11
$\begingroup$

I think the site is doing very well thanks to our moderators and people who answer difficult questions. I see more and more famous names on the site, and according to Lance it has already become addictive (I personally have started showing symptoms of addiction, as I am learning a lot by reading questions and answers posted on the site).

However there is an issue that I want to bring up. I just wish there was a way to give experts a higher starting reputation as I feel some of them get frustrated when they start using the site as they cannot even leave a comment. A person who has used other SE sites starts with 101, an expert in theoretical computer science who has a long list of papers in FOCS and STOC starts with 1. This does not seem right to me. I don't know if it is against SE rules or possible at all, but I think it would be nice to manually give these people at least the same reputation that people who have used SE sites beforehand start with. Of course they earn reputation quickly, but the initial frustration can be a bar for active participation of more experts.

| |
$\endgroup$
  • 2
    $\begingroup$ As far as I understand, the restrictions for new users of SE sites (like inability to leave comments and vote) is a protection against spammer, which I don't think applies here. A simpler solution could have been the following: a new user who registers using his/her university email address and website starts with reputation 101. But I doubt that people in SE will implement something like this in near future. $\endgroup$ – Kaveh Sep 28 '10 at 11:45
  • $\begingroup$ I think would introduce an unnecessary overhead. $\endgroup$ – Dave Clarke Sep 28 '10 at 11:50
  • 1
    $\begingroup$ I have just noticed that we can put bounty even on questions with accepted answers and then award it to whomever we select (though need to wait 24 hours before awarding it.) So this can be a solution without any change to the current SE system. But this can create other problems, like other people trying to answer the question while I am not really asking for another answer for the question. Any opinions? $\endgroup$ – Kaveh Sep 28 '10 at 12:59
  • 6
    $\begingroup$ I am very curious whether this frustration really exists. I think that there is things more important in our professional life (specially of a purportedly expert in TCS) than our reputaion in a Q&A website. In fact I don't really understand why there is such mecanism... I mean if we need that to have the site working, then there is probably no real TCS community. $\endgroup$ – Sylvain Peyronnet Sep 28 '10 at 21:12
  • 2
    $\begingroup$ The frustration is not about low SE reputation, its about limitations imposed by system on new users. I have seen a few complain on the site about the fact that they cannot leave a comment, and a few other people complained offline. I really don't think this is a big issue and after posting one answer they get enough reputation to comment, but still not being able to leave comment when one starts to use the site can be frustrating, and we may have more active participation from experts if they could leave comments when they first start using the site. $\endgroup$ – Kaveh Sep 28 '10 at 22:28
  • 3
    $\begingroup$ Blocking comments from low-reputation users may be effective in reducing spam, but I feel it also creates an initial barrier for people to engage with the site. When I first started out with MO I would greatly have preferred to post some small comments and requests for clarification, than wade in with a question or answer. Comments can be voted up but not down; they can be deleted easily; they don't linger around even after being deleted; someone new to the site is likely to be cautious in their first few contributions. So this seems to be a baby+bathwater situation. $\endgroup$ – András Salamon Sep 29 '10 at 13:42
17
$\begingroup$

My feeling is that the problem is not with initial low rep, but with people not voting enough. I'll confess that if I recognize someone's name AND they have a very low rep (because they just joined), I might be more generous in voting up reasonable answers/questions. But I compulsively vote up things anyway !

I've issued close to 500 votes at this point, and I should probably be voting more. And I say this without any intent to point figures (really!), but there are two users on this discussion thread with reputation scores close to 1000 who have made below 100 votes (actually it's 94, and 43).

So I'll assert two things:

  • reputation reflects your participation in the site as well as the answers you give. You can easily earn rep by being more active, and that's how it should be. My abnormally large rep has nothing to do with my knowledge level - it's to do with my activity level.
  • We should all be voting much more frequently, especially in the early stages of this beta.
| |
$\endgroup$
  • 8
    $\begingroup$ Another way to say what Suresh is saying: the reputation economy is based on votes. If people don't vote enough, reputation will be scarce and many of the features of the site will be ineffective. We will never get out of beta if a handful of people are the only ones keeping the reputation economy going. $\endgroup$ – András Salamon Sep 28 '10 at 16:25
  • $\begingroup$ Agreed. We should motivate people to vote more somehow -- there's already a badge that you can get for voting 300 times. On an unrelated note, I was surprised to note that I have issued more votes than Suresh. Yay me. $\endgroup$ – Robin Kothari Sep 28 '10 at 22:21
  • 1
    $\begingroup$ Dang ! I need to vote some more !!! $\endgroup$ – Suresh Venkat Sep 28 '10 at 22:58
  • $\begingroup$ I wonder what a similar calculation would like like for the top page of another SE site of similar age/size. $\endgroup$ – Suresh Venkat Sep 29 '10 at 18:44
  • $\begingroup$ A quick snapshot of the 35 top reputation users shows a total of 35663 reputation points received, 5184 votes made. The ratio of votes/reputation ranges from 0 to 0.61; in total for this group this is 0.15. Ignoring the small number of negative votes, one would probably want to see a votes-to-reputation ratio of at least 0.05 to 0.1 (the former if someone votes for questions only, the latter if they vote for answers only). 18/35 are on 0.1 or above, and a further 7 are on at least 0.05. $\endgroup$ – András Salamon Sep 29 '10 at 18:46
  • $\begingroup$ I agree that a score of around 0.1 is good to aspire to, since it means you're giving back (approximately) as much as you receive. But really it should be more: you should be giving back than you receive so that the system continues to grow $\endgroup$ – Suresh Venkat Sep 30 '10 at 7:04
  • $\begingroup$ @suresh I guess that is me you are referring to. Actually I had initially though I may have been up-voting to indiscriminately, but I guess that was way off. I'll go get voting some more. $\endgroup$ – Joe Fitzsimons Oct 1 '10 at 12:56
  • $\begingroup$ @Joe, no problem. I really don't mean to point fingers, but just highlight a general cultural problem. I'm sure that as everyone gets more familiar with the site, there will be more voting. But Andras made my point more succinctly and less personally :) $\endgroup$ – Suresh Venkat Oct 1 '10 at 17:09
7
$\begingroup$

I oppose to your proposal. Paraphrasing an answer to a related question on Meta Stack Overflow:

Your cstheory.stackexchange.com reputation is a measurement of how much the system trusts you. It is not, nor was it ever intended as, a measurement of how good a theoretical computer scientist you are.

If you are expected to be already familiar with the system, you get automatic 100 reputation points. If not, you do not.

Added: Honestly speaking, if some people were frustrated enough by inability to comment or other small things for the first week to quit the website, I guess that they would be frustrated for many other things and would not stay at the website anyway, and giving them 100,000 reputation points would solve little.

| |
$\endgroup$
  • $\begingroup$ I think this is exactly the key point. The system works so that those who are familiar with the system can use more advanced features, while those who are beginners will get a simpler and more restrictive user interface. $\endgroup$ – Jukka Suomela Sep 28 '10 at 17:01
5
$\begingroup$

It would be great if users with high enough (quite high, I'd say) rep could send a limited number of invitations to people they consider to be experts, so that those people can start as users with 50 (commenting) or 100 rep (vote down), in any case above the newbie-frustration threshold. I do, on the other hand, see the point of forcing brand-spanking-new users to try a little to get used to the SE mechanisms, no matter how famous they are.

| |
$\endgroup$
4
$\begingroup$

Perhaps there is a feature request which could fix this problem. If we had an option to simply give away part of our rep to new users (just as with bounties), then this would be less of an issue. I would happily give away some rep if it would help attract more experts to the site.

Actually the feature would be win-win for us and for the SE folks, since it would draw more experts, while still avoiding spam.

| |
$\endgroup$
  • 1
    $\begingroup$ I think your and RJK's suggestions are much better than mine. They would avoid the problems mentioned by supercooldave while removing the initial frustration. $\endgroup$ – Kaveh Sep 28 '10 at 12:30
  • 1
    $\begingroup$ This could be the basis of a working solution, with appropriate limitations and possible punishments for abuse of the system. For example, if I gave Alan Turning some start-up rep and Alan turned out to be a spammer or not the Alan I thought he was, then my reputation should suffer.... or maybe this is too complex and paranoid. $\endgroup$ – Dave Clarke Sep 28 '10 at 12:32
  • $\begingroup$ @supercooldave: we would be giving away part of our reputations as with bounties, so I think people will be careful in using the feature. $\endgroup$ – Kaveh Sep 28 '10 at 12:36
  • $\begingroup$ There might be people who would use this feature to game the system and increase their own reputation, but I think this is already possible with current system although more difficult. $\endgroup$ – Kaveh Sep 28 '10 at 12:42
  • $\begingroup$ The same feature request for a different purpose is on Meta Stack Overflow: meta.stackexchange.com/questions/23841/… $\endgroup$ – Tsuyoshi Ito Sep 28 '10 at 12:43
  • 2
    $\begingroup$ I think restricting transfers of reputation to users with very high reputation and only to transfer an amount that won't make the reputation of receiver go above 101 would avoid many of problems mentioned in the question Tsuyoshi has linked to. $\endgroup$ – Kaveh Sep 28 '10 at 12:50
  • 2
    $\begingroup$ Isn't this much ado about nothing? The hypothetical high-profile researcher will almost certainly gain enough rep with their first answer that they can comment and post links and the other basics: Given their expertise, their first answer is likely to be a good one, and the celebrity effect will cause an additional up-vote boost (cf. Bill Thurston's meteoric rise on MO). $\endgroup$ – Per Vognsen Sep 28 '10 at 14:04
  • $\begingroup$ That seems to be what has happened with some of the users here, but I can understand why they may want to be able to post a comment before posting a full answer. $\endgroup$ – Joe Fitzsimons Sep 28 '10 at 14:18
3
$\begingroup$

Several people (myself included) posted our first comment as an answer because of the reputation limitation. I wouldn't solve this problem by giving people more reputation to start with, but rather by allowing anyone with non-negative reputation to post comments. If spammers become a problem we can deploy anti-spam measures later.

| |
$\endgroup$
  • $\begingroup$ We are deploying an anti-spam measure, which is exactly the rep hurdle for commenting! Also, we cannot change the rep hurdle, because it is decided by the Stack Exchange server. If you think you can convince the people who run these websites, you can try writing on Meta Stack Overflow. $\endgroup$ – Tsuyoshi Ito Sep 29 '10 at 22:57
2
$\begingroup$

I recall reading answers from Alan Turing. I'm sure he has published in neither STOC nor FOCS, but should nonetheless be given a boost in reputation. But what if it is not the real Alan Turing? Any why should someone's publication record determine their reputation, anyway? And why stop at STOC and FOCS? What about POPL? Or TCS? And who decides anyway?

The reputation systems avoids having to centralise reputation assignment, which avoids the need for subjective judgement. The reputation system is fair, based purely on participation. Finally, the reputation system avoids problems with someone posing as a famous researcher.

| |
$\endgroup$
  • 1
    $\begingroup$ My concrete suggestion is the following: a university professor with valid university email address and website starts with reputation 101. There is no need to judge reputation or anything like that. All I am saying is there are other ways of checking that a user is not a spammer which can be used to avoid this initial frustration. $\endgroup$ – Kaveh Sep 28 '10 at 11:50
  • $\begingroup$ How would you differentiate between the valid address of a university professor and a student at the same university? Different universities have different schemes for assigning emails. And you'd somehow need to have a list of valid universities and what it means to be a valid emailing address for that university. This approach would not be trivial to realise. $\endgroup$ – Dave Clarke Sep 28 '10 at 11:52
  • $\begingroup$ That is the reason I have also included the website. :) Checking that an email is a university email is not difficult because of the .edu postfix. (The same applies to the national educational urls like .ac.uk) $\endgroup$ – Kaveh Sep 28 '10 at 11:55
  • $\begingroup$ Having a website hardly proves anything. And my email address ends with neither .edu nor .ac.uk. It doesn't have .ac or .edu in it actually. But I do have a website!! $\endgroup$ – Dave Clarke Sep 28 '10 at 11:58
  • $\begingroup$ Would you write on your website that you are a professor (assuming that you are not)? Sorry supercooldave, but I think it is starting to look like a game. If you want me to admit that no system is perfect, I agree. Even the current system allows spammer asking questions, right? I don't think we would get considerably more spam if we give these people a higher initial reputation. On the other hand we may get a large number of active expert users. $\endgroup$ – Kaveh Sep 28 '10 at 12:04
  • $\begingroup$ Writing on my website that I am a professor doesn't prove that I'm a professor .... anyway, I'll stop playing. I'm merely highlighting the weakness of your proposal. A spammer is unable to post (m)any links in a question or answer, so the spammer has to earn reputation before being able to effectively spam. And ultimately it is the links the spammer posts (possibly connecting to malware) that we need to ultimately protect against. $\endgroup$ – Dave Clarke Sep 28 '10 at 12:05
  • $\begingroup$ Thank you for pointing out (some of) the weaknesses of my proposal. $\endgroup$ – Kaveh Sep 28 '10 at 12:28
1
$\begingroup$

A person who has used other SE sites starts with 101, an expert in theoretical computer science who has a long list of papers in FOCS and STOC starts with 1. This does not seem right to me.

No, this sounds very right. If you're truely an expert your reputation can quickly shoot up to 100 or more by just answering a few questions.

| |
$\endgroup$
  • $\begingroup$ I have already said that they earn reputation quickly. The problem is the initial frustration when they start using the site. $\endgroup$ – Kaveh Sep 28 '10 at 11:42
1
$\begingroup$

As it is, the system feels inconsistent:

1) If reputation is there for protecting the system from newcomers, why can they give answers, but no comments? You can spam with answers, too.

2) If reputation really is only a measure of behaviour -- not knowledge -- why do upvotes on CW posts not give reputation?

I think a decision has to be made: should reputation be a measure of ability in the subject or ability in using stackexchange? The current implementation seems to try both.

As I see it, upvotes/reputation are/is mainly given for subject-specific well done answers and comments. If we can agree on that, there is no sense in using reputation as anti-spam measure. Implement Captchas for people below the current rep-boundaries, but allow everything. Editing and the like can remain coupled to reputation as it requires some subject-related trust.

PS: As a user of cstheory.stackexchange, I feel that those responsible for this subsite should address the meta-admins based on our critique, not us directly.

| |
$\endgroup$
  • 1
    $\begingroup$ Reputation measures quality of answers and activity on the site. CW posts are more like popularity contests and they would highly skew the way the site operates. Too many big-list posts would dominate the list of questions were CW to garner reputation. $\endgroup$ – Dave Clarke Oct 5 '10 at 14:05

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .