12

There's a guy on a mission to change all links to Java documentation from old version to Java 6 or 7 (the old documentation is not online any more). For example that one: https://stackoverflow.com/review/suggested-edits/3427037

Now, that may be correct but in some cases the new version of classes or methods may work differently so changing of the link will make change to the original meaning of the post. The frequency of edits suggest that they are not evaluated carefully.

Also, there's just so many of them coming - I'm not sure if it's a concern on its own. I guess not if all of them were correct.

I just wanted to bring it to attention here.

  • 3
    Er, wait. There is no publicly available 1.4.2 documentation anymore. If you want the documentation, you have to download it. So, I'm not actually sure how to handle this anymore. Normally, I'd go with the above, but that's not possible in this case. – Aza Nov 24 '13 at 4:38
  • I know, still, that massive amount of changes doesn't feel right... – Szymon Nov 24 '13 at 4:39
  • I'm more worried about the documentation links rather than the edits. If the edits fix things which need fixing, they're generally okay. – Aza Nov 24 '13 at 4:41
  • 5
    Is there any that can be verified as incorrect? Most documentation I've ever read also makes mention of the changes that occur between versions, so that only one version of documentation needs to exist. I don't know if Java does that or not, but things to look into. – animuson Nov 24 '13 at 4:42
  • 1
    An example of what @animuson is getting at can be seen in BitSet - the class has been around "Since JDK1.0" while certain methods have been introduced in 1.7 (valueOf). Given that Java is painstakingly backwards comparable (I am sure there are exceptions), this shouldn't be too much of a concern. – user213963 Nov 24 '13 at 5:03
  • 2
    There should be some easy way to deal with this, but I don't know what it is. – Blue Ice Nov 24 '13 at 5:39
  • "in some cases the new version of classes or methods may work differently" -- FWIW, for Java it is officially not so, search web for something like "Java backwards compatibility" – gnat Nov 24 '13 at 14:04
8

Fixing broken links is ALWAYS a good edit.

If you find a specific case where such an edit inadvertently points to documentation that isn't relevant to the post, then fix it and leave a comment notifying the editor of this.

Otherwise, don't worry about it. He's doing well.

-2

Yes, these are good changes in themselves, but, since the user:

  • Has less 2k reputation (which requires the edit to be reviewed), and/or
  • Are editing old posts (since an edit will move the applicable question to the front page),

These edits should be rejected as Too minor if the user is fixing only that and there are other issues with the post (and, ideally, edited to fix all the issues - "Improve" -> untick "Suggested edit was helpful"). If there's nothing else wrong with the post, approving it is fine.

A common issue I see with most / all of those posts (while possibly a bit opinion-based) is that they just have an ugly raw link: (possibly a result of how the user is identifying these posts)

http://docs.oracle.com/javase/7/docs/api/java/io/ObjectOutputStream.html

Rather than something prettier:

See the ObjectOutputStream documentation for more information.

  • I'd like to hear the downvote explanation before I upvote – John Dvorak Nov 24 '13 at 13:28
  • 10
    Disagree. Fixing broken link is always a non-minor change in my opinion. Might to radical change if the new link is not proper, but not minor. – Shadow Nov 24 '13 at 13:29
  • 2
    @ShaWizDowArd but other issues should be fixed as well. Isn't it so? – John Dvorak Nov 24 '13 at 13:29
  • 1
    @Jan true, but personally I reject only when editor suggested really minor things like removing thanks or "i --> I" without fixing other things. When editor fix whole block of code formatting but forgets "i" to "I", will you also reject just because of that? – Shadow Nov 24 '13 at 13:30
  • @ShaWizDowArd Understood. I'll try that strategy (accept unless it's blatant repfarming), too – John Dvorak Nov 24 '13 at 13:32
  • 1
    @ShaWizDowArd If the post isn't dependent on the link (which it shouldn't be), the link in itself is minor and, by implication, fixing it is also. (But, either way, I guess "too minor" is often opinion-based) – Dukeling Nov 24 '13 at 13:34
  • 1
    @Dukeling in case of jsfiddle or other source code hosts, you're right. But we can't expect users to always quote whole sections from official documentations, plus it might not be legal without asking the company first. Anyway, guess we can't have any real "rules" here, each and his own opinion. :-) – Shadow Nov 24 '13 at 13:37
  • @ShaWizDowArd, for suggested edits I disagree about "Fixing broken link is always a non-minor change". There's at least 3 people who need to review such suggestion, while the link might never even be clicked. And if clicked, any sane programmer should be able to find the correct link. – Arjan Nov 24 '13 at 14:51
  • @Arjan the team is working on a "broken links" review queue that all its purpose would be... fixing broken links. So this kind of proves broken links are big issue. Until such review is available, fixing those broken windows via suggested edits does more good than harm in my opinion. reference 1, reference 2 regarding future existance of broken links review. – Shadow Nov 24 '13 at 15:03
  • Sure broken links are a big issue, @ShaWizDowArd. I just don't feel suggested edits are the good way to handle broken links to common documentation. The automatic comments of the broken link detection would alert the author of the post, which is better than yet another review queue, I guess. (But then, I don't review, so I don't really care.) Just like your links, I'm unsure if there's any progress there: Does Stack Exchange crawl websites? – Arjan Nov 24 '13 at 15:15
  • Oh, @ShaWizDowArd, interesting comment here :-) – Arjan Nov 24 '13 at 15:18
  • @Arjan in the old times when the suggested edits queue was overflowing with suggestions I would have agreed. But these days it's rare to see more than 10 suggestions and there are way more reviewers than before, all happily reviewing. So I don't think those extra items are any problem. – Shadow Nov 24 '13 at 15:19
  • @Arjan nice find, didn't see that post before! – Shadow Nov 24 '13 at 15:21
  • 1
    ...unless, @ShaWizDowArd, folks who're reviewing suggested edits would be reviewing something else that would get them a shiny badge or something else ;-) – Arjan Nov 24 '13 at 15:21
  • 1
    And, @Shog9, I feel the quality of Stack Exchange would be helped much more by getting rid of posts that are just noise, rather than fixing links to documentation that I can find anyway once an answer has given me a clue about what to search for. I'd LOVE Stack Overflow to be my one stop for finding solutions again, like it used to be for a few years. (That said: I don't know how much time is involved in reviewing suggested edits and in the clean up.) – Arjan Nov 24 '13 at 17:21

You must log in to answer this question.

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