Does nobody grasp the irony that this very discussion is not threaded?
Does nobody grasp the irony that this very discussion is not threaded?
Code://try //{ if (a) do { f( b); } while(1); else do { f(!b); } while(1); //}
Of course it can, but the threaded mode has some important drawbacks:
1. When trying to read new posts, you need to skip about the tree. The posts I want aren't consolidated together.
2. You're limited to the site creator's definition of 'new'. Sometimes I don't want to see every new post since I last was in a discussion - maybe I want the last two hours' posts, or the last hour's, or the last fifteen minutes'. Easy to do on a flat post; I just scroll to the point where the posts I want begin. When the posts are all within a large tree, then any definition of 'new' other than the one being decorated is challenging to find.
I'd say it's better, yes. I rarely ever find a need to read only a particular portion of a discussion; I plan to read everything anyway, so there's not any additional material to read.There is only six posts and already you don't know if the replies are meaningful to the original post or only a tangent brought up by a poster. There is more material for me to scan through manually because the forum isn't doing it for me. Even if the forum users are consistent enough to mark there posts as replying to a specific individual, I still need to read that mark to know if or how it is relevant.
This is better? Really?
Further, threaded discussions have their own problems with trying to impose structure on data - conversations don't always fit that structure. You can make a post that simultaneously addresses points from different branches of conversation - or two branches as a whole could converge on a shared topic of conversation. In that case, the structure makes it harder to find all of the content that you're interested in because no single branch has everything. Further, when this occurs, even if you do plan to read both branches, the branch structure makes it harder to tell the chronological order of posts between the two branches, so a discussion that spans branches becomes inherently harder to follow.
I'd argue, in fact, it's easier to process unstructured data than structured data where the actual data doesn't fit into the structure imposed upon it.
You ever try a pink golf ball, Wally? Why, the wind shear on a pink ball alone can take the head clean off a 90 pound midget at 300 yards.
[Edit]
Now I can edit my freaking post again.
*sigh*
Whatever.
[/Edit]
I'm not sure how that's irony, but yeah the multiple views is pretty nice.Hybrid mode has been my choice for a long time. That's the irony.
Except, you know, when they are as exactly as I described in the post to which you are responding.When trying to read new posts, you need to skip about the tree. The posts I want aren't consolidated together.
So, once again, we see that people complaining about a particular format are actually complaining about the software doing the formatting.
*shrug*
There is a lesson here folks: don't use crappy forum software.
Soma
Last edited by phantomotap; 12-18-2012 at 07:30 PM.
You just argued for one of the strengths of threaded discussions. Threaded discussions are topical. You don't go around trying to answer different threads on a single post. Instead you produce one post per thread. This helps structure the information further, limit the size of the posts and reduce the amount of quoting. The chronological order of posts on different threads is absolutely of no importance when threads remain topical... which usually they do.
This is particularly interesting because there can be nothing visibly "thready" about a threaded discussion. You have threaded models that don't implement any sort of collapse/expand features and others than have all threads fully expanded by default. The former is typical of comments software for websites, the latter common on some web forum software (as well as email clients, bugtrackers, etc). You can have a flat display to a threaded conversation. You can't have the other way around unless every participant agrees on some posting rules.
This says nothing to me. I can as easily go around and say "it's easier to process structured data than unstructured data where the actual data fits into the structure imposed upon it. We will just go around in circles trying to argue each other out.
*All* data can be structured. The flat model offers it too a form of structured discussion. What you are ignoring is that you don't approach either model and do what you want of it. You adhere to it. You follow its implications. And this is why it's possible to have an healthy, comprehensive and fairly easy to follow threaded discussion on newsgroups, email and bugtracker and with it develop the kernel of an operating system, a web server and an open source operating system.
...
But above all how much more of this discussion do we really want to get into? The article author ignorance isn't just demonstrated by his baseless blanket argumentation, but primarily by the fact he insists in seeing a problem where there is none. Threaded discussions and Flat discussions aren't competing for 1st place except in his small head. There's no race. Neither there's a social or psychological advantage or disadvantage to one that can't be argued with an advantage or disadvantage to the other. There's no dichotomy. Both models apply equally well and we've learned to live with them. They are both healthy ways to debate and organize information.
What's really scary is this idea that suddenly I should defend that one should be put out of commission. How much of an idiot am I, for even entertaining the idea that something like that makes sense?
- What's the better color? Blue or Yellow?
- I guess it will depend on your favorite flavor of green.
Last edited by Mario F.; 12-18-2012 at 07:06 PM.
Originally Posted by brewbuck:
Reimplementing a large system in another language to get a 25% performance boost is nonsense. It would be cheaper to just get a computer which is 25% faster.
How many posts now have said something of the form "the threaded format can't show me the newest posts"?This is particularly interesting because there can be nothing visibly "thready" about a threaded discussion.
I think the "problem" is simply a lot of people responding haven't seen forum software that can present the information we want, as consumers of the forum, without doing the "visibly thready" or "visibly flat" thing.
Aquamarine all the way!I guess it will depend on your favorite flavor of green.
Soma
Only to the extent that users are willing or able to produce data in that structure. For example, I'm sure I break a lot of this site's "threaded" views (which I didn't even know existed) because I very frequently multiquote, as that's the way I think about a conversation. I don't think about a conversation in the threaded structure (in fact, I would find it awkward and mildly inappropriate to reply multiple times to say multiple things).
Different views can help alleviate one aspect of the problem - the problem of the view being inappropriate for the task you are doing - but they can exacerbate the problem of people violating the structural assumptions being made in the data because they are no longer aware of the constraints imposed by the structure of the data; they don't see the effects. For example, viewing a threaded discussion in a flat view encourages multiquoting and joining of different branches of conversation, which violates the assumptions upon which the threaded view relies. Especially when you're trying to build a tree hierarchy, it doesn't take a large number of people breaking those assumptions to impair readability - since the problem of a post in the wrong branch not only impacts that post, but the replies to that post as well.
You ever try a pink golf ball, Wally? Why, the wind shear on a pink ball alone can take the head clean off a 90 pound midget at 300 yards.
I would guess that if you multiquote, your post would be handled as a reply to the op, instead of being duplicated, or placed some other common thread (maybe who they replied to instead). It's not that disturbing anyway.
I actually tried to look for a multiquote in your history, for a bit, you don't do it as often as you think here.
Originally Posted by brewbuck:
Reimplementing a large system in another language to get a 25% performance boost is nonsense. It would be cheaper to just get a computer which is 25% faster.
I think you are arguing too heatedly about a single author's opinion. It is as if you take this too seriously.
So you don't agree with some of the author's opinions. Then you call the whole article a mess and denounce it?
Meh. Anyway, I'm just going to throw my lot in with Cat and the author. Certainly, flat approaches have certain disadvantages. But to me, those disadvantages do not really matter. I tend to read all posts or none anyway, no matter who replies to who and if they are sidetracks in the topic or not.
Threaded approaches are just annoying to me because of the reasons outlined in the article you seem to dislike so much.
Hmm... I think you are seeing too much. This has been a friendly debate, if I ever seen one.
I don't agree with the author's article, of course I denounce it. You agree with his article, of course you support it.
I mean, I made myself clear why I disagree with the entirety of his article, regardless of him making here or there some valid point. What you can't expect of someone who disagrees with something is to, err, support it.
Originally Posted by brewbuck:
Reimplementing a large system in another language to get a 25% performance boost is nonsense. It would be cheaper to just get a computer which is 25% faster.
Ah, so you want a meta discussion on this topic.
See? Great example of the advantages of a threaded model. That way you could try and argue i'm just bitter and I could try and defend myself saying I'm not. And everyone else could just ignore that stupid debate by collapsing that thread.
Originally Posted by brewbuck:
Reimplementing a large system in another language to get a 25% performance boost is nonsense. It would be cheaper to just get a computer which is 25% faster.
Oh, I can see the advantages of a threaded model. I do not disagree with that. I just find that there is no good implementation of the threaded model, which is pretty much what the article seems to discuss.
All threaded models seem to have a lot of drawbacks that flat models do not have. Of course it all comes down to your preferences when reading a thread, though, which ultimately decides which model is better for who.
I still find that the drawbacks listed in the article really kills the threaded model because I prefer to just follow all discussion, and not having to bother expanding nodes and things.