MozillaZine

First Survey Test

Saturday June 5th, 1999

In preparation for our upcoming vote in the Mozilla Throbber Animation contest, we are testing our new survey code. The survey (and the contest vote) are open to mozillaZine members only, to limit "ballot-box stuffing".

In our first test survey, we ask you to rate some options for our talkback forums. Please be sure to choose a rating for all of the options - your submission won't be accepted otherwise. You will notice some incongruity in the labeling of the choices - this code was initially designed as a survey for "wishes" for Mozilla, and we haven't finished the move to a generic survey yet. Changes will be made before the next test. We will be performing another test of this code in a few days.

Important Note: I neglected to mention it, but the ranking is on a scale from 1 to 20, not 1 to 10. Please take note of this when you vote.

To get started, click here. To view the results, click here. Send any bugs to this address.

Update: The survey is back online - the code to disable multiple postings had been commented out! Sheesh. That made the results of the survey invalid. So, I've recreated the survey, and you can vote again, and hopefully this time it'll work.


#1 Re:First Survey Test

by Anon

Saturday June 5th, 1999 11:58 AM

Reply to this message

Why not use something similar to CodeGuru <http://www.codeguru.com/> There discussion board has several features that you may find benificial. For example it only displays headers but it also limits the number of headers per page to about 25, and just splits the fourm across multiple pages. It also automaticly sorts the headers so that the most recent post appear first (including old messages with new responces). And since they are also an open source community they'll proably give you the code for free if you want it.

#2 Re:First Survey Test

by Anon

Saturday June 5th, 1999 12:35 PM

Reply to this message

I dislike strongly all the proposed changes. I like mozillazine as it is. It is still small enough to not need fancy moderation code; it rarely even needs to display less than all the messages. Please do not change anything. I have been reading mozillazine since the week it was first created (when it was orange and red) and have enjoyed it a lot.

One of the reasons I got tired of slashdot was because all the fancy moderation and forum code kind of destroyed the geek atmosphere and made the site feel more like a commercial portal type thing.

#3 Re:First Survey Test

by Anon

Saturday June 5th, 1999 1:52 PM

Reply to this message

Agreed, clean and simple is best. I can have a quick read through MozillaZine without worrying about some whizzy interface. Don't touch a thing.

#4 Re:First Survey Test

by Kovu <Kovu401@netscape.net>

Saturday June 5th, 1999 3:58 PM

Reply to this message

With threads you can see like 20 responses on your screen at a time, and only spend time d/ling those you haven't read before or want to re-read.

Most messages in threaded conversations contain a subject only and then "n/t" to indicate the subject is the reply and that there is "no text" in the message except "n/t". This encourages short replies, and if you want to rant you can do it and people only need listen to it if they want to by clicking on it. 100 replies always takes the same amount of time to d/l because only the subject is on the page.

Here is a really good thread structure: <http://www.seanet.com/~vb…oard/general/board1.shtml>

the main site is here <http://www.vbdesigns.com/demise/frames.htm> you would have to click on Say Something and then General Discussion to get there (frames site)

Once the message list gets a certain length begin a new page and then put hyperlinks up top to link to the older ones. Mozillazine could have one of these for each topic, we would only have to listen to replies when we wanted to once with an option to re-read at will.

Also, when you are constantly having to scroll to the bottom of a really long page, and the Home link is always at the top, that is a little incovenient--a newest first to oldest last structure would keep the latest message at the most accessible place of the screen, right under the topic, which everyone is thus reminded of constantly, encouraging people to stay on it. :) Also, the home link is right there, and so are the links to older pages of responses, which can still be responded to.

Look at that page link and see how quickly you a) get to the newest messages and B) scroll through 100 of them choosing which, if any, to look at. The admin remarks are obvious by their bright colors, which helps pick those out. Here, tell me how often you can see more than four responses at any one time? And you have to skip over all the old ones to get to the new ones.

I would combine this insert message window with the subject window at the top, tighten them up quite a bit, make them as tiny as possible (maybe just make this window a horizontal toolbar at the bottom of that window above), include the links to old pages, and then have the messages oldest last, minimizing the necessity for scrolling over old messages.

Just my $.20

#5 Re:First Survey Test

by Anon

Saturday June 5th, 1999 4:00 PM

Reply to this message

I like that invisible thing Mozadmin! That's cool!

#6 Re:First Survey Test

by Anon

Saturday June 5th, 1999 4:20 PM

Reply to this message

You should move the username and login up to the top of the form as I filled it out and realized right at the end that I needed a user account to submit.

#7 Re:First Survey Test

by Anon

Saturday June 5th, 1999 4:49 PM

Reply to this message

When you ask a question in the polls, take care to phrase it properly.

You asked a question, then said 'or whould we leave it the way it is', at the end.

This makes it ambiguous for us voters who can only vote for 'approval' or not.

#8 Re:First Survey Test

by Theran

Saturday June 5th, 1999 5:45 PM

Reply to this message

There should be fewer options in the list for surveys. I think that 1-5 would be a lot easier to deal with. I also agree that the questions should be more carefully phrased.

#9 Re:First Survey Test

by Anon

Saturday June 5th, 1999 6:31 PM

Reply to this message

Yes, simple is definitely best. I just want to click a single like to see all the talkback responses.

#10 Re:First Survey Test

by Anon

Saturday June 5th, 1999 6:31 PM

Reply to this message

like=link ;)

#11 Re:First Survey Test

by arielb

Sunday June 6th, 1999 4:13 AM

Reply to this message

I voted for threading as extremely essential. It's very hard to respond to someone if there are 20 messages in between. It is hard to follow other responses. In short it's a mess right now. Personally I like most of the slashdot model where you see the root responses in full along with the tree. I didn't like kovu's example because when I post something for the first time, I have to check everyone else's posts to make sure they didn't say the same thing. It's a nice compromise between flat mode and 100% thread tree.

Moderation? I just don't see a need for it. I've seen some pretty dumb posts in the polls section but that's probably because the polls are pretty stale.

I'd also like the use of cookies so that I don't have to type my name and password every time I post something. Slashdot and My Netscape don't have this problem

#12 Re:First Survey Test

by sdm

Sunday June 6th, 1999 7:48 AM

Reply to this message

Here's what I'd like: I'd like to log in, and see all the new messages in all the posts, since the last time I logged in. There are still active discussions on the posts that are old (in the sidebar), and I don't want to manually check all of them every time I visit mozillazine. I also don't want to have to remember how many messages there were on the newer posts the last time I visited, and figure out myself whether there are new messages or not.

#13 Re:First Survey Test

by SomeSmartAss

Sunday June 6th, 1999 10:17 AM

Reply to this message

If your going to have "Anon" posters, can we still se a name too (i.e put "user name (anon)") that way you can still have a small clue as to whose posting, even if they don't feel like creating a member account...

#14 Re:First Survey Test

by SomeSmartAss

Sunday June 6th, 1999 10:38 AM

Reply to this message

Oh! and one more thing... Instead of a 1 to 20 range, put a -10 to 10 range which, logically, makes more sense. I'm the type of guy who looks at results first, and the range made it look like there was only a posibility of "agreeing" with the proposal...

you could make the result bars two way with a display code as follows

<?php if ($result < 0) {

echo "<td align=right>" . (round($result*100))/100 . " <img src=\"wishlist_neglcap.gif\" width=\"" . ($result * 10) . "/"></td><td> </td>";

} else {

echo "<td> </td><td align=left><img src=\"wishlist_neglcap.gif\" width=\"" . ($result * 10) . "/"> " . (round($result*100))/100 . "</td>";

}

#15 Re:First Survey Test

by SomeSmartAss

Sunday June 6th, 1999 10:59 AM

Reply to this message

Oops, that first <img width> should be ($result * -10)

#16 Re:First Survey Test

by mozineAdmin

Sunday June 6th, 1999 11:01 AM

Reply to this message

SSA,

That's a decent idea, and something I'll definitely consider, but I want to keep the scoring as it is for now -- at least until the vote in the animation contest. I don't want the possibility of anyone getting a negative score for their hard work.

#17 Re:First Survey Test

by mozineAdmin

Sunday June 6th, 1999 11:07 AM

Reply to this message

Oh, another thing:

When we finally are done tuning the code, you won't be allowed to see the scores before you vote. This is to avoid people from attempting to skew the scores. During voting, scores will only be visible to people who have voted in that particular survey. We will release the scores at the end of the voting, for everyone to see...

#18 Re:First Survey Test

by SomeSmartAss

Sunday June 6th, 1999 11:39 AM

Reply to this message

"I don't want the possibility of anyone getting a negative score for their hard work."

Fair enough, but using the current system, where "1=extremely negative" and "10=moderate assesment" a score under 10 still effectively means "this Sucks!" 9 just sucks less than 2.

I think you need two systems of mesurement, one that actually ranks the given selections (1=like the least" "20=like the most"); a "contest mode" where the 1 doesn't nesesarily means it sucks, just that all the other ones are better. The other one, like the current system, should allow for opinions on the indevidual selections; a "wishlist mode".

#19 Re:First Survey Test

by arielb

Sunday June 6th, 1999 3:05 PM

Reply to this message

actually I don't even know why we're using a numerical range. "this really sucks" "this is bad" "this is ok" "this is really great" "undecided" makes more sense for this kind of question. oh well. (and btw threading rules! :)

#20 Re:First Survey Test

by SomeSmartAss

Sunday June 6th, 1999 6:00 PM

Reply to this message

"I don't even know why we're using a numerical range."

Numbers are easier to graph and manipulate. Even if mozAdmin was going to put qualitative descriptors in the picklist, there would probably be a numerical value hidden in the back, in some form or another. You still have to quanitfy it somehow.

#21 Re:First Survey Test

by Anon

Monday June 7th, 1999 6:41 AM

Reply to this message

I also like things the way they are right now. I noticed that the sites I visit and read the most all have flat comment display. Having to click on message links to read a threaded discussion is a pain.

To sum it up, KISS :)

#22 Re:First Survey Test

by mozineAdmin

Monday June 7th, 1999 7:40 AM

Reply to this message

"I noticed that the sites I visit and read the most all have flat comment display."

I tend to agree. For one thing, it keeps everyone involved in the discussion. It also tends to keep the discussion more focused. Well, most of the time.

If I did implement a threaded version, it would be an alternative to viewing the messages flat. However, that creates this problem: if you view a threaded discussion in flat mode, discussions are broken up, and it's difficult to track them. I could, technically, provide links to parent messages in the headers of a flat displayed forum, but then we're only accomodating, instead of necessarily providing the best experience.

#23 Re:First Survey Test

by SomeSmartAss

Monday June 7th, 1999 8:02 AM

Reply to this message

"If I did implement a threaded version, it would be an alternative to viewing the messages flat."

Hmmmm.... Perhaps you could but a little button or link that says "reply to this posting" by pressing this, you get shoved down to the response form, and a <INPUT TYPE=HIDDEN> element gets filled in with the original postings "id".

the replys would then be shown with the original message's "header" (title, submitter's name and date of submition) info, as well as its own in the yellow box above each response.

An option to view messages in straight chronological order, as they are shown currently. or "pseudo-thread" where, while not indented or anything, the replys get listed directly after the messages they were reffering to, allowing people to follow the train of thought.

This would create difficulties with the current overflow format though... perhaps you could listing the latest postings, and the message they replied to (if any) in "pseudo-thread" overflow mode, so viewing the last 5 messages, could, concievably bring back 10.

bitch of a select statement either way...

#24 Re:First Survey Test

by Kovu <Kovu401@netscape.net>

Tuesday June 8th, 1999 9:38 PM

Reply to this message

Just start a mozillazine newsgroup and have done with it :)

#25 Re:First Survey Test

by briank

Friday June 11th, 1999 6:01 AM

Reply to this message

So has it been decided at what score is high (or low) enough for one of the options to be chosen and implemented? Is it a majority 51-49% or does the score have to be a high rating such as 15?