|
I too have had the unpleasant experience of not finding my own stuff on CP.
Hans Dietrich wrote: If google can find it on the site, then so should the CP search engine.
I would like to agree, however we can't expect CP to spend the resources Google does in developing its algorithms.
OTOH I'm inclined (or sufficiently naive) to expect it could become much better with a limited effort, and that in turn would improve the "user experience" and somewhat reduce the number of junk questions.
If I can find the time, I'll do some research and make a suggestion to Chris.
|
|
|
|
|
Luc Pattyn wrote: If I can find the time, I'll do some research and make a suggestion to Chris.
I had to come up with a search algo for my site (link below) and discovered that it was trivial to let google do the work for me - a site search without the ugly google logo. As you say, google's the best.
|
|
|
|
|
Hans,
1.
if you have a solution that uses Google and works in real-time, i.e. it can find the data right after it has been added to your site, then I'm interested in learning about it for my personal use; and I would applaud it be available on CP of course.
2.
For CP I would hope to first find what is present, and then get even more functionality than Google can offer, e.g. I would like to first search, then interactively sort the results (when they are not too many), maybe by creation date, by last update date, by average score. IMO DataGridView functionality (such as in CP Vanity) is great; I implemented that on my own site (which uses full search, not word indexing, without problem, made possible by its very limited database size).
|
|
|
|
|
This probably wouldn't work for you. I'm not interested in finding something seconds after it has been added. I DO want to be able to find stuff I know is there and has been there for a while, which is my benchmark for an effective search.
|
|
|
|
|
OK, I have the normal Google search box on my site, it only finds what has been indexed before; I also have a pretty good idea how and when they crawl the web, the net result is it only finds what has been there for say 2 weeks.
That has its merits, however if I need to suddenly reply to a CP forum message I know I read only a few days ago, it wouldn't help me. OTOH, CP and Google having a special relationship may solve that. Most forum messages are reported by Google right away, all too often the question I read on CP and try to answer is the first thing that shows up in Google!
|
|
|
|
|
Chris Maunder wrote: what do you consider an "effective" search
I was looking for my article on DLLs, specifically Part 3, which I knew was called "XDLLPt3". The CP article search found no matches. google found this: Step by Step: Calling C++ DLLs from VC++ and VB - Part 3[^]
So to answer your question: an effective search is one that finds something I know is there.
|
|
|
|
|
Wow.
We never, ever thought anyone would search on filename. I just learned a new thing.
I will add that peice of meta data to the full text index.
cheers,
Chris Maunder
The Code Project | Co-founder
Microsoft C++ MVP
|
|
|
|
|
Hans should use QA then, it offers 5 radio buttons, none of them is numbered (don't click "vote", numbers may suddenly appear).
voting means a reader expresses his appreciation; he can judge positively or negatively the things he reads (articles, answers, comments, anything a user enters), and the average score can:
1. help CP to create ordered lists,
2. enable future candidate readers to make a quick selection.
having a scale (e.g. 1-to-5) allows for nuance.
voting here is not to be confused with an election, where all participants express their preference amongst candidates.
BTW: maybe scoring is the better term.
|
|
|
|
|
Luc Pattyn wrote: voting here is not to be confused with an election
On CP, you can be sure that voting will not be considered an election. Maybe sniping. Maybe "people acting immaturely". Maybe "wreaking havoc on the community spirit". Maybe "sabotaging the site good-will". Maybe "driving away prospective new members". Maybe "destroying the interest in contributing".
But not an election. Nope, you got that one nailed, Luc.
|
|
|
|
|
Chris wants CP to be very open, he dislikes all measures than are less than friendly towards new and existing users, even when some of them misbehave; he also goes a long way chasing and correcting misbehavior once it got signaled.
Personally I would consider giving zero voting weight (i.e. no voting rights) on white members, 1 on bronze, 2 on silver, etc. That would solve almost all voting abuse, but it would also muzzle well-intended newcomers. It must be very hard to do good by everyone.
BTW: the election was mentioned only because there is an occasion that only allows for positive votes; but then all candidates are being evaluated by all participants, whereas you and I read a different set of articles or messages on CP.
|
|
|
|
|
Luc Pattyn wrote: That would solve almost all voting abuse
Wow, what on earth are you smoking?
A simple "up-vote" meets all your requirements, with none of the negatives:
1. An up-vote allows a reader to express his appreciation; he can judge positively (an up-vote)or negatively (no vote) the things he reads (articles, answers, comments, etc.)
2. The accumulated score can 1) help CP to create ordered lists; and 2) enable readers to make a quick selection.
3. The accumulated scores of several articles/tips/etc. can be compared, allowing for nuance.
See? Once you forget about shooting fish in a barrel 1-to-5 votes, life becomes good.
|
|
|
|
|
Hans Dietrich wrote: A simple "up-vote" meets all your requirements,
Not at all. Author A writes an excellent article on a specialized topic, is so honest not to up-vote himself, and then gets two readers, one of them up-votes, as a result he is at 1 positive vote.
Another author B spends much less time writing a crappy article, votes himself up, creates four more accounts and votes himself up somewhat more, then all MVPs (or all 7 million members, it does not matter) read and reject that article, it ends up having 5 positive votes.
The voting system you suggest is worthless.
We need (and have) a scoring system, similar to an exam. If its crap, you get 0 out of 10 (or 1 in 1-to-5); if it is excellent, you get 10 out of 10 (or 5 in 1-to-5). Whatever the number of people who have voted so far, the overall opinion gets reflected by averaging individual scores; furthermore a weighting factor is quite acceptable to take authority into account.
|
|
|
|
|
The one thing you said I agree with is about weighted scoring.
All the rest is nonsense.
The current voting system is the #1 problem here. It's not going to get better with bandaids. In fact, over time, its problems have gotten a great deal worse, because CP gets a lot more traffic now. Ditch the idea that people here will act responsibly and maturely. They vote likes and dislikes. They vote iFad dujour. The more traffic and new members CP gets, the worse this will become.
We will never get back to the reality you're fantasizing about. The best thing we can do is minimize the havoc and bad feelings generated by immature "gimme" new members. We need to ensure they cannot disrupt the community we have.
Let's get rid of negative voting now.
|
|
|
|
|
Sorry, I disagree and have nothing to add.
|
|
|
|
|
Can I offer a suggestion?
Hans likes a solution that takes into account the reality of the average voter. Luc thinks about how he votes and needs a system that allows him to express how he feels.
I agree, and want, a system that does both.
How about if the voting system showed an 'thumbs up' vote, and a 'flag as inappropriate' vote. However, a third option is shown ("more vote options") that, when checked, changes the voting system into a 1-5 system for those who feel they need it.
2 problems with this that I can see
1. It's not as streamlined as it could be. Providing the extra option muddies the water
2. By making the defaul "thumbs up" I'm biasing the lowest common denominator
Even so, it's a solution. Any others?
cheers,
Chris Maunder
The Code Project | Co-founder
Microsoft C++ MVP
|
|
|
|
|
Chris Maunder wrote: Can I offer a suggestion?
That's OK for this once. 
Executive summary: disable all comment-less article voting; remove question voting; do an answer experiment (which I expect will prove the score becomes fake); the ultimate approach (I hope): use rep and score to filter votes.
Chris Maunder wrote: Luc thinks about how he votes and needs a system that allows him to express how he feels.
What I want most is a system that tells me how other people feel about something I am considering to read (and more so for longer items) if it is an article, or to believe is the right answer when in a forum. If the voting system does not allow me to express my appreciation, so be it; but it not telling me the community's appreciation of an article or answer is what bothers me most.
I'm not sure you're targeting articles or forum messages here. Anyway, your suggestion (which you have made before in suggs&bugs) sounds like a typical Belgian compromise (that is not always a compliment, believe you me) of two views, resulting in something nobody will like much. If I understand it right, the thumbs up would be an effective 5 (OK), the "flag as inappropriate" would be a 1 (or a "report to staff" without vote??), and "more" would lead to the original 1-to-5 voting scheme.
So that:
1. makes it slightly harder to vote anything below 5, making it a tiny bit less likely someone will down-vote for the wrong reason; Hans' excellent article could still be victimized;
2. will devalue a 5, since the more it works in avoiding a 1, the more it will cause people casting a quick 5 where a lower score would be the real appreciation.
So I'm afraid it does not solve anything; you'll get fewer false ones and more false fives (everyone rejoices?), hence less of a real score, harming the overall reputation of your article collection (CP loosing in the long run).
For articles (all kinds of them), I recommend you remove the voting bar, and move it to the "edit new message" page, so everyone who wants to vote has to explain and gets identified. For messages, I suggest you take an experiment with a twist: for a single forum and a limited period, take your proposal but make the thumbs up a 4, not a 5 (give it a special icon, I don't know, make sure to make it explicit), and see what happens.
More fundamentally, I still do not understand Hans' trouble with voting; if he writes a sh*tty article, he deserves a 1 (not that I expect him to do so). It he gets an unrighteous one, we need means to fix that, simply making it somewhat harder will not solve this; if he aggravates someone (not that I expect him to do so, but you never know), that someone will find him. Then the community, the site master, or the system should solve it.
Please remember this thread started by Hans seeing a question being down-voted. Well, the easy way out is disable voting on questions. I for one am not interested in votes on questions. I don't like your "bad question" widget (nor your "good question" widget), there are hardly no bad questions; there are questions showing ignorance, stupidity, laziness, etc. but they still probably are good questions. The only bad question would be the one that, after several iterations of "please explain, show code, what do you mean, etc" are still a complete mystery; and then its a bad enquirer, someone I would like to get rid of.
I am very much interested on the real score on answers. While I hardly ever ask a question, I'm here to learn; I learn by reading a question, trying to figure a good answer, then reading the answers. If I don't know the answer (I consider that an excellent question ), I'm glad to see to what extent others appreciated the answers presented. So if my hypothetical answer comes close to an answer with a low score, I know I was probably completely wrong; and if close to one with a high score, I feel OK.
Alternatives:
- implement a delay in vote processing, so unrighteous voters get less satisfaction;
- implement a delay in vote processing and some filtering (remove 5% of votes on both ends of the scale);
- disable votes for newbies, say white authorities (assuming bad behavior belongs to members that don't publish themselves); this punishes the good for potential wrong doings by the bad!
- and my favorite proposal (made before, maybe explained better here): disable votes that differ much from the current average (which starts at a fictitious 3 or 4) for voters that differ much in rep color from the author. So a bronze 1 gets rejected for a gold message/article averaging 3 or more. Which basically means: you need a relative level of authority for your disagreement being noted. This seems fair; established members can express themselves freely, newbies are limited in their disagreement. And I am willing to make an exception for an enquirer on a forum; if he feels an answer is not satisfactory, he should be allowed to say so no matter what.
|
|
|
|
|
Chris,
If you and Luc and I could meet for a glass of beer, I have the feeling we could come to an agreement in 30 minutes tops.
You, Luc, and others have proposed solutions to voting "problems". In my opinion, none of them solve anything. If you disagree with that statement, please hear me out; some solutions have been simplistic, some elegant, some baroque, some bandaids ("3 is the new 1"). But they all miss the mark. Let me state what I think is the problem, and then you can tell me how your solution (or anyone else's) will solve it.
Here's the problem: The 1-to-5 voting system has caused numerous people to post in this forum, asking "What's wrong with my article? What did I do wrong?" after receiving low votes for an apparently good article. [I am talking here about articles, because the problem is more relevant for them; it also applies on a lesser scale to posts.] The voting system has also caused several authors to quit CP (I know of two; there are probably more who just refuse to come back). In summary, the 1-to-5 system causes consternation, anger, bitterness, resentment, mistrust, and feelings that lead to "I'll show you; I'll start low-voting some of your articles"; and it also causes potential new authors to decide not to write that first article.
In short, the 1-to-5 system is eating away at the community spirit here.
I understand Luc's position, but I think what it gives us is simply not enough compared to what it robs us of.
You know what my solution is. It retains the ability to compare articles based on a numerical "goodness" score; it has none of the negatives I mention above.
So that's it. Please tell me how your solution addresses the problems I described. I know that Luc cannot do this, since his solution would, IMHO, make it worse.
|
|
|
|
|
I understood the goal was to create something better than the forums, aiming to replace them. While "Quick Answers" could potentially maybe be a meaningful addition to existing forums (I don't see how it would be quicker than regular forums, but it could), it being renamed to "Questions and Answers" (still not consistently everywhere!) is indicative of higher ambitions, its format only really supporting simple questions and answers is confusing. It is incredibly hard to believe, but I'm getting convinced CP is not aware how fantastic their forums are (yes, they do lag a bit in the looks department). So I don't know what the future will bring.
Hans Dietrich wrote: the QA forum has had the beneficial effect of absorbing questions that would clog the regular forums.
If you mean it acts like a "bad questions" forum, I don't see a need; regular forums know how to deal with them. In a polite way, most of the time.
[major addition]
I just noticed the main menu, and the current bug list, now refer to it as "Quick Answers"; has it changed, did I always read it wrong, what happened? and why are people calling it Q&A then, when it is QA?
I wish CP could make things a bit clear; I'm very active here, I read most of several forums, and I don't know; how is a newbie to know where to ask his question? why is the "how to ask a question" sticky not fully explicit on the matter?
OTOH if Q&A/QA is intended for simple questions getting a quick and good answer, then why spending all that effort in its voting, commenting, editing, removability, etc. To be quick, it should be simple, which it no longer is. I don't know what goals it is after, and I don't know what goal it reached.
[/major addition]
|
|
|
|
|
I've given this at LEAST 2 seconds thought, so here's what I've come up with for your sig:
QA repulses me.
I'm matter; QA is anti-matter.
QA: puke, puke, puke!
Gimme a Q: Questionable! Gimme an A: Awful!
QA? Not today.
QA? or Q&A? Make up your mind
|
|
|
|
|
Thanks for the suggestions, Hans; I think I'll make up my sig myself.
|
|
|
|
|
Luc Pattyn wrote: I'm getting convinced CP is not aware how fantastic their forums are
Thanks for that vote of confidence and kind words. I think our forums are the best out there, hands down. Slashdot is the only other system I look up to, but even that isn't quite what I'd like.
However, there were many, many threads about all the problems with the forums. The cross posting, the difficulties in coming to a definitive answer, the lack of tagging plus a bunch of other stuff. All we were hearing at the time was that we needed something better.
It doesn't actually matter which is better. What matters is to have a good, usable system. If that, eventually, means two separate systems to cater for the different tastes, then I can live with that. If, however, it means we take all we've learned from Quick Answers and plow it back into the forums, then fine. If it means providing Quick Answers with a UI that brings the speed and information density of the forums, then that's fine too.
The issue is that I get a lot of very strong opinions from very intelligent members that make it clear that people use fora in very, very different ways. For now I'm focussing on finishing up Quick Answers (2 final features + UI streamline) and then I'm heading back to the forums to see if we can make them better, so helpful suggestions on making the systems better is what I need.
cheers,
Chris Maunder
The Code Project | Co-founder
Microsoft C++ MVP
|
|
|
|
|
Chris, thank you for this clarification. All this is good news. I am not sure you wrote them before or after I made my big edit there, but it is good news either way.
Long ago I understood Q&A/QA was intended to replace the forums. A couple of issues, such as cross-posting, wrong-forum situations, and the slightly dated look of things would be reasons to do so. IMO having QA and forums side by side would not solve any of these. And I'm not sure having just QA would solve many of them either.
I have been a member of a beta testers group and an insiders group which came to be around the time Q&A/QA became relevant; I soon got frustrated by getting lots of e-mails, sometimes a dozen a day, and often on details where the overall picture wasn'[t clear yet, i.e. while not getting much information at all as to goals, status, plans, etc. So I opted out.
I am willing to participate once more, but I do hope the circumstances to do so would be better this second time around. I do have a couple of ideas and suggestions, as you might have guessed... I would prefer a systematic approach tough, which would start with goals. Having clear goals (written, discussed, agreed) makes everything that follows a lot easier.
|
|
|
|
|
It should be a comment not an answer. I posted about this the other day.
Maybe we need a 'Convert answer to comment' button with a selection tool to select the answer(s) the comment should apply to?
DaveIf this helped, please vote & accept answer!
Binging is like googling, it just feels dirtier. (Pete O'Hanlon)
BTW, in software, hope and pray is not a viable strategy. (Luc Pattyn)
|
|
|
|
|
I've often wished for a "spike" command like in the James Bond movie.
|
|
|
|
|
DaveyM69 wrote: Maybe we need a 'Convert answer to comment' button with a selection tool to select the answer(s) the comment should apply to?
That's actually on our todo.
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
