Community Highlights

설문

How much time should be spent searching for a solution before asking a question in the forum?



Jason Grubish
Jason Grubish 2022년 8월 9일 17:47 (2022년 8월 9일 17:47에 수정됨)
Do folks search their questions on Mathworks? I always end up using google to search my questions and expect to find my results showing MATLAB Central posts. Maybe I should just cut out the middle man and search here directly.
Adam Danz
Adam Danz 2022년 8월 9일 19:36 (2022년 8월 9일 19:36에 수정됨)
Adam Danz
Adam Danz 2022년 8월 9일 18:21 (2022년 8월 9일 18:23에 수정됨)
Good question (maybe this should be a poll)!
I use a custom search engine along with MATLAB Central search directives to search the Answers forum. For example, I start by typing "mlans" in my Chrome address bar. This custom key word tells chrome to search the Answers forum. Then I continue typing my query.
I also have one for the File Exchange.
The next most frequent method I use is google, followed by searching directly from MATLAB Central.
Rik
Rik 2022년 8월 9일 18:56
I tend to give up from the get-go. Mathworks is not a search engine builder, and judging by the results I have when I try, it is a non-trivial task. The problem is that finding what I mean is often similar (but not quite) to what I write in the search box.
It is nice to have the directives for when I know who wrote the question, but the search is often not fuzzy enough for me. I don't begrudge anyone who's working on the search engine, because everybody has gotten used to the excellent results you get from the big boys (at least once you scroll past the advertised results).
I have resorted to keeping a list of links for things that are hard to find and/or I use regularly (but not often enough to bookmark it).
Jason Grubish
Jason Grubish 2022년 8월 9일 18:53
Wow this is a great idea! I always end up using the Chrome address bar to search for matlab topics and most of the time end up typing "matlab" first anyways. Just created a custom search for that use case, thanks!
GeeTwo
GeeTwo 2022년 8월 2일 1:42
I usually look for a few hours before starting a question. As I build up the question, I continue to search and try things. After a few days without resolution I'm ready to post. I haven't got there yet.
Most recently, my question was "How can I merge nested cell row vectors of cell row vectors*?". I wanted to build a 1xn (or nx1) list of solutions (in this case strings) from a recursive function where each call may itself return a list, and the calling function modifies the recursive returns (in this case by prepending strings). Turns out the answer is [nested_vector{:}]. That expression makes NO sense to me, but it does close enough to what I need. It de-nests exactly one deep. If the cell vector only has non-cells, it tries to merge their contents. In the case of a row vector of strings, it makes one big string with no spaces.
  • * (column vectors of column vectors would have also worked)
snool
snool 2022년 8월 1일 9:29
There is an error in the poll.
The question is "How much time SHOULD be spent..." and the last answer is "Infinte. I'll never ask a question".
After reading the last answer people might think it is about what you actually do and not what you should do.
(Also the time span between a few minutes and hours is probably too big. I think the right answer is somewhere between these two answers.)
dpb
dpb 2022년 7월 31일 19:29
I agree w/ @Walter Roberson that I find the searching facility to be of little help -- I've got a bunch of things I know I've posted over the years and while I'm not even close to the volume Walter has, I never think to create a database of "things that might be useful again" so I rarely can find the post of interest even when I know it was mine.
The paucity of results per page that I've complained about since the introduction doesn't help; it just takes too long to page down page after page to make that feasilble --and the Q? titles, since are posted by the original poster are, as often as not, not even related to the real subject and remembering how it was posted is impossible.
It's the latter I think that limits the usefulness of the search engine as compared to googlesearch as much as anything, I think...it has a lot better chance the title subject matches the content and has a much larger viable population from which to choose. I've posited also since almost the beginning there needs to be more structure/restrictions placed on the title/subject.
Similar limitations hold for tags...the user doesn't have a clue most time -- I have a hard time coming up with those even for my own Q? despite 30+ year w/ MATLAB to make them really pertinent.
Erika McPhee-Shaw
Erika McPhee-Shaw 2022년 7월 18일 16:32
Have a strong feeling the answer is skewed by gender. BTW I am female.
Adam Danz
Adam Danz 2022년 7월 18일 20:46
Interesting postulate!
I searched for some info on this and read the first 2 research papers I came across (limited time to lunch break).
A 2002 paper looked at male/female participation in forums for online education [1]. They first polled female participants asking how often they think they post questions or comments to these forums relative to men and the distibution of responses were skewed toward "the same or less". Then they measured the actual paritipation rates and found no statistical difference between men/women.
But that's not specific to programming so I searched for another article, this one (2019) about participation in Stack Overflow which found that women ask more questions than men but men answer more questions [2].
I'd be interested in what you may find!
[1] Mazzolini, M., & Maddison, S. (2002, December). Does gender influence discussion forum participation in online education?. In ASCILITE (pp. 421-430).
[2] May, A., Wachs, J., & Hannák, A. (2019). Gender differences in participation and reward on Stack Overflow. Empirical Software Engineering, 24(4), 1997-2019.
Erika McPhee-Shaw
Erika McPhee-Shaw 2022년 7월 20일 22:54
So cool you did some research over lunch break! Much appreciated! I guess I did have in mind posting questions and answers in forums related to programs related to programming more than to online ed, and it's interesting you found research about both. I'm not too surprised by the findings from the research on gender and participation in StackOverflow, just based on listening in on coversations between my students who use that forum. Thanks again!
Rik
Rik 2022년 7월 19일 6:21
I always enjoy some lunchbreak-science. I don't know whether I would have expected any differences, but glad someone did the actual research.
Walter Roberson
Walter Roberson 2022년 4월 14일

There needs to be an increment between "a few minutes" compared to "hours". for example 15 to 30 minutes.

Matt J
Matt J 2022년 4월 15일 (2022년 4월 15일에 수정됨)

I'm actually pretty surprised that such a large fraction of people spend more than a few minutes. I would have thought the most significant choices would have been,

  • Don't search just ask.
  • Check only the automatic suggestions made by the 'Ask A Question' dialog.
  • 10 minutes or less
Adam Danz
Adam Danz 2022년 4월 15일

Considering that view counts far exceed the number of new questions per day, I think there are many users who find solutions in existing content without needing to ask a question. If the poll only considered responses from those who have asked a question or multiple questions, I'd expect the distribution to slide toward less search time.

Katherine May
Katherine May 2022년 4월 28일
Yep, this is exactly how I answer most of my questions. Even if the asked question isn't exactly the same as mine, the answers to them can still be helpful. It can take a couple of Google searches to get my search terms right, though.
Matt J
Matt J 2022년 4월 13일

An important factor is how well the Answers forum search engine works. For some reason, I find it very hard to locate posts on a particular question, even when I know they exist. If it worked as well as google search, I think that the auto-suggestions made when a question is being composed would obviate the whole issue.

GeeTwo
GeeTwo 2022년 8월 2일 2:11 (2022년 8월 2일 2:14에 수정됨)
I call this the "You Bet Your Life" effect. (I don't know if Jay Leno recreated this part, I got it from watching reruns of Groucho Marx.) That is, you need to say the secret password to get the special prize. The worst offender (of public documentation) I have encountered is ORACLE. To remove tables/relations from your workspace or columns from tables, you use 'drop'. To remove rows from a table/relation you use "delete". If you search delete you will never figure out how to remove a table or column, because all of the links tell you how to manipulate rows. If you search drop you will never figure out how to remove a row, beca.use all of the links tell you how to manipulate tables and columns (and indices IIRC). And searching 'remove' doesn't help for any of these cases. MATLAB has a lesser version if this malady; they don't seem to have specifically worked at being opaque like ORACLE, but they also don't seem to have done much to mitigate the effect. It would help if there were links at the bottom of the document to things that do something similar in natural language, even if they operate on a totally different data type. These synonyms would be even more useful with the data types they work on being associated with them.
DGM
DGM 2022년 8월 2일 17:09
I've also noticed that it seems like the related links at the bottom of synopses and webdocs tend to not refer across toolbox boundaries. I suppose I could see why that might happen, but it's just another obstacle when you're searching for something that you already know exists.
goc3
goc3 2022년 4월 13일

I agree 100%. Even if I recall the name of a user that posted a comment or specific verbiage from a post, I often can't find it searching a browser history. Beyond the difference in search capabilities, it may have something to do with what text from each page is searchable and/or available in meta-data.

Walter Roberson
Walter Roberson 2022년 4월 28일

The contributor: directive can help a fair bit.

On the other hand, I have made enough contributions that even with contributor: it can be difficult for me to find something that I know that I personally wrote.

DGM
DGM 2022년 4월 12일 (2022년 4월 12일에 수정됨)

I find this distribution difficult to believe; then again, maybe people who can't be bothered to click on the answers that appear as they create a new question also can't be bothered to click on a poll.

Adam Danz
Adam Danz 2022년 4월 12일

I have a feeling that the majority of visitors to the forum find a solution without needing to ask a question. The view counts per day far exceed the number of new questions per day.

But, yeah..., these polls aren't pure random samples from the community.

Kevin Holly
Kevin Holly 2022년 4월 28일
Yes, I was one of those people. I lurked the MATLAB Answer forums for years. *looks at questions asked on old and new account... At least I answer questions now. :D
DGM
DGM 2022년 4월 12일

That's a good point.

goc3
goc3 2022년 4월 12일

You make a good point. Though, I would say give the poll time. There are only 19 responses so far...

DGM
DGM 2022년 4월 12일

Haha. I didn't even notice the total count. I guess I can't say too much if I'm not paying attention myself.

goc3
goc3 2022년 4월 12일

That's a loaded question. I wonder if people will answer according to the question ("should") or what they actually do.

Adam Danz
Adam Danz 2022년 4월 12일

I often find that it takes days of thinking about a problem before I'm even ready to phrase the question.

Bjorn Gustavsson
Bjorn Gustavsson 2022년 4월 12일

But then you also have had "not too easy to search" for the right question already answered either/to?

Rik
Rik 2022년 4월 12일

For me the two are the same, but I'm aware I am very likely to be an outlier.

It also depends on the complexity of the issue. I'm currently working on a project where I could ask a question, but I'm not sure I should, and I'm not sure how to phrase it in a way that I'm likely to get help. In that specific case, I've been working on it for approximately 100 hours already.

goc3
goc3 2022년 4월 12일

I often run into the same problems, Rik. Sometimes it is due to proprietary projects, which should not be discussed publicly. Other times, the project is so complex that getting the average reader up to speed so that they could understand the problem at hand is not worth the time and effort.

More Polls

Check out the latest community polls.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!