What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 131 Stack Exchange communities.

I've found that the Data Explorer doesn't have access to Interesting and Ignored Tags. I found this feature request to add them to the Data Dump, but the reason given for not doing so is that they are not public information.

Since I can't think of any good reasons to keep these private,

lets publicize a user's Interesting and Ignored Tags.

EDIT: Well it doesn't look like this idea will get traction, but I hope people can see through the arguments against. Look at what's already public about an individual:

  • Questions
  • Answers
  • #votes
  • Tags (w/counts)
  • Badges
  • comments
  • edits made
  • reputation in total and per post
  • favorites
  • associated accounts

A whole gamut that could be used to simulate the tags of interest to the user, it just takes more complicated coding. By not making Interesting Tags public, we just giving the illusion of privacy, kind of like TSA gives the illusion of security at airports. It's privacy by complexity.

share|improve this question
    
Given how well that feature request was received, I'm not sure why you've posted this... –  Jon Seigel Jun 22 '10 at 18:51
    
@Jon, Desperate men do desperate things. –  Lance Roberts Jun 22 '10 at 18:53
5  
You can't think of any good reasons to keep them private... but can you think of any good reasons to publicize them? Why is this change useful? –  Grace Note Jun 22 '10 at 18:53
    
@Jon, I'm actually hoping someone can show me some Logic as to why these need to be private. It just escapes me. –  Lance Roberts Jun 22 '10 at 18:54
    
@ccornet, so that we can do some interesting Data Explorer searches, Data Mining is always a great reason. –  Lance Roberts Jun 22 '10 at 18:55
    
I can ......... –  Tim Post Jun 22 '10 at 19:29
2  
(-1) for the reasons in my answer. I'd give you a second (-1) for "Data Mining is always a great reason" because of the heinous enormity of that claim. –  devinb Jun 22 '10 at 20:23
8  
"Since I can't think of any good reasons to keep these private" - ah, yes, the Facebook motto. –  Shog9 Jun 22 '10 at 20:25
1  
@Shog9, we're not talking birthdates here, just whether you like VBA or hate waffles. –  Lance Roberts Jun 22 '10 at 20:39
2  
@Lance: again, the Facebook attitude: "Eh, most of this stuff is beyond trivial - who cares who sees their mini-game stats or cat pictures? If they wanted privacy, they wouldn't be posting it at all!" But of course, people do trivial things for non-trivial reasons, and who are you to take information that was never intended to be shared and open it up just to sate your idle curiosity? –  Shog9 Jun 22 '10 at 21:01
1  
So why not share it on an opt-in basis @Shog9? I wouldn't mind sharing them if it would help them recommend interesting questions/answers –  Ivo Flipse Jun 23 '10 at 15:20
1  
Regarding the edit... I stand a good chance of obtaining your phone number given your name and location, but I cannot obtain your interesting/ignored preferences given your questions and answers. I might assume that the tags you're active in correspond to those in your prefs, but there's a pretty good chance I'll be wrong... –  Shog9 Jun 23 '10 at 15:21
1  
@Ivo: Help who? The site itself already has full access to this information - and since the whole point of the preference is to adjust how the site looks for me, there's no need for an "opt-in". But I neither need nor want recommendations from arbitrary users who happen to wander by and view my bio page, so sharing this information with them seems pointless... –  Shog9 Jun 23 '10 at 15:24
1  
@Shog9: Since @Jeff isn't going to implement a StumbleUpon functionality, I would like some Stack App to do this for me. Just because YOU don't want to share this info, doesn't mean everyone doesn't want to. –  Ivo Flipse Jun 23 '10 at 18:25
2  
@Ivo: I fail to see the connection. Both favorites and aggregate voting data are already publicly available if you wanted to implement a recommendation engine based on user-ratings... And if you're going to implement an opt-in, 3rd-party service anyway, I don't really see why additional hooks into the user-preference system would be necessary. FWIW: I'm arguing on principle here - I personally do not use the interesting/ignored tags preferences at all; I just think this suggestion is a waste of time and sets a bad precedent. –  Shog9 Jun 23 '10 at 19:20

5 Answers 5

Lets talk about something called an expectation of privacy for a moment.

If I have data on SO that is not visible to others, I have a very reasonable expectation that that data will not be made public in database dumps. I've already entered that information with an expectation of privacy. That's not something you can just 'reverse'.

If its not already public, there is very little chance of it being included in a database dump. My questions and answers are CC-BY-SA, as the text is useful and I love to see it copied.

My interesting and ignored tags just feed companies that serve advertisements and figured out how the e-mail hash works .. and a few naturally curious people. The only people I want using that data are SO employees. I trust them, hence I provide information regarding the things that I like.

When it comes to DB dumps, what you see on the site is what you get. If that changed, I'd seriously consider not using the site anymore, so would a lot of people, hence its not very likely to happen.

Edit

As Kop suggests, I'd have no problem with aggregate stats. Just not on a user by user basis. I'm not disputing the usefulness of your proposal, it would be really cool to more accurately check what languages are rising in popularity with each dump. I'd just hate to see the dumps 'abused', beyond the current problem of people not following the license.

share|improve this answer
    
I have a problem with both individual and aggregate stats, as mentioned in my answer. –  devinb Jun 22 '10 at 20:23

Aside from Tim's excellent post on the expectation of privacy, there are a host of other reasons.

This that are fundamentally public are information is anything which you are giving to the site. This is the information in your profile, and any of your posts/comments/answers.

Things which are fundamentally private are information relating to preferences, or how you use the site. That is, it is really no one else' business which questions I choose to view. It is not their business whether I get notifications or not. It is not their business how often I visit the site. And lastly, it is not their business which tags I choose to mark as "interesting".

The internet is currently all aflutter with privacy kerfuffles. Gmail's taking a huge hit for what they are doing with individual user search data (Oh, it's also private what my "search terms" are on SO), Facebook is taking huge flak for what they do with individual user data. It has become very clear that people want to be able to "publicize" information rather than be forced to "depublicize" it.

Which leads to the conclusion of this, which is: Why do you want it public? All the possible reasons fall into two categories. You want to know it or you want to act on it.

  1. Satisfy your curiosity This is a terrible reason to make anything public, because many (if not most) people have curiosity about things that they have no business knowing. Satisfying curiosity is simply not a valid reason for making anything public.

  2. Change your behaviour This would or could have a terrible outcome. I've mentioned it before (especially relating to these issues) StackOverflow is about asking questions. The last thing we want is people to "game" the system. This includes things like "Explain the fraud-detection algorithm", "How can I tell who's online", and your request "Which tags are most likely to be ignored/interested". The reason this is terrible? It would be a very STRONG incentive for people to start mis-tagging their posts because they know that more people are interested in [SharePoint] than [CommunityServer]. So you tag it falsly, and thereby generate more interest. The other half of this is people purposefully leaving off relevant tags because they are on many people's "ignore" list. For instance, there may be many people ignoring [r-language], but that doesn't mean that people shouldn't be tagging their questions with it.

Now, if you're saying, "Obviously, people wouldn't do that, it would be bad if they changed their tagging practices just because of this information" then you fall back to reason (1): That you only want to know because you're curious. Which, as I mentioned, is invalid.

EDIT: I also elaborated here.

share|improve this answer
1  
While I agree completely with your points, I did find it humorous that you picked [sharepoint] for your interesting tag example. –  Grace Note Jun 22 '10 at 20:35
    
@devinb, nicely written answer, though I can't possibly agree with it. I'm not talking really important stuff here like birthdates. If someone cares about someone else knowing what tags they like, then they probably need to relax a little. If you look at the other Data Explorer searches, you'll see plenty of them that only exist to satisfy curiosity. But you're also missing the bigger picture, once the data is available to data mine, then trends can be determined that can make the site better. Having a few hundred thousand eyes instead of a dozen or so, will help us to find... –  Lance Roberts Jun 22 '10 at 20:45
    
improvements to the system. I always believe in making things better, and data is one of the keys to figure out how to do that. –  Lance Roberts Jun 22 '10 at 20:46
    
How will this make it better, though, @Lance? What trends from this will actually improve the system? What will this data provide that is tangible to the system? Should we be changing the focus of Stack Overflow because the current community is more interested in certain tags? Yes, trends can be used to help refine a system, but not all of them will have an effect. I honestly don't see how this avenue is one of the ones that helps. –  Grace Note Jun 22 '10 at 20:54
1  
@ccomet I've worked with both SharePoint and Community Server in the past. It seemed a good example of things which you could "claim" are related. And I find everything about SharePoint humourous. –  devinb Jun 22 '10 at 21:33
    
@Lance, you're bizarre correlating "more data" with "better everything in the world". –  devinb Jun 22 '10 at 21:34
    
Well played, devinb. Well played. –  Grace Note Jun 22 '10 at 21:34
    
That is, you've indicated that with more data things will just get better through "data mining" which makes me believe that you really have no idea what data mining means. Your theory would suggest that if you released the home address of every site user, the site would just get better. According to you "once the data is available to data mine, then trends can be determined that can make the site better". Why hide anything at all? Would the world therefore be perfect without any personal information? –  devinb Jun 22 '10 at 21:36
    
@devinb, things need to be hid to protect people. I could care less who knows my birthdate, but I want it protected because of identity theft. I list my address and phone number in the phone book so people can get ahold of me and mail me, but I don't plan on publicizing it in my profile. Most things have to be dealt with on a case by case basis. –  Lance Roberts Jun 22 '10 at 21:54
    
My point is that you have repeatedly asserted that "more data" is just better. But just now you admitted that it's still case by case. I've enumerated in my response, and in my response to the other feature request, exactly why this individual case is just simply a bad idea. It will not help, and it could quite possibly harm the usability of the search. –  devinb Jun 22 '10 at 21:55
1  
@Lance Even if devinb's examples are only "potentials" for harm, you haven't really provided any potentials for help. Things don't just "help", they help by having an effect. Having "more data" alone isn't going to help, what helps is having data that helps. As it stands, all that's been shown in this back-and-forth is that there is potential for harm. Give some concrete potential on what usable data you can identify in trends from user tag preferences. Give some examples of what possible trends could exist that would improve Stack Overflow. –  Grace Note Jun 23 '10 at 15:20
1  
@Lance: do what analysis? Are you gonna try to keep tags around that aren't/shouldn't be used otherwise but which are inexplicably well-represented in user preferences? Seems like a prime example of how too much irrelevant data can lead to bad decisions... –  Shog9 Jun 23 '10 at 15:56
1  
@Shog Exactly! I've been trying to find the words, and you're the one who found them. When people are given more information, they always try to incorporate it into their decision making, regardless of its relevance. They assume that more information is good and it ends up clouding the issue. –  devinb Jun 23 '10 at 16:07
1  
@Lance Less extraneous and irrelevant data is better. –  Grace Note Jun 23 '10 at 17:21
3  
@Lance: either you have a specific example in mind of analysis that this data would make possible, and we can debate that on its merit... Or you're back to claiming "more data == more better" and I'm back to asserting that idle curiosity doesn't justify publicizing formerly-private information. –  Shog9 Jun 23 '10 at 19:23

As Tim Post pointed out there is a serious privacy concern, but I think there is no harm in releasing aggregate stats on the interesting/ignored tags :O

share|improve this answer

It's not privacy by complexity. All you can do with the public data is guess what someone's interesting and ignored tags are. Because these preferences are entire independent of the person's voting/tagging/posting content. Yes, the majority of users tend to have interesting tags for the places they participate in the most, and ignored tags for parts where they might avoid... but they aren't tied together. The illusion here isn't of privacy - the illusion is that you will derive someone's preferences because of their activity. The reality is that our private settings are in fact private, and you can only surmise what we have with the current public data.

For example, I once set my interesting tags on Meta to be [status*] so that those qustions are highlighted for me. You cannot derive that from the data dump. Nor is there any reason for that to be public knowledge without my consent (see Tim's post about expectation of privacy). My preferences on the other sites are my own business.

share|improve this answer
2  
Other example, between my real interesting tags on SU, you can also find: android, iphone, blackberry, games, gmail. Why? because these tags are often used for off-topic questions, and I want to see them directly, to vote to close them. –  Gnoupi Jun 23 '10 at 15:35

Perhaps instead, would it be possible to access or import one's own interesting/ignored tags from associated accounts?

share|improve this answer
    
There's a feature request for that somewhere... –  Jon Seigel Jun 22 '10 at 20:20
    
Not quite what you want (sync is not nearly as good as import/access), but there's meta.stackexchange.com/questions/48631/… I didn't find any others. –  Grace Note Jun 22 '10 at 21:36

You must log in to answer this question.

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