There's little in what you say that I can disagree with. In fact, most of the difficulties were obvious to me before my post but I concentrated my comments on the higher issues rather than being concerned with the more mundane practical problems that you've rightly mentioned. It seems to me that we need to have solid objectives about what we actually what to achieve before we can clear away the unacceptable dross that plagues the web.
As I see it, there are several levels to dealing with the 'messy' posts issue (topics where there are large numbers of replies). In essence, these 'levels' represent different classes of problem and thus they require different (more specific) solutions.
1. First, there's the 'uncompleted reply' problem. This is where one intends to reply to an existing post and for various reasons one is unable to do so. Reasons may include the website's cut-off time for posts being exceeded or that the poster hasn't had time to reply to a post for various reasons. As I see it, this is a bigger problem than missing an initial post to a story because it can leave existing arguments in a specific thread to remain open and or be incomplete.
On more than one occasion on HN I've said I'd get back to someone on a matter and missed the timeout deadline for various reasons. This unfinished action leaves that thread incomplete in that there is no logical outcome or conclusion in respect of the earlier commentary. The problem is exacerbated if the person to whom I'm replying has made new comment that requires a definite reply. Also, the absence of any reply from me is liable to make me look like an unreliable correspondent in the eyes of that person for seemingly ignoring him when in fact matters were outside my control. This may not seem important but I believe such etiquette is actually important, especially so if the previous poster has gone to the trouble of posting information that actually needs a reply.
There are some simple solutions to this: a website can have a definite cut-off time, say 14 days, for new post to a story, but it could have a longer time for those who've already posted a comment and who later want to reply to someone who has already posted to his or her comment/post.
2. The Correcting of Errors and Typos Problem. For me, there's little worse than to revisit and reread one my posts several days after I've posted it only to find some glaringly stupid mistake that makes a nonsense of what I've written and that I can no longer correct it due to an inane website policy that stops latter editing. Mistakes such as the inclusion of wrong or inappropriate words easily happen when one misses, say, a wrong auto-correction entry. Similarly, with other minor errors such as the wrong use of 'its' and 'it's', the same goes for possessives (the correct use of which I actually do understand but it doesn't stop me from mistyping them on occasions).
As everyone in book publishing knows you cannot publish a book without an editor by virtue of the fact that we're all blind to many of our own mistakes, thus we simply cannot pick up certain errors no matter how many times we may proofread them. These days, through necessity, it's not possible to have someone sitting at our side correcting our errors whilst we type. That said, the problem of blindly missing our own mistakes hasn't gone away. In effect, even after 30-plus years, NO website has found satisfactory ways of accommodating/correcting typos after posting.
What web designers are effectively saying to us is that they are 30 years or more behind the times and they could give a damn if they never caught up. Right, there's no incentive from web designers or makers of web standards to bring web publishing standards in line and level with normal book publishing—well that's the view from their perceptive and of course we poor hapless users have never have counted for much anyway. Frankly, it's an unmitigated disgrace, it's yet another instance of the unprofessional nature of the software industry - if you want actual proof then read this: https://www.researchgate.net/publication/247573088_Software'... (despite this article being nearly 28 years old, almost nothing has changed to change its findings).
There are many ways around this problem and they don't involve rocket science. There's the limited editing approach as well as the ability to edit certain words on an ongoing basis, especially so those involving the apostrophe as well as words such as 'them', 'then', 'whose', 'who's' and so on.
I fully acknowledge that changing the wording or meaning of text no matter how small and insignificant the change after someone has already posted to or commented on said text is completely unacceptable and should never happen. Here too there are some easy solutions: if text is altered after a post has been made then it's dead easy to flag that post with say a dagger '†' whereupon clicking on said dagger would reveal the original text that the poster actually replied to. Moreover, many variations along these lines come to mind.
Again, the real problem is that both those who run websites and who write the code for website comments don't actually give a damn about such niceties. Nor do they actually care about the fact that their laziness and ineptitude is contributing to the ever-increasing number of errors that we see appearing across the web not to mention the increasing general sloppiness of the web per se.
Nor do they care or worry that their rotten workmanship is further contributing to increased web information entropy—or, as is more likely, they don't actually understand what increasing information entropy actually means. There are many things wrong with the web and we would be well served if those who actually ran the web set higher standards through example. Applying higher standards to accuracy and presentation is an easy point from which to begin.
3. I could not agree more with the fact that extended 'run-on' results in terrible signal/noise ratios. But again, little or nothing has been done to ameliorate this problem over the same 30-plus years. Moreover, the matter is made worse by websites allowing posters to reply to the first poster in order that their response appear high up in the posting hierarchy. This problem could be fixed in an instant if the typical tree structure that's used in most posts were modified to both better reflect the actual subject matter of the post as well reflect a more accurate relative temporal position in the posting list (both metadata from posts along with temporal data should be used to ensure that posts appear within the hierarchy in the correct and most relevant order). For this to happen it's likely that each thread in reply to any given post may have to split to accurately reflect both its metadata status as well as its temporal position relative to all other posts.
I know that's a badly worded explanation but it'll have to for the moment. Again, I also need to point out that there has been precious little development in the way forums have been displayed over the last 30-plus years. Whether it's phpBB,WordPress, MyBB or whatever, none of them do a sufficiently satisfactory job that would ensure that 'discussions and debate can meaningfully continue beyond a day or two'. The fact that none of these forum programs has either the means or wherewithal in this regard simply means that those who have programmed them have failed dismally in achieving a workable solution.
I repeat that the web is in an unholy mess: if we follow exiting example then we see that it inevitably leads upwards from the lowest common denominator and not as it ought from the highest of standards and from the moist excellent of ergonomic design downwards. I am strongly of the belief that for the web to improve significantly from its current ratbag existence then excellence should always begin at the top. Standards of excellence could start by websites actually agreeing on something as seemingly trivial as reaching an agreement on unified forum timeout procedures.
>As I see it, this is a bigger problem than missing an initial post to a story because it can leave existing arguments in a specific thread to remain open and or be incomplete.
I can see your point of view, however I'd expect the problem to naturally resolve itself if people where provided with the incentive to do so. As you've pointed out, these limitations are here almost always because of website policies. I am more interested in starting with the reason these policies exist in the first place, than starting with their removal. While I dislike these policies and the reason they exist, I would be cautious and not simply assume they're entirely inane.
What we have is a proverbial Chesterton's fence that had to be explicitly added to the software, and that has become widespread among forum/discussion platforms.
The more interesting problem to me here is finding a way of structuring online conversation so that people will find it natural to enrich an existing discussion without suffering from noise. The reason it is more important is that threads are mostly not abandoned because of cut-off time or hardcoded limits, mostly people leave when a thread falls out of view and becomes hard to find again.
The problem we really have is a problem of noise, organization, and visibility of information.
--
Let's look at another adjacent domain for a minute. There exist websites serving structured, user-contributed, long-lived information that people continue building upon for years, and not hours.
When people use software like Mediawiki, contributors have no trouble finding the right page to lookup information in or to add content to. This shows that a website can scale to very large amounts of information without a big duplication problem, and without needing to limit the lifetime of user-generated content to a few days.
If we contrast this to forums, the same topics are discussed again and again, month after month, but each individual contribution to a thread is lost after a few days because of the structure of discussion forums.
Unfortunately, while traditional wikis are great for building articles, they do not make for appreciably better forum software.
I believe there must be a way to solve the problems of long-lived discussions in forum software, by taking inspiration from how other sort of platforms manage to scale to vast amounts of information without excessive noise, duplication, or time limits.
But this has to be a more fundamental change than just adjusting the limits.
--
>The Correcting of Errors and Typos Problem
I'm unsure how important this is. I see typos as an annoyance, that almost always a sufficiently interested reader will be able to correct on the fly. Language is very much redundant, so that people can omit a "not" in a sentence and a reader still guess that the writer meant for a negation to be in there.
A post so compact that a typo confuses the meaning entirely occurs occasionally, but it is rare, and empirically I've rarely seen a conversations being materially impacted by a typo. Except perhaps for people who insist on reacting to the noise more than they react to the information.
--
I'll also note that even when talking about more substantial errors than typos, this isn't a fundamental or universal problem.
The problem of people editing old posts to change their meaning is well-known, I think, but also rarely matters very much. The only times I've seen that done, it was being played for laughs. A deceptive person would be caught very quickly (be it by archives, screenshots, or other witnesses), and the problem solves itself.
Popular forum Reddit, for instance, happily allows me to edit a month old post. Hacker News has more stringent limits. There is a tradeoff there, and it is sometimes annoying to see a limit being more conservative than it needs to be, but I don't see a fundamentally hard problem being held back by a lack of innovation there.
>Nor do they care or worry that their rotten workmanship is further contributing to increased web information entropy
I agree that ability to edit a post feels very nice, but I think the emphasis you put on this is excessive compared to the harm it does.
To be convinced that this problem is as important as you make it out to be, I'd want to be shown that it substantially affects the quality of discussions, first.
--
>I could not agree more with the fact that extended 'run-on' results in terrible signal/noise ratios. But again, little or nothing has been done to ameliorate this problem over the same 30-plus years.
Right, and this is really THE problem I think is most important. I think we've had some slow improvements, in the form of going from flat discussions (traditional forums) to increasingly threaded discussions, and sorting of comments by votes.
A Hacker News or Reddit thread can accommodate a thousand comments without becoming impossible to navigate, where a 30 year old forum may have a hundred page, resulting in no visibility for anything but the first and last posts, no ability to have multiple discussions in parallel, and all the other obvious problems that directly result from the structure of the software.
I do think we've done a small amount of progress, but it has been excruciatingly slow, and there is room for a much better system that scales to higher amounts of information, and allows people to naturally build upon previous work, without artificial limitations becoming necessary.
> This problem could be fixed in an instant if the typical tree structure that's used in most posts were modified to both better reflect the actual subject matter of the post as well reflect a more accurate relative temporal position in the posting list (both metadata from posts along with temporal data should be used to ensure that posts appear within the hierarchy in the correct and most relevant order).
Right, I agree with this sentiment completely. I think more metadata is a way of addressing this, or at least some more interesting sorting method than letting people arbitrarily gain visibility by responding in the right place. Metadata may be a solution, but the best sort of metadata doesn't require adding many annotations to each post (or people will simply not bother). It should be there structurally.
Visibility (and sorting) are very important to the quality of a discussion, because that directly determines whether relevant or low quality posts will be seen and replied to. That seems obvious, but when software is a chronology (like a traditional forum) or even a tree of threads sorted only by votes and time (like HN), I think there is room for improvement.
>I know that's a badly worded explanation but it'll have to for the moment.
It would certainly be appreciated if you have something more detailed in mind, and if you find the time to elaborate. I'm interested.
>none of them do a sufficiently satisfactory job that would ensure that 'discussions and debate can meaningfully continue beyond a day or two'.
I agree with that, despite what I've said above about some relative level of progress having been achieved. Discussion software should be able to scale in terms of amount of information, and in terms of time. These two properties are only somewhat correlated.
>The fact that none of these forum programs has either the means or wherewithal in this regard simply means that those who have programmed them have failed dismally in achieving a workable solution.
This suggests that crafting a better solution would not be challenging to you. I'm not sure I know exactly what the better solution is, but I agree that we can do better, I'd like to see more people thinking about this, and I think this goes far beyond determining the optimal value for forum timeout limits. It's the structure that needs to evolve faster than it has.
The impact of all of this is not a trivial thing. There are a few billions of eyeballs on the internet, and the quality of discourse directly affects public opinion. If we organized information better, we would create tremendous value.
In some sense, Google understood the importance of this. Wikipedia succeeded in its particular niche. Much of the rest of the world does not seems to care very much about organizing information.
As I see it, there are several levels to dealing with the 'messy' posts issue (topics where there are large numbers of replies). In essence, these 'levels' represent different classes of problem and thus they require different (more specific) solutions.
1. First, there's the 'uncompleted reply' problem. This is where one intends to reply to an existing post and for various reasons one is unable to do so. Reasons may include the website's cut-off time for posts being exceeded or that the poster hasn't had time to reply to a post for various reasons. As I see it, this is a bigger problem than missing an initial post to a story because it can leave existing arguments in a specific thread to remain open and or be incomplete.
On more than one occasion on HN I've said I'd get back to someone on a matter and missed the timeout deadline for various reasons. This unfinished action leaves that thread incomplete in that there is no logical outcome or conclusion in respect of the earlier commentary. The problem is exacerbated if the person to whom I'm replying has made new comment that requires a definite reply. Also, the absence of any reply from me is liable to make me look like an unreliable correspondent in the eyes of that person for seemingly ignoring him when in fact matters were outside my control. This may not seem important but I believe such etiquette is actually important, especially so if the previous poster has gone to the trouble of posting information that actually needs a reply.
There are some simple solutions to this: a website can have a definite cut-off time, say 14 days, for new post to a story, but it could have a longer time for those who've already posted a comment and who later want to reply to someone who has already posted to his or her comment/post.
2. The Correcting of Errors and Typos Problem. For me, there's little worse than to revisit and reread one my posts several days after I've posted it only to find some glaringly stupid mistake that makes a nonsense of what I've written and that I can no longer correct it due to an inane website policy that stops latter editing. Mistakes such as the inclusion of wrong or inappropriate words easily happen when one misses, say, a wrong auto-correction entry. Similarly, with other minor errors such as the wrong use of 'its' and 'it's', the same goes for possessives (the correct use of which I actually do understand but it doesn't stop me from mistyping them on occasions).
As everyone in book publishing knows you cannot publish a book without an editor by virtue of the fact that we're all blind to many of our own mistakes, thus we simply cannot pick up certain errors no matter how many times we may proofread them. These days, through necessity, it's not possible to have someone sitting at our side correcting our errors whilst we type. That said, the problem of blindly missing our own mistakes hasn't gone away. In effect, even after 30-plus years, NO website has found satisfactory ways of accommodating/correcting typos after posting.
What web designers are effectively saying to us is that they are 30 years or more behind the times and they could give a damn if they never caught up. Right, there's no incentive from web designers or makers of web standards to bring web publishing standards in line and level with normal book publishing—well that's the view from their perceptive and of course we poor hapless users have never have counted for much anyway. Frankly, it's an unmitigated disgrace, it's yet another instance of the unprofessional nature of the software industry - if you want actual proof then read this: https://www.researchgate.net/publication/247573088_Software'... (despite this article being nearly 28 years old, almost nothing has changed to change its findings).
There are many ways around this problem and they don't involve rocket science. There's the limited editing approach as well as the ability to edit certain words on an ongoing basis, especially so those involving the apostrophe as well as words such as 'them', 'then', 'whose', 'who's' and so on.
I fully acknowledge that changing the wording or meaning of text no matter how small and insignificant the change after someone has already posted to or commented on said text is completely unacceptable and should never happen. Here too there are some easy solutions: if text is altered after a post has been made then it's dead easy to flag that post with say a dagger '†' whereupon clicking on said dagger would reveal the original text that the poster actually replied to. Moreover, many variations along these lines come to mind.
Again, the real problem is that both those who run websites and who write the code for website comments don't actually give a damn about such niceties. Nor do they actually care about the fact that their laziness and ineptitude is contributing to the ever-increasing number of errors that we see appearing across the web not to mention the increasing general sloppiness of the web per se.
Nor do they care or worry that their rotten workmanship is further contributing to increased web information entropy—or, as is more likely, they don't actually understand what increasing information entropy actually means. There are many things wrong with the web and we would be well served if those who actually ran the web set higher standards through example. Applying higher standards to accuracy and presentation is an easy point from which to begin.
3. I could not agree more with the fact that extended 'run-on' results in terrible signal/noise ratios. But again, little or nothing has been done to ameliorate this problem over the same 30-plus years. Moreover, the matter is made worse by websites allowing posters to reply to the first poster in order that their response appear high up in the posting hierarchy. This problem could be fixed in an instant if the typical tree structure that's used in most posts were modified to both better reflect the actual subject matter of the post as well reflect a more accurate relative temporal position in the posting list (both metadata from posts along with temporal data should be used to ensure that posts appear within the hierarchy in the correct and most relevant order). For this to happen it's likely that each thread in reply to any given post may have to split to accurately reflect both its metadata status as well as its temporal position relative to all other posts.
I know that's a badly worded explanation but it'll have to for the moment. Again, I also need to point out that there has been precious little development in the way forums have been displayed over the last 30-plus years. Whether it's phpBB,WordPress, MyBB or whatever, none of them do a sufficiently satisfactory job that would ensure that 'discussions and debate can meaningfully continue beyond a day or two'. The fact that none of these forum programs has either the means or wherewithal in this regard simply means that those who have programmed them have failed dismally in achieving a workable solution.
I repeat that the web is in an unholy mess: if we follow exiting example then we see that it inevitably leads upwards from the lowest common denominator and not as it ought from the highest of standards and from the moist excellent of ergonomic design downwards. I am strongly of the belief that for the web to improve significantly from its current ratbag existence then excellence should always begin at the top. Standards of excellence could start by websites actually agreeing on something as seemingly trivial as reaching an agreement on unified forum timeout procedures.