Scrolling Problem

Like Gmail's 'never delete an email' promise? :)

小心！ I have been gaining on Gmail's capacity limit .. my % is steadily creeping up.

Great, glad it's working for you, too. We'll keep an eye out for any other problems, though.

'Is there a physical limit?' There shouldn't be, save for disk storage space, but I don't think Pretzl can come up with lesson topics faster than we can buy more storage space :P

Wow - you're good. :)

I just tried it myself and it worked perfectly. In all honesty the last ten times I have tried (I usually try to see what new and weird idea Pretzl has come up with) it didn't work at all. Well, I guess you can consider that problem solved .. I will let you know if it happens for me again.

'Not the longest thread' - is there a physical limit? I'm guessing one of the first Newbie lessons might be the longest.

Just looked at the lesson suggestion thread, and this time it worked for me every time I tried it. It's actually not the longest thread on the site, either, so I don't think it's anything related to that... maybe it was a temporary problem when we were both looking at it?

Do you happen to have a link handy where the problem occurs for you? I'll keep poking at it and see if I can make it fail too (I'm almost sure it wasn't working for me last night on my iPad when I checked).

It certainly could be true. I remember reading an article some time ago that said Shanghai would be one of the last cities in the country to offer FIOS, because there was already so much infrastructure in place that they would roll it out in under-covered cities first, and then into the bigger cities, with Shanghai being last. It might also be a thing of new vs. old buildings. I live in a pretty old complex, and when I tried to get the fastest service China Telecom offered I was told that the phone network in my apartment block didn't support it. My impression from a couple trips to Kunming is that a lot more of the urban spaces are newly built, perhaps with better wiring in place.

Hey John, thanks for looking at that. I was talking to a local recently who works in the media here and he said something along the lines of 'the Internet in Kunming is better/faster than in Shanghai'. Have you ever heard of this? Is there anything in it? Now I'm aware from poddies that there are apparent differences from place to place, both in 'speed' from time to time (i can recall others saying that the Internet was incredibly slow when I was having no problems) and the GFW.

ChinesePod is certainly learning. I certainly am, at least! (but no, seriously, we're honestly trying to get more focus on UX and customer service). I really want to see is problem fixed, and as I said before I don't set tech timetables, but I'll do what I can.

As for issues in the past, I can't really comment as I wasn't around (I didn't come back to CPod until July) but I can do my best to get communication working more smoothly in the future.

RE: the lesson suggestion thread, I'm not sure what's going on there, but I'll take a look at it in the morning (when I have access to the source code to see what's up) and get back to you. Seems like I'm having the same issue. It may well be the longest thread on the site, so it may have hit a wall we didn't know existed.

Hi John

I want to add my thanks to you for showing an interest in this, and for taking up the cause with good grace. I don't understand a word you've said but I'm still impressed (it all sounds plausible.) It suggests that ChinesePod is learning. As RJ mentioned, 'Nobody even acknowledged that the problem existed'. Quite true. There was a period so long ago.. (was it January?) when many poddies were complaining and I can recall staff saying (finally, after much prodding) that they couldn't reproduce the problem. No-one bought it and it has left a bad taste in the mouth - like a serve of 臭豆腐 on a stick. I have left well alone for months now - maybe now I have had my say I will do that again. It's a great site and a great business - keep up the good work.

But while I am here, one final thing (perhaps related?): on Pretzl's Lesson Suggestions page, it is so long that I now NEVER get to my destination if I click the link for a new comment. It takes me to the top of the first page and that's where I stay. What is happening there I wonder?

Hi RJBerki,

Turning off the collapsing function is definitely an option. We're sitting down either tomorrow or Monday to see what technical options there are, sorting out the options, etc. I'll let you know what's decided.

I think the threading, though, is a net positive. We've gotten good qualitative feedback on it, and it has resulted in a marked increase in comments. Not having multiple conversation interwoven into a single (paginated) thread makes continuing those conversations much easier. I agree that this should have been implemented better, but I don't want to throw the baby out with the bath water, so to speak.

Not acknowledging the double page load issue was my fault -- it was acknowledged internally almost immediately (Calkins posted quite soon after the update was posted), but I didn't match the internal email with an external comment. That's something I'll be more careful with in the future. It took a couple of days to fix because it was introduced with some other changes, and the tech team wanted to make sure that rolling part of the update back wouldn't break the rest of it, and whatever changes were needed to fix it without affecting anything else had to be worked into the tech team's schedule.

As for why it wasn't caught before deployment... I really don't know. I'm not part of that process, and I don't want to speak for other people.

I really hope that we can get this right soon, because I think the threaded comments and ability to link to individual comments within them is a pretty great feature for facilitating conversations. For what it's worth, I do very much appreciate your patience and feedback, too.

Thanks John for taking the time. I am often much more disappointed in how problems are handled at cpod, and how slowly they are handled, than by the problems themselves. Can you turn off the collapsing function until a more elegant solution is found? Just a thought. My other thoughts go like this. Why was threading rolled out before they knew how to do it without breaking something? Surely this came up during the analysis of what can or will be affected by the planned changes or at least seen in some beta test. My other thought is that you could do away with the threading nobody wanted anyway. It serves only one purpose that I find useful and that is that it makes the "replies to me" function possible. On the other hand, I didnt need "replies to me" before things started getting buried due to threading.

As thrilled that I am that the double page load and loss of back button functionality has been fixed, why did it take so long? This happened once before so the tech team has had some experience with this. Nobody even acknowledged that the problem existed and it took several days to fix. The other part of that question is, after whatever change they made that caused this blip, why wasnt it noticed by the tech team. Dont they do a post change check to make sure nothing has been broken? Ok, Im done. My professional experience just hasnt prepared me for the lack of urgency I experience, or seem to experience at cpod. Im on your side, really. This is such a well done language tool and I would like to see it be "great" in ALL ways, and not become the web programming skunk of the world. Thanks again and I hope you dont take offense but I call them as I see them. Even the good stuff. And there is plenty good here too.

RJBerki, I totally agree regarding needing to communicate better, especially on technical issues. As I said in my above response to paulinurus, I was hoping to improve communication just a little bit, on this one issue, with this post. I certainly didn't mean for the throwaway intro to be taken as mocking anyone.

As for the question you posed, the double bounce seems to be because the code isn't calculating whether or not the comment you're looking at is in actually in the collapsed portion of a subthread (and the double bounce seems to only happen in subthread comments). There are, I think, four kinds of possible comments:

1. Top level comment, on the first page of comments (example). These work perfectly.

2. Sub comment on the first page, in a thread that doesn't have more than the limit that's shown before older comments are collapsed (example). These also work perfectly.

3. Either of the above two types, but on any page but the first (example). In this case, the first page is loaded, and then the calculation is made and there is a redirect. This shouldn't be done this way -- the calculation should be made in advance, and an HTTP redirect should be used to bounce you directly to the correct spot. However, in this case there is no bounce once the correct page is reached (other than the normal browser bounce when you use an internal hyperlink, but there's nothing that we or anyone else can do about that).

4. A comment on a sub-thread that has enough comments on it to have collapsed some of them out of view. This is where the real problem is. If the comment is inside the collapsed comments, it's basically the same as #2, it just takes longer because the page has to load, and then the collapsed subthread is expanded, and then the page slides down to the right spot (example), because until the collapsed comments are unrolled the comment that the internal hyperlink is pointing to doesn't exist on the page. However, the issue comes when the comment is one of the ones that's shown (example). What's happening here is that the page loads, the comment that the internal hyperlink is pointing to already exists on the page, and the browser slides down to it (just like in #2, or the first case in this point). However, after that happens, the thread is automatically expanded, which displaces the comment down the screen, and then the browser slides down again. If this occurs on any page but the first page, then you have the bad effects of both the redirect from #3 and this bounce. Definitely not optimum UX.

Ideally, the problem with #3 will be fixed with a transparent HTTP redirect (that doesn't break the back button), and then the code would determine whether or not the collapsed comments need to be unrolled, and only unrolling them if they have to be (in which case you get the first case of #4, which is basically only the normal browser bounce that happens with any internal hyperlink, and is worth it to be able to link to comments directly without having massive page load times caused by giant subthreads).

For what it's worth, I'm copying all of what I've written above to our tech team. I don't set tech priorities, but I do have a voice in improving user experience, and I want to see this fixed (the programmer in me likes elegant solutions to tricky problems like this).

If my response to another user offended you, I am truly sorry. I think (and if I'm wrong, xiaophil, please let me know!) that he took it for what it was meant to be, which was a lighthearted intro to a pretty serious look at what the issue is. I think that we often do a poor job explaining the technical causes behind some of the issues on the site, and I was trying to in my own small way address that. Again, sincere apologies if my tone was inappropriate.

John, I have to say that you have quite an attitude responding to poddies complains. Do bear in mind that we've gone though 9 months (and perhaps more to come) of hell using the community site, so much so, it does at times feel like you guys do have a conspiracy to test our patience.

thanks John for letting us know the double page load followed by bouncing that was introduced at the beginning of this week, was fixed this morning. Thanks also for communicating. I think you would find that users would tolerate a lot if you let them in on why, and respond rather than ignore. Cpod is a complex website, you do manage a lot, but that is part of what is required. One thing that puzzles me though is that if this is a position calculation issue, why does it land immediately in the correct spot - then bounce - then return to the right spot? It seems to have it right, then goes through all these gymnastics only to decide it was right in the first place. The calculation of position must have been done, or it wouldnt always start in the right place. It seems to me like more of a double calculation, a repeat once it has already been done correctly. What am I missing?

Hey John, thanks a lot for "coming clean". I hope I didn't sound too testy, as I really don't feel that annoyed, but it has been something that I have been curious about, and obviously I am not the only one. Anyway, your explanation is quite hard to understand, by necessity I'm sure, but let's just say that you have successfully convinced me that it is indeed a problem that is more than meets the eye. Thanks for the detailed reply!

Hi xiaophil,

"Perhaps some CPod staff member can come clean about this problem?" But then all of the work we've invested in constructing a carefully planned conspiracy surrounding the page bouncing issue will be for naught! :D

But in all seriousness, the core problem *is* pretty tricky to solve. The issue is, more or less, that the position of a comment on each page needs to be calculated, and it's not static. If the comments were single-threaded, it would be easy, as you could simply count the number of comments in the thread ahead of the current comment to find its ordinal position in the comment list, and then divide by the number of comments per page to get the correct page. However, because comments are threaded, and we want to keep the number of comments per page capped at a certain number (to reduce database load and bandwidth usage for us, and loading times for our users), this is not trivial.

Now, the calculation for top-level comments under this system is still pretty easy. The real issue is with threaded comments. First, the page position of the threaded comment's parent needs to be calculated, and then the position of the comment itself in the subthread needs to be calculated. Because we fold down all but the latests comments in long sub-threads, this needs to be checked for and, if the comment being linked to exists within part of the thread that would normally not be shown, that collapsed section needs to be expanded for the comment to be shown. Because of the way pretty much all browsers handle in-page hyperlinks (the # part of the link), this results in the bouncing as the browser attempts to center the window where the content you're going to see is, and then when the collapsed section is expanded the page suddenly gets longer and the window bounces again.

I agree with you, though, that this needs to be fixed, and is fixable. The most egregious issue, with the double page load followed by bouncing that was introduced at the beginning of this week, was fixed this morning. We do indeed plan on overhauling the comment system on the backend so that we can do these calculations on the fly, and eliminate (or at least reduce as much as possible) the bouncing, but coming up with a system that will remain performant on a system with as many comments and threads as ChinesePod is not a minor engineering detail (if you look at Facebook, for instance, you can't link to an individual comment within a thread, and there are likely very good reasons for that). It's something we want to get right, without abandoning or degrading the threaded comment system we currently have, which I think most users agree was a major step forward in making the comments on ChinesePod more usable.

I hope this sheds some light on the issue.

Just now it didn't work for me, either. Trying it again here …

And now it is working. It wasn't working for a couple of hours at least.

I can open older replies. And this is a 'Reply'.

You do realize that this got worse about 4 days ago and now includes a broken back button function. Surely we can at least get back to where we were when the original complaint was made.

After entering this thread, the url changes in time with the first flash. "page=1' which isn't there at the start, suddenly appears near the end of the url. Surely that is the cause of the first flash.

Anyway, for Firefox users, to 'go back', right click on the 'go back' button and select 'Community >> Chinese Pod', or whichever page you want to return to.

Its sad that users have to look for their own fixes to problems that CPod deem aren't a priority.

Users that don't look at these threads are not going to have a complaint. And I really don't think the majority of new users look at comments.

It took me ages after discovering this site before I started looking at threads. Even then I was simply browsing through lesson-threads while studying the related lesson. Those users are not going to see the problem - this issue only arises after clicking on a particular comment on the Dashboard/Conversations page.

Any survey of users should only include those that look at comments through the Dashboard/Conversations page (or those that would if not for the issues).

Have you seen a glut of new users since groups/threading/dashboarding? Because that was half the point.

... regarding the scrolling concerns from our users ... Unfortunately, this is not a priority issue for our tech team to resolve at this time.

Peter I appreciate your candor about this, even though the news is disappointing and surprising to me.

Our "scrolling concerns" described in this thread consist of 1) the original complaint, + 2) the extra "flash" and adjustments, added two days ago. Then there is also 3) the inability to go back (leave the page) via the browser Back Button, also new. So are all of these non-priority now? The tech team will get around to them but not now, is that right? 2) and 3) are bad--ridiculously so as RJ has said.

What I find most disappointing here is that not one CPod staff member has come up to say they empathize, sympathize or can see what we're on about (a natural, human reaction, if only you guys would try to experience the same site we are), with the frustration, discomfort and inconvenience your users--maybe it's only 7 of us, but still--have expressed. None of you really sees what we're on about? What we complained about doesn't bother you? RJ says "Has anybody there actually experienced what it is I am talking about?" I wonder this too.

And I've never seen any indication that any technical person is laughing up their sleeve at anyone, I think that's unfair. The worst they could be accused of in that respect is poor communication, and recently they've come on leaps and bounds.

... the fact that so many people have indicated their frustration makes this a serious issue.

It's only about 7 of us here Bob. I don't think we're making up the majority of users.

I agree that this is the biggest problem with the site at present, but suggest anyway that CPod ask new and trial users for their feedback, and listen well to them. If no one new to this site thinks there's anything wrong or user-unfriendly, then it could just be that us old buggers / seasoned complainers are out of our minds.

Peter,

I know you are just the messenger, but this is how I feel about that:

I reported this two months ago. Surely even low priority items could get some attention in two months. I could have lived with it the way it was, but now its ridiculously worse. I suggest another reason why they don't fix it. They don't know how. In most industries, if a product was this badly broken, no one would go home until it was fixed. Cpod IT evidently exists in a vacuum. They seem blissfully unaware of convention or best practices. They have a lousy work ethic and no sense of urgency. I get the impression sometimes that they are laughing up their sleeves at bosses they despise. To make matters worse, those charged with the task, have lousy communication skills. Questions are either ignored or you often get an answer to a different question. This shows how much they are paying attention. Now you are telling me this is OK with Cpod. No plans to fix it. Not important. Has anybody there actually experienced what it is I am talking about? Does anyone actually log on as a user through an external web portal? Maybe they just don't know how bad it is? As it is, I must close my eyes after each page change to avoid needing an airsick bag. It boggles my mind that their standards are so low that this is not a priority for them. That they care that little and now are almost bragging about it. Its the little things that make customers happy. Fixing these things would go a lot further than creating a dashboard if you want customers.

Xiao_liang, Its not just this issue that makes me hesitate when I think about giving this company more money. Lets just call it the straw that broke the camels back. Cpod seems for some reason to be a company that as a whole, is less than the sum of its parts. I am not privy to enough information to say exactly what is wrong, but I would say some leadership is badly needed. Maybe its like the "bad service restaurant" (see lesson of same name) and they are hoping we start to find entertainment value in their lack of web prowess and customer care? Its so bad, its good. The "BSW". The lessons are good after all. Win-Win.

Aha, that was something I didn't realise because of how I read the site. Turns out I always read threads by opening them in a new tab, then close the tab when I'm finished.

It's like trying to read during a bungee jump.

I'm basically a man who is happy with the way things are, but frankly speaking, this "boucing" feature is one of the few defeats of Chinesepod I can't put up with.

I click on a comment then realise I am not interested. 'Go back' doesn't work, so I try to scroll to the top of the page to hit the Community tab. Before I get there, I am transported back down to the bottom of the page. So I try to scroll up again, and the same think happens again. By the time I get out of there, I am fuming. And this happens multiple times in a session.

And its not just that scenario. Before I read a comment, I like to scroll up a bit and read the previous comments, just so I can put the new comment in context. I have to do this scrolling 3 times before I am successful.

That 1/2 second is a gross underestimation. The two hyperspace shifts typically play out over a number of seconds. Often, I have already begun reading the earlier comment, before I am rudely interrupted.

I think the fact that so many people have indicated their frustration makes this a serious issue. In fact, I think it is the biggest issue with this site at present.

Breaking the back button is one of the cardinal sins of web design.

Honestly, I'm not sure why it's such an issue. It's a presentation thing, but it doesn't affect the ability to read comments. It's like 1/2 a second before you read it. I agree it doesn't look too professional, but I'm sure we could find something a bit more serious to quit over. ;-)

Hi RJBerki, we've reported to our tech team regarding the scrolling concerns from our users. They said this is probably caused by the coding to hide comments with more than 5 sub-comments. Unfortunately, this is not a priority issue for our tech team to resolve at this time.

John and Peter were posting at roughly the same time yesterday afternoon (my time). At this time, your wife's (or your?) comment was close to the top of the conversations page. They each responded to threads either side of it, while leaving this thread untouched.

my god, are you people really this incompetant or are you mocking us (the users)? I have been here 3 years but at this point when my subscription runs out in August, I will not be renewing. I think you owe all your users some time credit for shoving the dashboard down our throats and the trainwreck that has resulted.

this is now much worse. And as Go_manly says, now you cant return to the community page after viewing a post. Hitting the back button only results in another ride on the scroll jumping roller coaster, which brings you right back to where you started from. The only way back is to hit the community page link at the top of the page. From bad to worse.

The original flash happens in comments with more than 5 sub-comments. Typically only the most recent 5 comments in a thread are shown. After reaching the desired comment, it then decides to open up the whole thread for viewing, before returning to your comment.

Why does the opening of the thread need to be played out in front of our eyes. Can't it be done in the background? Honestly, I would rather open the thread manually, than have this awkward mechanism do it for me.

The new issue has arisen since CPod started playing with the page-ordering of comments. It seems this extra flash is actually a change of page. And when you were actually taken to the correct page in the first place, you still get the extra flash just for good measure.

An added problem with this extra flash is that you can't use 'go back' to get out of the thread. It simply takes you to the thread 'pre-flash', which is followed by another flash as you are taken forward again. All you can do is click the Community tab and start again.

Actually I didn't mind the originally-discussed issue, but this extra flash is bad. Hope it gets fixed soon.