filter past are frankly just advertising, or "SEO" in various guises, people selling their useless web app, pushing their own affiliate links, breathless chortling over the lastest gray areas to investigate for reducing their taxes, or promoting their own blog. For a while there, we were even getting almost daily front page reposts of Wall Street Journal editorial psuedoscience.
I suspect there's a big chunk of us here who really could not possibly care less about getting rich quick.
I think there's a fine line to walk. "Hacking", if that's still what this place is about, means an intense curiosity to understand, the joy of discovery, and an equally intense desire and joy in creation. I don't see HN reflecting that as much anymore.
Nowadays, it seems like there are much more people complaining about how articles about economics or making money aren't 'hackery' enough. It's what I dislike about Reddit for the few time I read it; HN is (used to be?) full of people who either run or want to run software businesses, and want to provide actual value with software (and have it validated by getting people to pay for it) rather than doing some code experiments for the fun of it with no other users than themselves in mind. There are many 'programmer' forums, the amount of 'software business' forums is minuscule.
advertising, or "SEO" in various guises, people selling their useless web app, pushing their own affiliate links, breathless chortling over the lastest gray areas to investigate for reducing their taxes, or promoting their own blog
- SEO: yes (same)
- people selling useless web app: this is just how you look at it; in many cases, it's people who want feedback on their experimental ideas. Which I think is fine.
- ppl pushing affiliate links: no, but they didn't used to be here, either
- tax optimisation: yes
- blog promotion: as long as they're blogs on the topics of the forum, yes
Many things on the list are described in a negative way, so it's only natural that they're undesirable. But I don't see them that often (like affiliate links - they're quite rare as far as I can tell) and I used to think that Youtube was useless, too.
I've probably posted it before, but it seems like a good time to reference Richard Bartle's analysis of social dynamics in MUD communities again: http://www.mud.co.uk/richard/hcds.htm
I think this is as good an overview of online community dynamics as any other I've seen in the last ~20 years online. By accident or design (more the former than the latter, I think), this yields quite similar insights to grid-group cultural theory, as discussed in the following: http://changingminds.org/explanations/culture/grid-group_cul... http://www.soa.org/files/pdf/2009-chicago-erm-ingram-02.pdf and http://www.amazon.com/gp/product/0955768144
I have not got around to reading the book in the last link, but I like his other work that I've seen.
(Also, since I'm sure anyone commenting on the state of HN will have their account scrutinized, my first account is nearly 2 years old, and I've been looking at HN for around 3 years total.)
I have a low 5-digit UID on slashdot, and registered well after I started reading, so I was sort of in the second wave there. The discourse quality fell off and leveled out, but my experience is that low digit UIDs generally contribute at their historic high quality levels.
In my opinion, a low UID filter on slashdot would be far more effective than the mod/meta-mod/complain about mods and meta-mods response this system created.
I don't know how this might translate to HN; "Filter by age of account?" What if by default you only showed comments and posts from accounts older than two years? (This would disqualify me, by the way, so maybe one year..) :)
People could select to see comments from younger accounts if they liked, or perhaps very well seasoned accounts (two years on the site) could promote up comments to general visibility. This would be patently unfair, in that new awesome commenters would not get to contribute to the discourse immediately, but it would almost certainly immediately increase the quality of discourse.
Slashdot's system has some echos of this, but they made it far, far too easy to become a moderator as they wanted to spread the work out. They also I think created a mental expectation with the '-1' score which created a trolling subculture, that is, reading slashdot at -1 is a cesspit, and so nobody does it, hence people use it as a cesspit, hence.. What I'm suggesting is making it the strong default to do the equivalent of reading slashdot at say 4 and higher, but using age of account as the heuristic.
This would be a reasonably easy chrome extension, but obviously easier to do on your side. If you demo this yourself, I'd be curious to hear if you think the quality is improved.
In real life, we might call this an apprentice system, that is, you have to sit in the back, and only occasionally get your answers while you prove you wish to belong and that you deserve to belong. It's a known effective way of maintaining quality of members in a desirable group, from religious to musical to .. I imagine secret societies?
Re: xenophobia, it's not precisely that, as I'm not suggesting a date-joined-cutoff, just that it takes a long time for HN to be convinced you should contribute, rather than just learn.
/. 2774/4UID user here, seniority doesn't necessarily correlate with quality as high performance users are a subset of early adopters just as they are with later adopters.
Do you disagree with my experience on slashdot, out of curiosity?
What it means for an account to be GC'ed is up for experimentation, but a reasonable default might be that it doesn't get any voting or submission rights. Arguably it's a bigger problem that stupid comments get upvoted than that stupid comments exist at all.
Here's a community that has really strict moderation: http://whirlpool.net.au/wiki/wp_modfaq
The user is able to see reason for being censored and can complain, etc.
Or something like that, if you get to negative score, you can't vote anymore until you go back to positive with nice comments and good submissions.
Concurrently, you could use the votes (esp negative) of the users tracked and identified as described above as notifiers to the moderators. That way you leverage more volunteers, but with less power (and they wouldn't know they had it, either).
Combined with a 'you get two posts for each post that scores +2 or higher, or 1 post every 3 days' to limit the amount of posts and incentivize people to may high quality posts (or at least posts that they think will score well).
I'm more concerned with the a decline in the quality of good comments. It's anecdotal of course, but it seems to me like the top 10% of comments I read and discussions I participate in here has declined. This is really the part that I care about most.
I care about the median or worst comment mostly because of how they affect the best ones.
For the record (and in no particular order):
1. http://news.ycombinator.com/item?id=2440812
2. http://news.ycombinator.com/item?id=2441254
3. http://news.ycombinator.com/item?id=2440990
From that viewpoint, if the articles posted haven't changed much, not much have been lost. The perceived quality may be only representative of increase in number of users, and new users trying to earn karma or new users unfamiliar with guidelines.
So may be the high quality comments and articles are still there, but some low quality content/comment is skewing the perception.
I have seen it too many times. I was in a discussion with a prominent user here and noticed that every one of his comments would shoot up 4-5 points within seconds of being submitted while mine would go down as many points.
In a later thread, around a month ago, I intentionally requested the page as frequently as I could to measure comment scores, and found that this users comments were spiking up to the top of the thread within a minute.
Comment counts are also self-perpetuating, which you have fixed now, but it would mean that with multi-account a simple 2-3 votes would spike a comment up with further votes from other user (eg. 'if he has had 4 votes in 8 minutes this comment must be good' - groupthink)
Multi-accounting should be easy to fix with IP tracking and some smarts (ie. I assume a multi-accounter would copy+paste the link to the comment). And although the comment counts are no longer there, the ranking and valuing each users vote as being equal is still there.
I'd change the way the karma is calculated in order to reward people who are consistently good rather than heavy posters. I suggest using the geometric mean of the post scores.
As such, you wouldn't have to give a karma penalty to comment. A weak comment would reduce the karma automatically.
I recommend the geometric mean since the scores are most likely Zipf-distributed. It could be calculated in the background as a running arithmetic mean of the logarithm of the scores, which shouldn't be too expensive.
I'd display the karma not as an absolute number, but as color-coded rectangle. The color would be based on percentile of the score described above, or perhaps as a function of this percentile. The post count (or better, the post frequency) of the user could be represented as the length of the rectangle. You could then display this "badge" next to the user name in threads. Since it would be a continuous thing, it would not divide the comunity in have and have nots as the badge system introduced and retracted earlier. In order not to discourage people from posting, users below a certain threshold (say the 50th or 60th percentile) would all have the same value displayed.
I'd also grant the various rights based on the basis of the percentile rather than based on hard numbers.
Every time a new right is granted, users should get guidelines on how to use it, i.e. don't upvote based on agreement, don't downvote unless the post is offensive/shallow.
Commenting in topics with little points could perhaps weight more than commenting in threads with a big score. This would encourage posting in less popular threads.
Downvoting should cost karma. This could be implemented as a comment with a low score. You'd keep two counts: a post count and a downvote count. The karma would be calculated based on the sum of these numbers, but the post count would still be displayed. The value of a downvote would be calculated according to the number of post, otherwise it would cost less to people with lots of comments/votes. The karma of the voter may also be taken into account.
The up and down votes should be weighted as a function of the karma of the voter. They could be implemented as a floating point number.
Removing the number of votes per post prevents to detect a highly upvoted answer to a little upvoted comment. If the votes are weighted as described above, the vote count has little meaning anyway, but the post score could still be displayed in a meaningful way by using a color-coded system similar to the one described for the karma.
I'd suggest two distinct scores
log( comment score ) / log( highest score in the thread )
log( comment score ) / log( highest score ever at the time of thread submission )
Making the first value relative to the max score at the time of voting would allow that value remains relevant as time passes, even if the community grows or shrinks. It would require to add a "max score" field to submissions, though.
Absolute karma score is a measure of how many people like your comment out of how many people have read it. There's just no way to determine how many people have read it.
One solution would be to compare how much total karma the comments in the post have accumulated when you post. Then compare it to how much karma there is once reading has stopped.
The karma doesn't have to be calculated immediately, btw. It could be evaluated when the thread is closed for new comments.
Edit: and rampant, childish downmodding.