8

Examples:

I just realized that these question, although they could of course be applied to Open Data, have no clear focus on any "open" aspect. If they referenced a concrete open dataset, it might be a different story. As it stands, I guess that these question would be better suited for Stack Overflow or Database Administrators (in case of database-specific questions).

What do you think?

1
  • 1
    Isn't the mere fact of data being made public in non-useful forms reflect on the open nature of not-quite-open-in-reality data? Already it feels like this issue must surely have come up more often in this forum than at those developer/dba sites.
    – Roger_S
    May 16, 2013 at 20:34

5 Answers 5

6

No developer hacks with open data any different to non-openly licensed data. So if you follow this argument through, you should avoid discussing any developer techniques.

But that seems stupid. There is a reasonably well-defined activity of data hacking/wrangling, which is having a big surge of interest because of all the open data being released right now. Having a forum to discuss how to best to approach badly formatted CSV files, merge multiple datasets together, how to convert to linked data, techniques for mash-ups on maps, ideas to produce cool visualisations etc. is valuable. Creating this new site and community around this activity has value that you would have say with a tag on StackOverflow, even though there is a bit of overlap. I'd rather discuss how to deal with that overlap, and that might mean carefully writing the scope.

The other sort of questions seen on the site - asking about finding or dealing with particular (open) datasets - is valuable too, but I think it would be a far less interesting site in general with just them.

5
  • 2
    These are challenges which, if included in the scope, I would agree add value to the site. I think the limitation should be that if the inquirer is discussing certain challenges, then reference the exact data (web location or otherwise) for which those challenges are arising from. That way the community can more effectively assist and the question becomes more grounded in the context of this site.
    – Taliesin
    May 9, 2013 at 15:28
  • So if someone asks a question about a specific dataset is it on topic? If rcackerman had asked "How do I exatract data from xyz.gov's address dataset?" instead of "Good tools to parse repetitive unstructured data" would it have been on topic?
    – John
    May 11, 2013 at 1:55
  • 2
    Yes, it would be on topic. But I think the scope of the site should be wider, to include more general questions about tools.
    – D Read
    May 11, 2013 at 9:00
  • 2
    I think requiring that the question be about a particular dataset makes the entire exercise less valuable. Part of the purpose of SE sites is to provide a Google footprint for particular problems and their solution, so that people that have the same problem in the future can take inspiration from those solutions. If the problem is with a dataset, by all means, make the question about the dataset, but if the problem is general, why artificially constrain the question? Other people might have the same problem, but with a different dataset. May 14, 2013 at 19:28
  • Not sure what you think about it 10 years later, in particular when the website has now very low traffic, but doesn't what you describe also overlap with softwarerecs.stackexchange.com? The two questions mentioned in the original post would fit perfectly there.
    – J-J-J
    Mar 13 at 15:39
4

I said something similar in the other meta thread about the PDF extraction question, but I think the value of an Open Data community is at least in part that people can ask questions of others who are likely to have faced similar classes of problems. A lot of us have direct experience with getting data out of PDFs, since this is unfortunately a common way governments release data, and this pool of people therefore has a concentrated set of people who've dealt with this problem that Stack Overflow may not have.

About the PDF question in particular, I'm also not sure it's fair (again, as I said in the other thread) to categorize it as a "programming" question. The solution may well have ended up not requiring code.

1
  • Someone asking a question about extracting PDF should mention what kind of tool they're looking for (some library in a specific programming language? which language? a tool that does not require code? etc.). Otherwise, the risk is to end up with a long, unusable list. Incidentally, this kind of question is already in the scope of softwarerecs.stackexchange.com.
    – J-J-J
    Mar 13 at 15:46
2

I think you have just about answered your own question. If the programming question does in fact target a specific open data set then it would be suited for this site.

For the examples you reference here, Stack Overflow would appear to be the more appropriate forum for those questions. However, for questions targeting sophisticated database handling Database Administrators would be a better option.

At the end of the day, questions on this site should somehow be related to the topic of Open Data, but if we are to move unrelated questions let's help the inquirer by pushing the question to the site with users best suited to answer the question at hand.

EDIT: While you are targeting programming questions here, the same question could be asked regarding generic database questions.

1
  • Thanks, I added the database aspect to my question. May 9, 2013 at 14:08
2

@Taliesin answers pretty well but misses one thing. What exactly are open data specific questions? My understanding is that open data should be machine-readable.

Therefore, as I stated in a similar question targetting specifically data extraction, any question related to formatting data to be easily readable and therefore Open, is related to open data.

To summarize, questions about data extraction techniques are on-topic because it targets opening up data.

3
  • 1
    Machine readability is one important aspect. However, take for example the IMDb Top 250: It is very machine readable, and still it is nowhere close to open. Why? Because the data is not published under an open license. To make the concept of Open Data clearer, I strongly recommend Tim Berners-Lee's 5 Star Scheme where he explains the different aspects of high-quality Open Data. May 9, 2013 at 14:55
  • Yeah, I don't discuss all aspects of open data. I just mean some programming questions are related to open data because the programming will result in opening up data. But indeed, there is another deal, the license.
    – Vince
    May 9, 2013 at 15:05
  • 1
    I don't follow the argument. Opening data means making it public. But data extraction may be aimed at private or confidential use of the data, so it does not necessarily imply opening data.
    – J-J-J
    Mar 13 at 15:41
0

10 years late to the party, but I think the two questions you mention suffer from two things:

  1. They lack focus

The people who asked these questions do not mention their level of technical knowledge and their requirements (e.g. which programming languages or OS they are using, if they have material or computational constraints, etc.). So the risk for them is to get suggestions of tools they won't be able to use.

Now, you could argue that the answers could be useful to other people than the original poster. The problem (besides ignoring the needs of the person who originally asked the question, which is a bit cavalier) is that it would be useful only to people willing and having the time to explore many answers with possibly a lot of irrelevant or outdated suggestions for their specific situation. I think that either these people are a tiny minority or they don't exist at all, in particular when they have the alternative option to ask a question tailored to their specific situation.

  1. They are already in the scope of other websites

If we solved the problem above and made these questions more focused, then they would be in the scope of https://stackoverflow.com or https://softwarerecs.stackexchange.com.

Consequently, I don't think that questions about tools with no direct relations to open datasets are in the scope of opendata.SE.

You must log in to answer this question.

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