[DISCONTINUED] MATLAB Answers Wish-list #4 (and bug reports)

Edit: due to the increasing size of this thread, it is continued here.

What should you post where?

Wishlist threads (#1 #2 #3 #4 #5 #6): bugs and feature requests for Matlab Answers

Frustation threads (#1 #2): frustations about usage and capabilities of Matlab itself

Missing feature threads (#1 #2): features that you whish Matlab would have had

Next Gen threads (#1): features that would break compatibility with previous versions, but would be nice to have

@anyone posting a new thread when the last one gets too large (about 50 answers seems a reasonable limit per thread), please update this list in all last threads. (if you don't have editing privileges, just post a comment asking someone to do the edit)

This topic is for features you would like to see for the MATLAB Answers facility itself, and also for bug reports about the MATLAB Answers facility.

This topic is the follow on to the first Wish-list for MATLAB Answer sections and second MATLAB Answers Wish-list #2 (and bug reports). Those grew large enough to become unwieldy; and Mathworks has made enough changes to make a number of the past points no longer of relevance. More recently there was the limited purpose New design of the forum - grey on white which turned into a bug and wish list; I have renamed that for continuity.

I suggest one wish (or bug report) per answer, so that people can vote their wishes.

[Rik]

I took the liberty of opening a new thread and editing this question.

[Walter Roberson]

Thanks, Rik.

[Adam Danz]

For easy navigation and for the historians,

Wish list #1

Wish list #2

Wish list #3

Wish list #4

Wish list #5

Anyone with edit privileges should feel free to update this whenever needed.

Answer 1

[Jon]

Maybe this functionality already exists and I'm not aware of it, but why isn't there a notification of some kind on my profile page that alerts me to new comments, questions, and answers that I'm involved in? An unread message alert, something like this:

Then when you click on "Questions", those that are "unread" are at the top of the queue. And you can mark them unread if you want to come back to them later. I have a hard time keeping track otherwise, especially if I leave for an extended period of time.

[Walter Roberson]

Work around: do a search on

contributor:jon

However, that will find all jon as it is a substring search on the contributor name. It is a bit more effective on names that are more distinctive.

You can search for your answer, Jon, by searching on

answered_by_id:86349

This number has nothing obvious to do with your Author number, and I do not as yet find a way to search by author number. And if you wanted to search on your Questions you would need a differerent id number yet. The only way to find those numbers is to left-click on a link to you and let it pop up the small summary about you, and then go and hover above the link to your Questions or Answers and read off the URL to discover the ID. Or click on the link and then read the tag out of the search box on the resulting page.

Yes, this could be made easier!

I am the only roberson around so it is easy for me to search for my own material, but you might have some difficulty if you searched by name.

[Laura]

This is a great idea!

[Walter Roberson]

A couple of months ago, Mathworks implemented "Activity Feed". If you look under "My MATLAB Answers", it is the first entry.

Notification is only for Questions that you are "subscribed" to. There is a configuration option to determine whether by default you want to subscribe to any question that you ask or answer or comment on. You can always toggle any given question on or off -- including toggling notifications on for a Question you have not responded to at all.

The subscriptions are not back-dated for anything you answered or commented on in the past, but it looks like they might have been turned on by default for any Question you asked.

Answer 2

[dpb]

Again the complaint of excessive white space (futile, I know :( ).

[Walter Roberson]

Whitespace did get reduced, in time.

[dpb]

A little, yes. Still way more inefficient than a classic newsreader interface for the quick perusal of what would be worth looking at in more depth, though.

[Adam Danz]

I'm also a bit puzzled by the purpose of the marginal white space.

[dpb]

Just following "modern" web page design w/ packaged toolset I'd guess. They used up the RH third of the page with the automagic links and ads, etc., but that only covers the top little bit of the main page listing so all the rest are forced to the same narrow page as well...

I really don't see any reason at all for 2" of blank space on the LH side, though...

I'd take just the subject line and two or three lines of the Q? in the short version as the screener view, though...

Frankly, the newsreader view of a split screen w/ just the list of thread at top and a preview window for each thread in turn on bottom would be even better -- then can open the particular thread to get all the bells and whistles if interested; otherwise never need to switch windows/tabs.

[Adam Danz]

The margins appear to be even wider on a wide monitor. The screen shot below (shrunk to 25%) shows this page in Chrome with 100% zoom (ie, no zoom).

Answer 3

[Walter Roberson]

Please add a keyword to Search Answers that specifies sorting order. I frequently want to link to a discussion from a few days ago, so a lot of my searches are of the form

contributor:roberson Some Terms of Relevance

and then I have to search and click on the sorting order drop-down to sort to Date Updated (Newest to Oldest) which does a page refresh. And then afterwards I have to Back twice, with the search needlessly being done again as I get back to the search results sorted by the default Relevance.

I suggest a keyword such as "sortby", such as "sortby:recent" with "sortby:newest" being equivalent to that. The default would be sortby:relevance (with sortby:relevence perhaps being accepted as an alternative for those of us who are sometimes weak in our spelling ;-) )

Answer 4

[Walter Roberson]

When we delete a Question for spam, either by Delete with a spam flag or by clicking on This Is Spam for postings detected as probable spam then when the deletion is finished, the page is forwarded to the Answers home page. The painting of the home page requires that a time-consuming database lookup be done to find the most recent messages and present them. As it is common when doing spam control for us to have 6 to 20 Questions open in tabs (to allow the pages to paint in the background while we deal with them in sequence), that database lookup and repaint is slow and unnecessary, often adding 17+ seconds to the deletion process.

I therefore recommend that when a posting is deleted for spam, that instead of going to the homepage, that it just refreshes, taking us to the "Can't find your Question?" page; we can Back or close the tab or whatever we find appropriate from there.

Remember, deleting spam is not the normal work-flow for anyone but the staff and the decaSpartans (the 3000-ers ;-) ), and that when engaged in spam control the work-flow is seldom to want to see the Answers home page for each Question killed off.

[Walter Roberson]

Thank you to the developers who made it possible to select large number of spam posts and delete them on one go!

Answer 5

[Walter Roberson]

Needed: a Ban Hammer. Or at least a Spam Hammer. So that we can flag a poster as a raw spammer and toss all their postings, past and future, into spam quarantine.

Signed: the overnight spam janitor. Spamnitor? Spamitor? Spaminator?

[Star Strider]

I’ve deleted a bunch of them as well recently. It’d be nice to be able to ‘freeze’ a spammer’s account to immediately prevent further postings from it, and automatically flagging it so that TMW will then take appropriate follow-up action.

Answer 6

[Paul Metcalf]

One idea for MATLAB Central is the ability to follow an author. Currently you can only follow a file but there are some authors for which I'd like to receive a notification if new files are added...

[KAE]

In addition to file authors, I would like to be able to follow contributors on Matlab Answers, and optionally receive an email every time the chosen person asks/answers/comments.

[Rik]

I find the e-mail notifications annoying, so I just check the activity center, but I would think following a contributor would show up there anyway.

This is a very good idea. I currently have a few bookmarks that I check from time to time, but that is very prone to missing stuff.

Answer 7

[OCDER]

Suggestions:

• when answerer is working on a question and someone else posts a new comment/answer, show a popup before answerer submits a redundant answer/comment "There is a new answer/comment. Look before submitting your answer/comment?" Otherwise, people end up working on an answer, pressing submit, realizing a key answer/comment was given 1 min ago, and then deleting recent post.
• prevent people who ask a lot of questions but provide no feedback that the answers work from posting new questions until previous ones are resolved. Ex: "You've asked 15 questions, received 40 answers, but accepted 0 answers. To ask more questions, please accept answers or close them. If you're not getting correct answers, please read tutorial here on how to ask questions to get answers."
• for new users to this forum, a quick quiz before getting started, since people aren't reading the tutorial. Ex multiple choice: Which is a good question? A. help me? B. what's wrong? C. can you do this for me? D. Why do I get an error when using fopen? I have attached the code and full error message...

[Cedric Wannaz]

+1 / I don't upvote though, otherwise your post goes up in the thread and it becomes more difficult to find ;-)

[OCDER]

Thanks for the support. Just saw your detailed answer and felt you, and everyone, should be thanked properly for a LOT of patience :)

[Rik]

Especially that first one is something I get a lot. I try to reduce the chance of that happening my using a slightly tweaked search and sort when going through the list of unanswered questions.

[madhan ravi]

It's an interesting one @OCDER

Answer 8

[Rik]

Maybe it would be a good idea to ask for a confirmation before allowing users to answer their own question. Many people mistakenly use the answer field, which results in weird threads and a lot of talking about not-the-problem (and maybe a bit of frustration).

I think requiring a few reputation points before being able to answer your own question would be a bit harsh, but it would be alternative.

The main downside to the first option would be that people who don't understand the structure, will probably not read a warning popup.

[Image Analyst]

I think that's an excellent idea. The number of times people truly answer their own question is like less than a tenth the times where they add an answer that should have really been a comment to someone else's answer.

[dpb]

And, I already voted +1, but I'd suggest the "confirmation" be "NO" by default and be specifically worded such as to make clear it's a comment unless and only unless it fully actually answers the question.

[Rena Berman]

(Answers dev) Thanks for the feedback!

[Guillaume]

As I've suggested elsewhere, another option is to give editors the ability to move posts between answers and comment and even to promote comments as new questions.

[dpb]

I'd agree 110%+ with just a button to automagically re-categorize A-->C (and vice versa, sometimes someone writes a comment that is, in fact, a very good Answer) keeping poster ID, etc., ... I've moved quite a number over the years but it's tedious so not nearly as many as coulda'/shoulda' ...

Answer 9

[dpb]

I wish the ability to Delete or Close a Q? while someone is preparing an Answer could be disabled -- several times I have just finished a detailed response to a Q? and go to post and discover the Q? is no longer visible. It's really frustrating.

Latest was just yesterday on a Q? that had to do with stock market weekly averages -- both Walter and I had posted comments trying to lead the poster down the path (in somewhat differing directions, but that's useful, too) and I finally decided to just illustrate how retime could solve his problem in one line and had prepared a really nice (I think! :) ) demo and BAM! Black hole... :(

If the poster subsequently chooses to delete or some other editor thinks inappropriate or whatever, well, ok, I can accept that as being inevitable occasionally but in this case as in the others I can recall I think it was a legitimate-enough question (albeit from a naive user) that the answer/demo would be worthwhile even if the OP never came back.

[Walter Roberson]

In cases like that, the site refuses to post but leaves your content in the buffer. I then select the content, Copy, and refresh the page. If the question was only closed instead of deleted, I then reopen, and then post the content. But if it is deleted, then I either shrug and move along or else save the code for later re-use.

[Cedric Wannaz]

I sympathize. This happened to me many times actually (and especially close to mid-term periods I'd say). I ended up printing threads as PDFs before answering (as well as the OP's profile), and emailing the OP if the question was deleted or edited out (with the option to paste back the original content in the latter case). In parallel, I chose not to spend much time answering questions unless the OP had a valid identity or at least an email address with the [Contact] option enabled, and I posted a proposal somewhere for enforcing this option to be enabled.

Yet it seems that it is a no-win situation, because there will always be incivilities that make us loose our time and motivation, and "holding back" our nice answers just in case kills the pleasure of answering.

[dpb]

I am too impatient to take that much effort/time but in this particular case I did save the code as being useful at some future time.

It would still I think be useful option for editors to have the facility to be able to prevent the nuisance factor--can't control what they do to us later, but at least could wait 'til we're done! :)

[Cedric Wannaz]

It would be neat if accounts had some kind of "trust-ability" colored flags that allow contributors to evaluate the amount of time/effort they are willing to spend without having to investigate too much:

Rep > 100 : green

Rep < 100 :

• red: deleted or edited out thread(s)
• orange: no [Contact] option, unverified identity
• yellow: [Contact] option, unverified identity
• green: verified identity (institutional email address or e.g. cross-check with LinkedIn)

But still, there will be students using their institutional email accounts, posting, and deleting their account once they get an answer ..

[Walter Roberson]

I do not use an institutional email address, and I am not in LinkedIn.

[dpb]

Nor I...

[Cedric Wannaz]

Yea well, I'm not sure that I would answer your questions anyway folks ;-)

Back to my initial logic, you have a rep over 100 (=> green) and soon 100k for Walter (and we figured out that the rep won't wrap before max int32 already).

This mechanism would essentially be for evaluating new accounts with no or low rep. The flag could disappear as soon as rep > 100.

[dpb]

Chuckles...

Quite often the posts come from those with 0 points and probably 80% only ever appear once or twice.

I'd just like to have things not disappear irregardless while working on it.

[Rik]

@Cedric, that would be neat idea. Implementation might be quite a pain, so even if it is considered it might take some time. For the reverse we have the mvp, staff, and rising star badges, so it would be nice to also have some indication at a glance if a question is likely to be a waste of time. Sometimes you can look at the name and recognize them for low quality questions, but that doesn't filter the single-hitters.

[Guillaume]

I certainly think that posters shouldn't be allowed to close or delete a question once it has been commented on and even more so if answered. People with low privileges shouldn't be allowed to close questions even if it's their own.

[Jan]

Closing own questions is rarely useful, so I would not miss this feature. But there is obviously a need to delete own questions. This helps to reduce noise and to save the work of the answering members, when no answer is required anymore. I loose about 1 or 2 messages per week by questions closed/deleted during writing an answer. But if the OPs do not have the power to remove their questions, I'd loose much more time by creating answers, which are not needed anymore. I do not see the optimal time point, where deleting an own answer is efficient, fair and useful, so I'd prefer the current policy.

Newcomers use flags often for posting comments and I've spent some time to move them to the comment section in the past. Allowing flags for Rep > 5 would be an improvement in my opinion.

[dpb]

How about giving mvp,staff, ...??? the facility to lock the Q? while responding and until some reasonable time after the response/edit is posted? If the OP still then decides to delete, well, ok, can't make the horse drink once lead to water, but...

[Jan]

@dpb: A lock would be useful. But then I have less collisions with users deleting their question than with Walter, who posts a sufficient answer while I'm still typing. Implementing such a lock might be very hard. Consider, that the current database is not ACID proof: There are several threads with 2 accepted answers. Sometimes an answer posted at mathworks.com needs 10 minutes to be delivered on mathworks.de.

[dpb]

I wouldn't try to lock edits, multiple answers aren't all necessarily bad; often they're sufficiently different approaches as to be worthwhile tutorial info. Even if they're the same or nearly so, "no harm, no foul!". If desired, a "came-in-second" responder can always delete theirs to clean up the thread if wish to at any time.

I'd just have it prevent deletion/close for the duration of the edit and some period after...all the Q? I've had such an experience with are of sufficient content I don't see them as really needing deletion even if the OP were to think "what's the point?" when what he's getting is prompting rather than spoon-fed answer which is what I think happened in the specific case that (finally) prompted the request.

I'll grant it may be difficult to actually implement owing to the distributed nature, but "if you don't ask"... :)

The number of comments indicate it's obviously hit enough as to be a topic worthy of consideration for possible enhancement.

[Guillaume]

My preference would be to not allow posters who don't have sufficient privileges to delete or close questions that have had a comment or answer. If they really want to question closed/deleted they can ask for it (in a comment or with a flag) and a person with enough experience can make a judgement whether or not there is nothing of value in the comments/answer and close/delete it if so.

Posters can still close/delete their own questions if they've had no feedback. Yes, it doesn't solve the fact that it can happen while you're busy writing an answer. It can be frustrating but it's fairly rare.

The advantage of that is that it requires no change in infrastructure. The mechanism is already implemented.

When you delete a question (or an answer for that matter) that have been commented/answered, you not only delete your own content but also somebody's else content, content which they may feel is valuable.

[dpb]

That's good point and would solve probably 95+% of the cases...and all the ones I've come across.

It's not the end of the world kind of a problem but is (at least to me) really, really annoying when it does happen.

[Walter Roberson]

I feel that there should not be any trust ratings built in to the system. I think that it is too subjective and exposes too much risk of abuse.

However, sometimes it would be useful if you could attach private notes, visible only to you, to profiles. This would allow you to create an entry like "trashed questions after answer" or "asks good questions" or "knows a lot about CANBUS" for your own use. For this to be most useful, appearances of the username should be marked with a "you made notes about this person" icon.

[Rik]

@Walter I bet some people already have a system like that on their own, but having it on this site in the first place would indeed be really useful.

[dpb]

Walter, I'd suggest that suggestion (so to speak :) ) is worthy of its own Answer as a forum enhancement.

Answer 10

[James Tursa]

I would like power-users to have the ability to conditionally accept an Answer during the initial 7-day wait period. There have been many occasions when a Question gets answered and it is obviously worthy of accepting. The original poster may not acknowledge it, or may simply provide a comment of thanks without formally accepting it. Rather than waiting the 7 days and then spending additional time coming back to the Question to accept the answer (if it hasn't been already), I would like to have the ability to click on some type of conditional acceptance button immediately. If no additional answers are subsequently posted and no answer has been accepted then when the 7 days expires the conditional acceptance would turn into a formal acceptance automatically. If additional answers do get subsequently posted then the conditional acceptance could be automatically deleted (maybe with an e-mail notification to the power-user to re-visit the Question).

[Star Strider]

An alternative would be a ‘Conditional Acceptance’ flag, so that a Comment such as ‘Thank you very much! That completely solved my problem!’ would be flagged, then those with the appropriate privilege level could check the ‘Flagged Content’ and Accept any Answers with that flag after the seven-day embargo.

The problem would be in the frequent event of similar or overlapping but not identical Answers that were posted at about the same time, or two distinctly different Answers, that due to an ambiguity in the Question, would both seem to be equally ‘Acceptable’. It might be necessary to be able to Accept more than one of the appropriate Answers.

I believe that it should be possible for the OP to Accept more than one Answer, if the OP wants to. That could lead to more Accepted Answers, because it would not force the OP to decide between two or more equally Acceptable Answers, especially if the OP Accepted the ‘wrong’ Answer initially.

[Chen Lin]

Low "accepted rate" is a big concern for us. Nice to hear discussions on this topic. I like both ideas. James' approach automates the acceptance process, while Star Strider's approach leverages existing workflow. Hope other power users can share more thoughts.

Having more than one accepted Answer seems to be a neat idea. I think it's also beneficial to other members. Visitors tend to read ONLY accepted answer but other answers might also be very valuable. I would suggest we limit the maximum number of accepted answers (2,perhaps).

Here are some other ideas we got from power users and our team.

• The landing page would be a good place to draw newcomers' awareness to accept an answer.
• Notify OP to accept best answers by email
• Give Points to OP for accepting best answer as well as other activities such as comment or vote (should be a new point system, a little off-topic?)

[Image Analyst]

Sometimes users want to accept multiple answers but they are limited by whatever one they happened to click on first. How about if they click any "Accept" link that a popup box comes up with a prompt that asks "Which answer(s) solved your question?" or "Which answer best solved your question?" or "Which answer did you use?" and have checkboxes with the Answerers names and they can select one or more of those?

Also, some past power users seem to have drifted away. For example, look at some of the flat lines here and here. It might be good to reward people for being active by either computing reputation based on only the last 2 or 3 years, or else have some kind of weighting system where old answers are increasingly weighted less than more recent answers. That way new poser users could see their names rise to the top of the list faster and people who don't use it anymore would eventually drop off the list, and probably don't care if their name is on the list, and perhaps don't even deserve to be on the list anyway.

[Star Strider]

I agree with Image Analyst on the first paragraph of his latest Comment, but not on the second. Reputation earned should be reputation kept. People here have lives, and with their lives, have other interests and obligations. (For example, Matt Fig, who also authored several useful File Exchange contributions, is according the the ‘Contributors’ page, off earning his Ph.D. at one of the finest engineering schools in the U.S. I would not penalise his Reputation status for that, since he may return when his time permits.) Also, it seems that traffic on Answers has decreased in the past few months, so some may find fewer Questions, or a lower frequency of Questions, in their areas of expertise. We all have our particular areas of interest. I suspect people show up here when their time and motivation permits.

One certain way to discourage people from contributing would be to put a time filter on Reputation points. One certain way to keep talented people contributing would be to demonstrate that their efforts — even if their participation is sporadic — is rewarded with privileges appropriate to their permanent Reputation.

[Image Analyst]

Maybe users (or non-users) could just be hidden from the list if they went longer than 3 years without logging in, but their points remembered, so that the rest of us wouldn't see them. However if they logged in again, they could "renew" their points for another 3 years, and of course, they and all the rest of us would see them again. Just throwing out an idea to encourage newer users to "climb the list" and see their progress, while long-inactive users could still retain their spot on the list should they ever decide to participate in the community again.

[Walter Roberson]

There is a lot of student traffic, so traffic is partly cyclical based upon the typical school year. We should expect it to rise again in the next few weeks as assignments become due near the end of the late 2015 term.

My experience is that the overall traffic is rising. Over 200 questions a day is not unusual now. Those who choose to specialize in what they answer can do searches (but might miss something of interest to them), but the more generalist among us would tend to get overwhelmed . I know when I get overwhelmed here, I tend to stop trying to even read everything and tend to instead skip around and only check out and answer whatever might at first glance look interesting, in a much lower volume.

Just like how in Google searches most people don't read past the first page of results, I suspect that a lot of the volunteers do not read past the first page or two of "Most Recent". It gets to be too much.

I often answer questions before bed (and if I can't sleep, get up and answer more.) It is not uncommon that by the time I wander off, about 2/3 of the first "Most Recent" page are questions I have contributed to. When I wander back in the morning, 5-ish hours later, it is common that only about 3 of the first "Most Recent" page are questions I have worked on, reflecting updates from other people. It would probably be normal for most of the volunteers to see a completely different set of questions on the first page every time they visit.

As the number of students using the facility increases and as courses tend to teach much the same thing, the ratio of "interesting" questions for those with expertise in other areas will tend to decrease even if the absolute number of questions they might be interested in increases. Even if the distribution is like a sum of a few normal distributions (i.e., multimodal), when traffic doubles, for specialists it might go from "finding a needle in a haystack" to "finding two needles in a haystack twice as big."

One thing that I do find is that as traffic increases, so does the number of people who say "Give me code that implements this obscure paper!". I've got the "My Term Project Is Due On Tuesday" Blues. It is discouraging.

I am Closing an increasing number of questions on the grounds of No Question Was Asked. It seems a bit harsh sometimes, but with there being as much traffic as there is, the volunteers don't survive long without some Boundary Setting.

I do not have solutions here, but I think we are seeing burn-out, together with the more selective volunteers having difficulty finding questions of interest to them due to the volume.

[Star Strider]

I should have clarified my observation about traffic. The volume may have increased, but the character of the Questions seems to have shifted toward less substantive — or less frequently substantive — content, as Walter noted.

I go back through the last two or three pages every morning (covering the previous 12 hours or so), looking for Questions of interest to me and for those that have not been Answered. I keep up occasionally during the day, and check my Profile page to see if any follow-ups are needed.

[Star Strider]

Afterthought — It seems that a large number of ‘Questions’ of late could easily be resolved by the four-letter quasi-expletive used years ago (and still used in the gaming community): RTFM! = Read The (Friendly) Manual!

[Walter Roberson]

Not knowing where to look is a real thing. MATLAB is not a small language

[Chen Lin]

Walter is right. The traffic is cyclical. But if we trail 4 quarters(T4Q), we actually see a significant increase each quarter.

Answer 11

[Adam]

One or two things with respect to notification preferences and Activity Feed that I would like:

• Having a preference to not automoatically add Answers threads I comment or answer on to my Activity feed would be great. I click the unfollow (or whatever it is called) box literally every time I am adding a comment or answer. Just having a one time preference to have the default be off and then I can switch on as I wish would be nice.

• Being able to choose what I get e-mail notifications about. At the moment it seems I can just either switch them on or off so I switched them off because I don't want to get spammed about stuff from Matlab Answers, but I would like to have e-mail notifications on for a couple of File Exchange submissions that I follow. I tend to never use my Activity Feed (mostly for the reason above as I don't always remember to untick the follow option) so don't see updates on these toolboxes I am using.

[Guillaume]

That all seem reasonable, For your second point, the simplest would be to have a tick box next to each item in the expandable list that you get when you click on See emailed activity details.

[dpb]

Fully concur to both...

[TADA]

Before you follow a thread, there's a check box you can uncheck when you comment/answer in order not to follow that

[dpb]

Yeah, but ya' gotsa' do it every stinkin' time...which is patently annoying and rude can't set the preference.

Hence, like Adam, I pay no attention at all to activity feed 'cuz it's mostly/almost entirely noise don't care about and have the email notification 'off' as well.

[Adam]

@TADA - Yep, I click that (almost) every time I respond to a thread (and when I don't it's because I forgot), hence my request!

[Rik]

+1 I agree with the preference thing, and the point about granular email notification control.

It would also be nice to have a way of keeping track of unread threads. Sometimes a comment/thread gets drowned in the activity feed (for those of us who do use it).

[Walter Roberson]

When Activity Feed was put in place, there was a preference about whether you wanted to automatically subscribe or not . I cannot seem to find that preference now.

[Adam]

I looked for one straight away when it was introduced, and did actually mention it in feedback to the Devs when they e-mailed us editors back at that time, but I've only just got around to complaining about it in this thread now, after clicking Unfollow a few thousand times!

[Walter Roberson]

By the way, I use Activity Feed a lot. We are currently averaging over 175 new Questions every day, which does not count activity on older questions, so it is easy for updates to scroll off of the front page of 50... off of the first 4 pages even. I am following close to 15000 Questions since Following was introduced in October 2017, and I have apparently Answered more than 40000 historically, so it is not practical for me to keep mental track of which questions I have interacted in just by title...

Using contributor:roberson as my filter unfortunately also matches questions I have only editted...

[Adam]

My usual modus operandi on visiting Answers is to head straight to My Matlab Answers -> My Answers to check up on any that had activity either since I last checked or if it was my activity in which case obviously I know there is nothing new. Then I do the same for My Comments. Occasionally I check My Questions too, although not often if I haven't asked one in a while. Finally I hit Browse and look at the first page (occasionally, but rarely, the second) to see what is new.

What I would like to use my activity feed for is question threads that I find interesting for some reason, whether or not I have contributed. It gets cluttered so fast though if I forget to unfollow even a handful of threads I comment in or answer so I've never developed a habit of using Activity Feed as I'd like to.

I contribute to far fewer threads than you in general though so checking My answers and My Comments doesn't usually take long since my last login.

[Rik]

And that is precisely why a thread-view for the activity feed would be useful. That way you don't have to post a comment in order to get it in your My Comments list, but you avoid the level of detail that the activity feed provides. Personally I don't understand why the extremely detailed recent activity feed is exposed to users as well.

[dpb]

I use Adam's protocol--the browser link recall with my user name retrieves each; then the word "sort" retrieve link to the sort I once did on unanswered by most recent. I skim the first page of it and that's it...

I've used the activity feed on a couple of occasions when I was curious who might have voted for an answer... :)

What is difficult as Walter points out is that once have answered as many as even I, a search limited to your own Answers can't seem to find ones you remember you wrote on a given subject that would be useful to link to...so they still go into oblivion.

The fact TMW still hasn't done anything about the waste of space so that one has to page and page and page to get to anything beyond the first 50 means I just don't...

[Guillaume]

That's interesting. I used to have the same workflow (check My Answers and My Comments) but when the activity feed was introduced I ditched that and now use the activity feed exclusively. Yes, I wish it was more compact (in particular, collapse multiple consecutive edits into just one change) but for me it works well enough.

The fact TMW still hasn't done anything about the waste of space so that one has to page and page and page to get to anything beyond the first 50 means I just don't.

Mathworks has to go through a fairly rigid process to make change to the website so it takes time. But they are listening to us.

With regards to searching on Answers, yes it's pretty much useless, it can never find what I'm looking for. Google is actually more often useful than mathwork's own search.

[Walter Roberson]

dpb:

On the main Answers web page, the "Recently Added" view, to the right of "Sort by", there is now "View:" followed by two icons. Click the right hand one to get the view that uses less space.

[Walter Roberson]

I use mostly Activity Feed. However, because it shows so much (every different Edit shows up for example) then afterwards I end up going to My Answers and My Comments to see if I have missed anything. I also sometimes search

 contributor:roberson -asked_at:30

so that I can catch updates to old questions that I might have missed... especially for ones that are from before October 2017 when the subscription service went live.

[Adam Danz]

Very interesting to read other people' workflows.

@Walter Roberson, I've been searching for search-options such as the ones you use in the comment above. I'm aware of contributor: tag: question: answer: comment: but I'm curious what other options are available. What is -asked_at:30?

[Walter Roberson]

https://www.mathworks.com/matlabcentral/answers/help?s_tid=gn_mlc_ans_hlp

[Adam Danz]

Perfect, forgot about that page. Thanks, Walter.

Answer 12

[dpb]

Seems excessively strict given that the "extra-x" format is now default for all MS Office products; what's the reason isn't allowed--just never added to the "allowed" list, I presume? Just had a Q? wherein the poster ran into the problem; renamed the file but didn't change the format and caused issues because of incompatible format in Excel.

Run into when try to attach any .xlsx file--error box contains following text (I didn't bother to capture image, just the text):

Cannot attach this file because:

    File format is unsupported. Use any of these formats: .bmp, .csv, .fig, .gif, .jpg, .jpeg,
      .m, .mlx, .mat, .mdl, .pdf, .png, .txt, .xls, .zip

Attach a file:
Choose file.xlsx

[Rena Berman]

(Answers dev) Hi dpb, Thanks for bringing this to our attention.

[Rena Berman]

(Answers dev) You should be able to attach .xlsx files now.

Answer 13

[Greig]

A "Follow Question" button.

How many time have folk encountered a question that is useful or interesting and they won't too keep track of it, but don't necessarily have a comment or answer to contribute?

Basically a mechanism that allows you to follow a question, without having to answer it or comment on it, or bookmark it in your browser. An extra option, called "My Following", in the "My MATLAB Answers" menu would be an obvious place to link to a list of threads that you are following. Obviously, an "Unfollow" option would be needed too.

[Walter Roberson]

Should following generate email notice of updates?

[Greig]

I would say that having the option to get email alerts would be useful.

[Walter Roberson]

A couple of months ago, Mathworks implemented "Activity Feed". If you look under "My MATLAB Answers", it is the first entry.

Notification is only for Questions that you are "subscribed" to. There is a configuration option to determine whether by default you want to subscribe to any question that you ask or answer or comment on. You can always toggle any given question on or off -- including toggling notifications on for a Question you have not responded to at all.

The subscriptions are not back-dated for anything you answered or commented on in the past, but it looks like they might have been turned on by default for any Question you asked.

Answer 14

[Walter Roberson]

If there were a search tag based upon who last modified a post, then I would be able to search

contributor:roberson asked_at:7 -last_updated_by:roberson

to select postings in I had contributed but I was not the most recent contributor. Especially if this was combined with my proposed sortby:newest I could be filtering down to things that have been recently updated by someone other than me.

Answer 15

[Image Analyst]

Far too often the original poster posts a new "Answer" to a question posed by someone in a Answer or Comment, instead of replying to that specific Answer or Comment where the question was posed. I suggest that the "Answer this question" caption be changed to mention the original poster, like "Answer the original question by Walter Roberson". Hopefully that will lessen the number of times people answer their own question when they should be replying to someone else.

I'd also like to see "Comment" be changed to "Reply" to be consistent with the wording used in almost all other online forums.

Answer 16

[Chad Greene]

For new users, it might help to force askers to answer yes to these questions before their question can be posted:

• Did you describe your problem in detail? {{Include a help bubble that error messages should be included in the question, etc.}}
• Have you provided a minimal working example? {{Include a help bubble describing an MWE for those who are unfamiliar with the concept.}}
• Have you formatted your code for readability? {{Include a help bubble describing how to format code.}}

I don't think it would be necessary to make askers say yes to these questions forever, but it may get rules of etiquette and effectiveness into their heads if they have to think about their questions the first few times posting on the forum.

Answer 17

[Cedric Wannaz]

For people who don't check the "Allow users to contact me securely" option in their profile, there could/ should be a button that allows us to send them an automated message "there has been some activity on your thread" (and e.g. a field or drop down for us to enter/select a thread ID).

The rational is that if someone wants time from us, it should be acknowledged/imposed that he/she accepts to receive reminders about open threads or just about added comments. This may help reducing the amount of questions for which the person asking never comes back (if only for saying that he/she doesn't need an answer anymore).

Some option for sending electric shocks remotely would be nice too..

[Rik]

Unless contributors have manually disabled email alerts, your wish has been fulfilled.

[Cedric Wannaz]

Awesome! .. but where do I click for sending electric shocks? ;-)

Answer 18

[Adam]

Hopefully this is still work in progress, but the options for asking a question are a little odd at the moment. I notice that 'Release' has been added as a field which is great, but it is disabled until you enter a Matlab Product, which itself is not compulsory and is rarely filled in by people. The ability to ask the question is enabled straight away, without tags, product or release so why is release disabled until you enter a product? It just means that one of the most useful pieces of information for a question is very unlikely to be filled in most of the time.

[Image Analyst]

I wish the release were available as a number. Currently when we do this

MATLAB_Version = ver

we get this:

If we want the version as a number, to see whether we can run a newer function or not, we have to do this:

MATLAB_Version = ver('MATLAB');
versionNumber = str2double(MATLAB_Version.Version);
if versionNumber > 9.0  % or whatever....
etc.

[Rik]

I use an equally convoluted method:

function v=getVersion
persistent cur_v
if isempty(cur_v)
    cur_v=version;cur_v=str2double(cur_v(1:3));
end
v=cur_v;
end

The problem with including a function now, is that you'll have to wrap it in a try-catch block.

[Adam]

I wouldn't have a clue what it meant though if Matlab Answers started showing version as 8.2 and I wouldn't waste my time trying to work out which release that equates to, I'd just ignore it, especially since every toolbox has its own version numbers that are different for a single release.

[Image Analyst]

Sorry, I thought you were talking about the MATLAB release in MATLAB, rather than the MATLAB Release when posting a question to the Answers forum.

In the past few days we've talked with the MATLAB Central web developers about the release thing, including my peeve, the existing release number vanishing if I add a toolbox to the product list. Hopefully these are on their to-do list.

[Steven Lord]

Image Analyst, I know this is a bit of an aside since the original question dealt with specifying release when asking a question on Answers, but the version number of a product is not necessarily representable as one number. For example, in release R2018a Global Optimization Toolbox has version number 3.4.4. Concatenating the individual digits is not a good alternative, as products with versions 11.1, 1.11, and 1.1.1 would all be represented as 111.

If you want to determine whether the version of a toolbox you have installed is older than the version specified by a version number character array (such as that returned by ver) use the verLessThan function. It knows how to handle version numbers like '3.4.4' or '7.0.1' (as shown in the help text.)

[Walter Roberson]

Side note: I created a Question a couple of days ago. The Tag was mandatory.

[Image Analyst]

Steven, I didn't know about that function. It could be made useful. To be useful it would also have to take Ryear strings. Why can't it? We mostly deal with the Ryear release strings, not the actual numeric version numbers. For example I might know I have R2018a but have no idea if it's version 9.3 or 9.4 or 9.5.1.2 or whatever. Why can't the version input from that verLessThan() function take release strings? Then it might actually be convenient and useful.

if verLessThan('matlab','R2014b')   % instead of verLessThan('matlab','8.4')
    % -- Code to run in MATLAB R2014a and earlier here --
    s2.EraseMode = 'xor';
else
    % -- Code to run in MATLAB R2014b and later here --
    s2.FaceAlpha = .25;
end

[Steven Lord]

I've added that request into the enhancement database with a link to this discussion.

[Anusha Sridharan]

(Answers dev) Thank you for the feedback. The reason that release is not enabled until a product is selected is because, there are many products that do not exist in certain releases of MATLAB(for instance Text Analytics Toolbox was only introduced in R2017b). Selecting a product or (multiple products) enables us to filter the release drop-down with only the releases that all the selected products exist/existed in. This helps maintain valid combination of products and release for each question.

[Adam]

What percentage of questions actually include a product though? It seems to me very rare that people bother to include the product (all the more so if we exclude the cases where the product is just 'Matlab', in which case no filtering would happen anyway). For example, as a very small sample, of the 50 questions on the front page as I type this, just 4 include a Product field other than Matlab and another 7 have the Product as Matlab.

Given how useful the information given by the release version is this doesn't seem to be a worthwhile trade-off for 8% of questions (albeit a very small sample, as mentioned!).

Does the release need to be filtered? Surely people will select the correct release anyway if they select it at all (or it could default to the latest, which is presumably valid for all toolboxes).

Either that or force people to include the product too before being able to ask a question would be good.

[Anusha Sridharan]

(Answers dev)

Thanks again for the feedback.

This is a new feature and we will consider updating it as we continue to get input from the community. Based on your suggestion, we will explore the possibilities of making Product a required field.

With respect to the release drop-down being filtered, we do think this is important as it provides users with a smaller, pre-filtered list that is easier to use.

Thank you again for taking the time to give us feedback on this feature.

[Walter Roberson]

Most of the time I can deduce the Product(s) involved based upon the question and function names mentioned. Not always, as there are some products I seldom look at, and also details like which exact SimEvent specialization (such Simscape Multibody) is not always clear.

It is a lot more useful to know the user's MATLAB release.

Instead of filtering release based upon product, filter product based upon release.

[Image Analyst]

But don't throw away the release like you do now. For example if someone said they're using MATLAB R2016b and I decide to add a toolbox to the product list, now the release vanishes! Even if that toolbox existed in that release. So now if I add a toolbox, I have to remember what the release is so I can add it back in. If we moderators forgot the release and just added the toolbox, now the release info is gone. Obviously not a good situation.

[Anusha Sridharan]

Thank you everyone!

We appreciate all the great feedback around the release field and will consider these suggestions when we revisit this feature.

[Walter Roberson]

Is it worth having "Full Suite" as a separate product when it is at best an obscure name for a campus-wide license (referenced that way only once that I can see) ?.

I see there is also "Standard Suite". It looks to me as if "Standard Suite" and "Full Suite" are the two Student bundles, rather than products.

[Adam]

I was reminded of this discussion again today on seeing this question. Most of the time people still simply don't include version information in any form, but when someone here is explicitly adding it into his question text it is surely a clear sign that the method for including this information via the dropdown lists is fundamentally not working.

I really hope this is being strongly considered for redesign because version information is so fundamental it should be in every question and should be so obvious how to do it (or compulsory) when asking a question that even a total newbie to the forum can easily add the information.

To take another very quick sample, 6 months on, at a rough count ~15% of the first 100 questions currently on the first 2 pages of questions have a Product entered that isn't Matlab and 21% include Matlab as the only product. It is maybe better than it was, but still not good for the use case of having to enter the product before it suggests valid releases (given that, as mentioned earlier, every release is valid when the product is just 'Matlab').

Answer 19

[Rik]

Since the new editor has become way more fancy:

A while back I read in some help page for the Answers forum that a referrence to a function should be the name in monospace with a hyperlink to the doc page. It would be helpful to have this automated via a button in the 'insert' part of the toolbar (complete with the doc-page search bar).

Also, are there plans to try to expose the text of comments/answers to the browser spell checker? I rely on it embarassingly often, so it would be helpful to still have it. I just tried if it was specific to Chrome on W10, but Firefox has the same issue (or feature?)

[dpb]

+1 on both...particularly for some of our non-native English speakers, it might help significantly in clarity of posing their Q?.

[Walter Roberson]

I have noticed that Firefox on desktop and Android do not recognize the input area as being text areas, which is probably why the spell checker is not active.

[Guillaume]

I have already complained on the answers-editor mailing list (not sure you're on it Rik) that hyperlinking to the doc is now a lot more complicated than it used to be. Certainly, now I don't bother making the hyperlink monospace. Too many keypresses required. I will keep on complaining until it's fixed.

[Rik]

@Guillaume, I missed your comment, but no, I'm not on that mailing list (yet?), so I'll confine my complaints to this page ;)

[Walter Roberson]

II noticed that on desktop that if you type <> around a plain url then it is converted to a link, but that the old syntax of <url text> is not supported . Neither one appears to work with mobile .

[Guillaume]

I don't know how I missed it the first time I looked at it, but in the help page there is a list of keyboard shortcuts. It list 3 different methods of inserting urls. The old <url text> is now [text](url).

This voids my main issue with inserting urls with the new editor. I just have to get used to the new shortcuts.

I still think that the url field in the CTRL+K dialog box is much too short.

I also think that the pop-up that shows the target when you hover a url in a post is much too narrow, wrapping a long url over many lines, making it much harder to decipher.

[Rik]

Por-tip: if you want to use a copy-pasted comment, you need to finish the keyboard shortcuts yourself. So pasting in a full [label](url) will not work, but if you only paste in [label](url and then type the close paren it does work.

Or maybe it is just time to make an elaborate autohotkey file...

[dpb]

The old <url text> is now [text](url).

I fail to understand why there was any need to do that.

And why both "underscore-text-underscore" and "star-text-star" yield italicized text instead of italicized and bold, respectively, is hard to follow reasoning behind???

Answer 20

[Walter Roberson]

When making a response (Answer or Comment) to a topic originated by Mathworks Support, the browser will almost always time out, with the "Saving" notification going back to "Submit" as if the response had not been saved. This leads people to click Submit again, often several times. Usually the response does get saved, so we end up with one copy of the response for each time the person clicked Submit, but at least once for me the response did not get saved.

My work-around now is to select my response and Copy it in my browser and then refresh the page to see if the response did get posted, and if not then Paste and try again.

[Image Analyst]

You can use Lazarus https://addons.mozilla.org/en-US/firefox/addon/lazarus-form-recovery/ to avoid copying to the clipboard. Simply right click, select Recover Text, and pick the recent text that you typed in. It stores around 20 prior texts you entered and you can just pick one and have it enter it again. It's like it automatically saves your message after every keystroke to it's own private collection of clipboards.

[Walter Roberson]

"Lazarus Forms Recovery could not be verified for Firefox. Proceed with caution"

(and I do not get a Lazarus menu item when I right-click.)

It worked for me in the past, but I stopped paying attention to it partly because Lazarus has never worked with Google+. And now I can't bring up the manufacturer's page.

[Star Strider]

I use Firefox for pretty much everything. ‘Lazarus’ installed without problems (or notifications from Norton) on Firefox 40.0.2 (the latest version) earlier this morning. I’ve not used it yet.

@Image Analyst, thanks for the link!

[Walter Roberson]

Firefox evolved into aa security model that does not support Lazarus unfortunately .

Answer 21

[Adam]

This seems to have been something that changed in the last few days. Why has the 'Home' link on Matlab Answers suddenly changed to take me to a page I find useless?

For the last year I have been here I have clicked on 'Home' as my standard way of returning to the full unfiltered list of all questions. I browse this sometimes to answer one and sometimes because the question is of interest and I want to read someone's answer.

Now this link takes me to a page where I have to click to 'Ask a Question', 'Answer a Question', read 'Top Questions' or 'Try Answering These'. This is very annoying.

It does at least seem that when I clock 'Go' next to 'Answer a Question' that this does now take me to the unfiltered list (yesterday it was taking me to a filtered list of only unanswered questions), but I'd still prefer that 'Home' takes me to the Matlab Answers page of all questions as it intuitively should.

[Walter Roberson]

There was an update within the last couple of days designed to make it easier for new users to find information.

If you go directly to the old location (such as by bookmark) then you will continue to have the same interface as before. The location is http://www.mathworks.com/matlabcentral/answers or as appropriate for your country. The new interface has a /index on the end of that.

[Adam]

Yeah, I do have the main page bookmarked so that does still work, I guess I just have to get used to going to my bookmarks instead of 'Home' now.

[Star Strider]

From ‘Home’, in the upper left, click on ‘My MATLAB Answers’, then right-click on ‘My Answers’ then ‘Open Link in New Tab’ to create a new tab with all your Answers in the ‘old’ format. Then bookmark that.

[Adam]

The problem is that is just answers I have given. I use those links a lot to keep track of any replies to questions I have given an answer to or commented on, but to do something as simple as go to the full unfiltered list of questions it seems there is now no single-click link other than our own bookmarks.

The 'Answer a Question' link at the top is one that filters to only unanswered ones so that is probably what I was clicking yesterday when I ended up there too.

[Star Strider]

I certainly agree that ‘Answers’ now is definitely worse than it was a year ago, before TMW started ‘improving’ it. I don’t understand the reason for all the changes. It wasn’t ‘broken’, so no ‘fixes’ were necessary.

[dpb]

IMO it was broke, just "less broke"... :) (or :( more realistically, I guess).

[Walter Roberson]

I use firefox. I drag the URLs of my most commonly accessed pages into the toolbar, so I can reach them with a single click. I go into the Bookmarks menu, find the Bookmarks Toolbar section, find the appropriate URL, edit the Name section to make it something short. For example I have

UK Tag FAQ up mark Ask floop

where the first takes me to the UK Answers page, the second to the Guide to Tags, the third to the FAQ, the next to the Answer about formatting with markup, the next to the Answer about how to ask a good question, the next to the FAQ about looping over files, and so on. The UK one I click on to start my session; the other ones I drag into an Answer or Comment to copy in the appropriate link.

[dpb]

Good idea to mark the FAQ's, etc., ... I've been too lazy on the front end but should.

[Chen Lin]

Thanks for all the great feedback. I’m the new Community Product Marketing Manager, and appreciate your insight into new features and designs. Hoping I can shed more light on the “why” behind the new landing page.

The new landing page is designed to orient new and casual users to Answers. Historically, we saw a very high abandonment rate by newer users clicking on the “home” button. When we examined other Q&A forums in the industry, it appeared that other websites were friendly in introducing the forum, and welcoming new visitors. The page was designed with this goal in mind.

We’re hoping the question autosuggest feature will also cut down on duplicate questions and that the ‘Get Started’ tips will help novice users ask better questions.

The URL for past home page did not and will not change; we did this so any bookmarks and workflows would remain intact.

Thanks again for your thoughts. We plan to add some features in future releases that we think will be of interest to you. Please continue to share, either publicly or my messaging me directly.

Chen Lin