I added a feature-request earlier that revolves around my use of tagging to draw .NET developers to questions that are language-agnostic (as far as C# and VB.NET are concerned).

The request is based on my assumptions that there are people qualified to answer the question who have [.NET], [C#], or [VB.NET] in their "Interesting" tags list, but not necessarily a particular combination of the three.

I'd like to see what user's have in their Interesting and Ignored tags lists as part of the data-dump, even if it's been anonymized so that the actual lists for a user can't be tied back to the user themselves.

  • 3
    Please don't abuse the tag system. – Ian Elliott Aug 18 '09 at 19:00
  • 4
    I'm referencing the fact that previously he said that he places completely irrelevant tags on certain questions to draw more attention to them. By what he's asking, it seems like he's trying to find out what tags people have marked as interesting, simply so he can abuse this further. – Ian Elliott Aug 18 '09 at 19:18
  • If you tag questions properly, the right people will come. Stop trying to game the system. Nothing I hate more than people appending a tag I follow because they use it in another part of their project. If you really think that I won't retag a question as soon as I see that it has useless/misleading tags, anyway, you're out of your bloody mind. – Eric Aug 18 '09 at 19:24
  • 1
    @Ian: I'm requesting the data because I'm interested in creating the venn diagrams to see if SO users follow their language or their framework. I'm not trying to abuse the system, just understand which of multiple tags (all of which are accurate for the question) is more effective. – STW Aug 18 '09 at 19:29
  • 1
    @Yoooder: And how do you not see that is an admission of everything Ian accused you of?! – Eric Aug 18 '09 at 19:29
  • 1
    @Eric Because it would be USING ACCURATE TAGS! It's not trying to abuse or game the system provided that the tags used are accurate and relevant to the question! – STW Aug 18 '09 at 19:31
  • @Yoooder: The correct tag is .net, not c# if it is a .NET framework question. – Eric Aug 18 '09 at 19:32
  • @Eric: it would actually be the opposite of abusing the system; if all VB.NET and all C# followers also follow the .NET tag, then tagging C# or VB.NET would be pointless and redundant for .NET questions. – STW Aug 18 '09 at 19:32
  • @Eric: I think you're missing the point of the .NET analogy. The vast majority of C# questions in SO are focused on working with .NET by means of C#, very few are actually C#-specific questions--even if they represent an example of the situation in C# they can get perfectly valid and accurate answers in VB.NET or any other .NET language – STW Aug 18 '09 at 21:13
  • 1
    It's meta, the downvotes usually mean they just disagree with you. – jjnguy Aug 18 '09 at 21:22
  • 1
    For a partial list of why you may have been downvoted: meta.stackexchange.com/questions/15251/… – jjnguy Aug 18 '09 at 21:23
  • 4
    My down-vote says "This is a really, really bad idea." The dumps are about content you can remix and re-use with proper attribution. My preferences are none of your business, even if 'anonymous' ... – Tim Post Jun 22 '10 at 0:24
  • @Tim, Why in the world would you care about people knowing your tag preferences? Who cares if you like VBA? or if you don't like waffles? – Lance Roberts Jun 22 '10 at 0:32
  • (-1) for the reason in my answer. – devinb Jun 22 '10 at 20:42

If it was already public information, I would be all for it.

Unfortunately it's not currently public information.


It's unlikely they will put anything in the Public Data Dump that is not already publicly visible.

  • 1
    That's only true to an extent. There's a bunch of information that isn't publicly visible in the dump (votes for instance) – Ian Elliott Aug 18 '09 at 20:18
  • I can appreciate needing to keep it private, even if it's anonymized it would be interesting to see – STW Aug 18 '09 at 21:09
  • @Ian how are votes not publicly visible? – devinb Jun 22 '10 at 21:42
  • @Ian, you can see down/upvotes separately by clicking on the vote count (this may require some minimum level of reputation) – bdonlan Jun 23 '10 at 0:04

I'm all for more data, so if they included this type of information in the data dumps I wouldn't have a problem with it. I'm sure some of our resident statisticians will be sure to whip up some fun graphs.

  • 4
    MORE DATA! I wanted to do some cool stuff with the data dumps, but there's not enough data...nearly everything should be public, except for the openID used to log in. – Thomas Owens Aug 18 '09 at 19:09
  • More data is definitely good, although at some point I won't be able to index all of it (my database can't be nearly as large as SO's), so there will come a point where I have to start dropping data regardless. – Ian Elliott Aug 18 '09 at 19:20
  • Ian: That's true. But depending on what you want to do with the data, you need different portions of it. Therefore, everything should be included and then the consumer of the data can decide what to use and what to ignore. – Thomas Owens Aug 18 '09 at 19:27
  • @Thomas, I was never suggesting it to be any other way, just mentioning the fact that there are limitations (although more so for me since I'm actually building a framework off the dumps - I would LOVE more data, but am disappointed that I'll have to pick and choose what I need at some point :(.) – Ian Elliott Aug 18 '09 at 19:35
  • Ah. Yes...tradeoffs. I wish I didn't have to deal with any myself. – Thomas Owens Aug 18 '09 at 19:38

The general principle is that we make stuff available in the data dump that is already public on the website.

Since the list of Ignored and Interesting tags is not public, I don't see why it would be included..

  • so simple solution, make the data public – Lance Roberts Jun 22 '10 at 6:12
  • 9
    The problem is that making data public that was not originally public has a way of undermining users' confidence and trust in the system. This is why so many people have a less than stellar opinion of Facebook; they took data that users thought was private, and made it public without the users' intervention. – pkaeding Jun 22 '10 at 19:58

I am kind of ambivalent about this change:

  • On one hand more data = more awesome, we could mine some extra info out of it.
  • On the other hand this data is not public

You can already get a much more interesting breakdown of the tags a user participates in using the data dump, many of us do not bother carefully crafting an interesting / ignored list.

I am less worried about abuse with the data, much richer information is available in other places.


I mentioned part of this in my answer here

When given new information, people will act on it. It cannot not have an effect on the way they behave. Thus, since we know that it will have an effect, we must consider whether the effect will be good or bad.

Firstly, lets consider the positive effects of having the aggregate interesting and ignored tags in the data dump.

  1. You can tell which tags are "interested" the most. We already know which tags are the most popular from the "Tags" page, so this doesn't generate truly useful information.
  2. You can tell which tags are ignored the most and not ask those questions. This doesn't really help if you have a question about those issues.

Lets consider how a user would incorporate this new information into asking their question. The user already has a question, that's why they came to StackOverflow in the first place.

Three options

  1. Question is "interesting" The user tags the question correctly and moves on, their net gain is zero. The data simply confirmed that their question would have a good chance of being answered, which they could have known by looks at the [tags] page.
  2. Question is in the middle The user now has the opportunity to use their new information. They incorporate some random tidbit from a subject many people have as "interesting", and then add that tag to the question. This will generate more views even though the question is not related to that tag at all. This actually makes the "interesting" tags less useful, because the fact that people are maliciously gaming the system means that the users can't trust the tags anymore.
  3. Question is "ignored" The user doesn't want their question ignored, so they do not tag it correctly. Even if that tag is literally what their question is about. Something like [NeuralNetworks] or [r-language] or any other niche tag. If users stop tagging their questions correctly because those tags are ignored, this is a disaster for searching, both from Google/Bing and from within StackOverflow itself. It also degrades the integrity of tags as a whole.

Ultimately, this information will cause no gain and will almost certainly cause some degree of harm (probably minimal, I'll admit), so there is no reason to include it.


To those who assume that exposing this information would be subject to abuse: yes, it could possibly be abused--but I think that the benefits would offset the risk by a fair deal. For one thing, the community would continue to retag questions that were wrongly tagged just to draw attention.

Also, the currently provided data can be easily used to see the list of tags applied to a user's questions. I can easily look at user X, query their posts, and aggregate the tags of their posts into a rough profile of what they are working on, it's likely to be a subset of what they are interested in.

BUT one C# developer might realize that C# is just the language he uses to work with the framework, so he might tag all his questions with just [.NET], another might tag their BCL questions with [C#] and [.NET], and another might confuse the language for the framework and tag all their questions as just [C#]. However, the same group of user's may all have both tags in their Interesting Tags list.

So far the only good reason not to include this data is Brad's answer. The opponents of the idea seem to be paranoid that the data is ripe for abuse--because what I really want to know is whether I can get more spam-hits by tagging questions with Viagra or Enzyte

  • I stopped reading at potentially. – Tim Post Jun 22 '10 at 0:34
  • edited, for your conscience :) – STW Jul 23 '10 at 16:49

You must log in to answer this question.

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