I’ve just incremented the OSNews version number from 3.0 to 3.1. I’m not really sure if now is too early or too late to make this change, but some fairly large backend changes have occured recently that prompted me to revisit whether or not this is really the same site that 3.0 was at rollout. Anyway, there’s a nice new feature that comes along with the change to 3.1 — THREADING. Read on to find out about it.

I think many readers will be happy to find out that we have implemented the oft-requested comment threading. Threading comes in two modes: “Collapsed,” in which you can view a linked comment title beneath a parent comment, and “Expanded,” in which you can view the entire comment, associated via indentation. Here is some detail on that.

First off, threading is only available to registered users. This will not change – the editors all agree that threading is an incredibly unfriendly, illogical, counter-intuitive system for tracking comments for those not familiar with it. Although it may seem to make sense, it carries several drawbacks.

   1. Threading does not provide an easy way for a user to decipher which comments have been added since they last visited the site. Currently you must either remember it or somehow divine it. This is the largest drawback, I believe, and the only way around that is to track which comments you’ve seen or to provide that information in a cookie. Since we know that several of our tin-foil hat wearing users do not like cookies, we are forced to track the information ourselves, and its not in our best interest to additionally hammer the database until v3 optimization is complete. Either way, I think everyone agrees it’s unreasonable for a large website to track user views for each story, even if it’s simply the highest number comment you’ve seen.

   2. I experimented with “reparenting,” which is displaying a reponse to a comment which has been voted below your threshold. I found it to be more confusing than anything else, as it was not immediately clear that the comment was a reply. Instead, collapsed parenting has been added, which means that comments below your threshold are ignored unless they have replies which are at or above your threshold, in which case they are collapsed. Doesn’t that sound really confusing? That’s why we’ve avoided threading for so long, because those of us who know how to use it often forget how much assumed logic there is in deciphering what one sees.

   3. Because of the basic usability design of the site, it is not easy to provide a consistent interface that can accurately capture all replies without making the comments per page variable. This means the first few pages are likely to be very long as the replies will always be higher, and the last few pages shorter. I’m looking into ways to fix this disparity, but I don’t think it can be done without rewriting large portions of the comments engine.

For these reasons and more, threading is neither extremely practical or usable enough to be our default interface. We provide it solely as an alternative for our registered users. I can tell you after using it for a few days, it’s very useful, although not always as easy to use as flat view.

For those who are wondering, I intend to add an on-the-fly view toggle to the comment page, although it will almost definitely rely on a cookie. If cookies are a problem… well, you have your preferences page.

Lastly, and perhaps most importantly, the threaded view has not been fully tested in all browsers. The threaded view has been tested in Gecko (Mozilla/Firefox), Opera, MSHTML (IE), and KHTML (Safari/Konqeror). Please consider all other browser support experimental. Additional browser support will continue to be added over the next days/weeks.

We are always asked the same few questions every time we solicit feedback, so let me provide some official answers here and now, and let this story be added to your browser bookmarks for everyone to see. These questions have been answered in the comments before, but I will provide the answers here, lest someone mistake the comments for unofficial.

Why don’t you use valid HTML/valid XHTML/XML&XSLT?
I have had many long conversations about this with many of our users, and it’s generally a stalemate each time. We have very skilled users who have a lot of very valid opinions on this. However, the fact remains that the goal of this site is to provide a website that renders properly everywhere, everytime. The typical argument goes something like this: “If you use valid code, it will gracefully degrade in older browser who will simply not get the style.” Unfortunately, this is simply not true.

We figure if OSNews reports on hobbyist OSes we’d at least like to render the full site on their Netscape 2 compatible browser as well. That’s why there’s no critical javascript on the site – you can even get the full experience on Netpositive.

We’ve gone through painstaking testing to make sure we render anywhere we can, in some cases using individual browser identification. We monitor which browsers hit our site, and know that we often have hits from obscure browsers.

I get lambasted for this viewpoint often, and there are many who very publically tell me our reasons are “lame” or “weak,” but they aren’t. We don’t want the site to be a lesser experience anywhere, and right now, it isn’t. This is the purpose of OSNews, and we believe a large part of what sets us apart from other sites.

Strangely though, for the record, I do maintain a validating version of OSNews, and at some point, we may make it public. However, as of today, the site is intentionally broken (with respect to compliance), and we’re well aware. If should look ok in your browser right now, no matter what browser you’re using.

If you wrote valid code, you’d have fewer versions of the site to maintain.
We currently maintain only one version of the core site, although many sections/pages have a mobile version and a desktop browser version. That’s it: two.

Why don’t you use Slash/Nuke/Scoop?
We understand that OSNews has replicated many features of popular sites like Slashdot, but that doesn’t mean that Slash is the way to go. It just means that they have implemented logical “stuff” that users wanted, as have we. We like the simplicity of our site, and will always listen to users who have suggestions on logical ways to further simplify the interface.

Why don’t you use Perl/Python/Ruby on Rails?
This is my favorite question lately. We get a lot of suggestions to rewrite the site in a new language, but we like PHP and will continue to use it for the forseeable future. If you like another language, we encourage you to use it on your site and discuss it here.

It amazes me that so many people have taken the time to email me about rewriting OSNews using Ruby on Rails. While I have only played with this framework on the side, and must admit it seems very powerful, it’s still very young and should probably be proven to scale on large, high-traffic sites before anyone would consider deploying it.

Why can’t anonymous users post with a username?

Perhaps at some point in time we’ll add this feature. Right now, if you want to use a moniker, you need to register. There are several other benefits to registration, including tons of site customization.

Why can’t we moderate admin comments?
Why don’t moderators use two accounts?

Put simply, there are definitely (definitely) people out there who will mod down admin comments *because* they are admins. We know because our posts were very frequently reported as abuse in the previous incarnation of OSNews. We’ve made an internal effort to not get involved in flamewars and high-tempered debates to avoid these problems, but we will not be enabling admin moderation.

Furthermore, it our site, and we said so.

Why not start anonymous comments at -1?
Again, we’ve discussed this, and we simply have too much anonymous participation to handicap those comments, many of which are valuable. However, we have discussed, and may soon implement, a system by which more trusted users start with a higher score.

Why have my comment votes increased/decreased?
We have complex algorithms that now determine your “trust.” Trusted users are often given more votes, while untrusted users’ votes are regularly reduced. We do this to help weed out the trolls who take joy in ruining a community. I think it’s likely that at some point in the near future, new accounts will start out with 0 comments votes and have to “earn” them via high quality comments, moderation, accepted story submissions, and story rating.

Why can’t I log in/reply on the mobile site?

Mobile improvements are coming. The mobile site will be getting additional features at some point to bring it more in line with the main site, although it’s unlikely that it will support logins.

Since rolling out “OSNv3,” we’ve steadily added new features. Chiefly amongst them are the following, in no particular order:

  • 1. Improved comment template
  • 2. Index caching to improve performance
  • 3. Custom timezones
  • 4. Email on reply to your comments
  • 5. Profile biography space
  • 6. Improved user pages
  • 7. True comment permalinks
  • 9. Friends and Fans (which will be unveiled soon)
  • 10. Improved “Trust” system
  • 11. On-the-fly threshold toggle
  • 12. Moderation review
  • 13. “Read Thread” view
  • 14. “Reply” function
  • 15. OSNews Digest (will be re-activated shortly)
    Just added today – finally…

  • 16. Hovering over a topic icon finally displays the name!!
  • 17. Story title in title of page!

    Although tempted, I don’t think I’ll be implementing this humorous suggestion from Roguelazer, who says, “Add 3 new fields to our User Profiles. They can be “combo box” fields. The first one will be “Stance on GNU GPL”. The second one will be “Stance on Gnome devs”. The third one will be “Stance on OS X/Windows” … Through this addition, most of the comments here can be avoided, just by clicking somebody’s user profile. Brilliant!

    As always, suggestions, feedback, flames, gift certificates and evites can be directed to osnews-crew at osnews dot com, or to me personally at adam at osnews dot com.


    1. 2005-08-25 8:50 pm
      • 2005-08-25 8:50 pm
        • 2005-08-25 8:53 pm
          • 2005-08-25 8:54 pm
            • 2005-08-25 8:56 pm
            • 2005-08-25 8:59 pm
            • 2005-08-25 9:00 pm
            • 2005-08-25 9:05 pm
            • 2005-08-25 10:39 pm
            • 2005-08-25 11:41 pm
            • 2005-08-25 9:00 pm
        • 2005-08-25 9:25 pm
        • 2005-08-26 12:46 am
      • 2005-08-25 10:18 pm
      • 2005-08-26 10:25 am
    2. 2005-08-25 8:50 pm
      • 2005-08-25 8:50 pm
        • 2005-08-25 8:55 pm
          • 2005-08-26 9:52 am
      • 2005-08-25 8:55 pm
    3. 2005-08-25 9:02 pm
      • 2005-08-25 9:11 pm
        • 2005-08-25 10:01 pm
    4. 2005-08-25 9:05 pm
      • 2005-08-25 9:10 pm
        • 2005-08-25 9:11 pm
    5. 2005-08-25 9:05 pm
    6. 2005-08-25 9:09 pm
      • 2005-08-25 9:18 pm
    7. 2005-08-25 9:13 pm
    8. 2005-08-25 9:19 pm
      • 2005-08-25 9:21 pm
        • 2005-08-26 1:39 am
        • 2005-08-26 7:26 am
    9. 2005-08-25 9:22 pm
      • 2005-08-25 9:29 pm
        • 2005-08-25 10:13 pm
      • 2005-08-25 10:12 pm
        • 2005-08-25 10:30 pm
          • 2005-08-26 12:18 pm
    10. 2005-08-25 9:26 pm
    11. 2005-08-25 9:46 pm
    12. 2005-08-25 9:51 pm
      • 2005-08-25 9:53 pm
        • 2005-08-25 10:07 pm
        • 2005-08-26 9:26 pm
        • 2005-08-27 9:10 pm
      • 2005-08-25 9:55 pm
        • 2005-08-25 10:03 pm
    13. 2005-08-25 10:00 pm
      • 2005-08-25 10:03 pm
        • 2005-08-26 9:29 pm
    14. 2005-08-25 10:08 pm
    15. 2005-08-25 10:42 pm
    16. 2005-08-25 10:48 pm
      • 2005-08-25 11:02 pm
        • 2005-08-26 2:13 am
    17. 2005-08-25 10:56 pm
    18. 2005-08-26 12:30 am
    19. 2005-08-26 12:41 am
      • 2005-08-26 3:02 am
        • 2005-08-26 5:49 pm
    20. 2005-08-26 1:42 am
      • 2005-08-26 3:02 am
    21. 2005-08-26 3:22 am
      • 2005-08-26 3:33 am
    22. 2005-08-26 3:40 am
      • 2005-08-26 3:42 am
      • 2005-08-26 7:42 am
    23. 2005-08-26 4:23 am
      • 2005-08-26 3:13 pm
    24. 2005-08-26 7:02 am
      • 2005-08-26 7:04 am
    25. 2005-08-26 8:14 am
    26. 2005-08-26 8:24 am
      • 2005-08-26 10:28 am
        • 2005-08-26 10:32 am
      • 2005-08-26 10:29 am
      • 2005-08-26 12:19 pm
      • 2005-08-27 6:52 pm
    27. 2005-08-26 1:44 pm
      • 2005-08-26 2:06 pm
    28. 2005-08-26 10:12 pm
    29. 2005-08-27 6:57 pm