Bugs and Suggestions

21-Dec-09 23:10

FAQ says:
What is the purpose of the 'Minor Changes' checkbox on the post page?
This option is only visible when updating an existing Quick Answers entry. By checking it you are marking the current edit as consisting of only cosmetic changes such as formatting, spelling or grammar. You would not check this if you are adding/removing content or reworking the post. Code Project uses this flag to filter what is sent out on RSS subscriptions to these entries.

First of all, I have to apologize about not using it the first time, I just forgot it. And that's actually the point.

I think it could be better using it just on the other way. I mean... Minor changes could be the default value of an edit. Most of the times, the edits are because of Typos, Format changes and so on. EDIT: At least when the edited message is your own message.

Relevant content changes, corrections or additions are less usual than the others and then, in my opinion, is when the check-box should be used.

Many people are not going to read the FAQ, and even people that have read the FAQs and are "legal" users (respecting the rules and the correct way of doing things) may forget to activate the Tag when doing a fast edit.

Regards.
--------
M.D.V. Wink | ;)

If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpfull answers is nice, but saying thanks can be even nicer.

modified on Tuesday, December 22, 2009 10:29 AM

Thiru Thirunavukarasu22-Dec-09 5:20

22-Dec-09 5:20

Thanks for the comments Nelek. But I'm afraid checking the 'Minor Change' option by default would inadvertently mark "real" changes as minor. As you correctly noted most people don't read the FAQ and won't bother adjusting this option. Though I'm sure many minor changes will be marked as non-minor I think that's better than having non-minor changes marked as minor..

Luc Pattyn22-Dec-09 6:47

22-Dec-09 6:47

I do not know what the relevance of "minor change" is, however offering a default will always result in a lot of false attributes. So I suggest:
1. you compare the item before and after the change; if size was altered by more than 10%, or more than 20% of the paragraphs were altered, it could be called a major change.
2. and when in doubt, you explicitly ask the author after he saved the content, i.e. make it sequential.

Smile | :)

Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]

I only read code that is properly indented, and rendered in a non-proportional font; hint: use PRE tags in forum messages

Thiru Thirunavukarasu22-Dec-09 7:04

22-Dec-09 7:04

Luc Pattyn wrote:
I do not know what the relevance of "minor change" is

When we have subscriptions in place (or when sending out emails) we'll hide minor change edits. We also hide them on the version history page.--In both cases to reduce noise. The value of this considering the complexity it adds is debatable. I first came across this feature in DokuWiki and thought it'd be useful here as well. But I only ever used that for myself - not with millions of other people. Practice may very well show us otherwise. Sniff | :^)

It'd definitely be good to add some smarts around detecting minor changes programmatically. I'm not sure if detecting the percentage changed would do it. And asking the user when saving might be too much of a burden? I personally wouldn't mind being asked this question but I wonder if other members would complain. Maybe if we brought up a pretty jQuery dialog it'd help things..

Anyway, thanks for the thoughts. We'll definitely need to re-evaluate how this works.

Luc Pattyn22-Dec-09 7:41

22-Dec-09 7:41

I think I understand now. As its main purpose is notification filtering, this would be my way of thinking:

1. have an automatic "change" evaluation, resulting in a number (say a scale of 0 to 2; i.e. 0 = any change; 1 = size changed by >=10% or >=20% of paragraphs altered; 2 = size changed by >=20% or >=40% of paragraphs altered)
2. have some way for the author/editor to bump the value upwards (add 0/1/2 saturating on maxvalue of course)
3. give the subscriber the choice to what change level he subscribes.

How this would be used by writers:
- a formatting/typo change gets a low value and does not deserve a bump, so it would result in 0 or 1 automatically; no worries.
- a content change (say adding an example) would result in 1 or 2 automatically; no worries.
- fixing an error (say adding "not" to a sentence) would result in an automatic 0, but deserve a full bump since everyone probably wants to know it. This is the only case requiring human intervention.

readers subscribe to change level according to relevance
- the original author, an editor, or a user would want it all (subscribe at 0)
- a casual reader may want notifications only for larger changes

Net result: the author/editor doesn't do much, it is the subscriber who decides what he will get notified about.

Smile | :)

Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]

I only read code that is properly indented, and rendered in a non-proportional font; hint: use PRE tags in forum messages

modified on Tuesday, December 22, 2009 2:05 PM

Thiru Thirunavukarasu22-Dec-09 10:48

22-Dec-09 10:48

Interesting. We have a task to automate Minor Change some how and I'll add your suggestions to the notes.

However I disagree that the only time we need human intervention is at 0. An editor may change significant parts of the post for readability. I think this should be marked as minor.

How about just requiring the editor make an explicit choice on whether their edit is minor or not? If they forget to make a choice we reveal (hidden by default) a detailed explanation on what Minor Change is all about.

Luc Pattyn22-Dec-09 11:05

22-Dec-09 11:05

Thiru Thirunavukarasu wrote:
However I disagree that the only time we need human intervention is at 0. An editor may change significant parts of the post for readability. I think this should be marked as minor.

I disagree, for several reasons:
1. An editor can make mistakes, if he changes a lot and I have visited before, I'll want and read it again, even if he intended not to change the meaning.
2. I will skip badly formatted questions (e.g. code without PRE tags), and may want notification when the formatting gets improved, while not being interested in minor changes.

Anyway, my suggestion remains: automatic evaluation, and the default is to accept that (so bump defaults to zero). If you think you must support a negative bump, by all means give it a try (but I wouldn't).

Thiru Thirunavukarasu wrote:
How about just requiring the editor make an explicit choice on whether their edit is minor or not? If they forget to make a choice we reveal (hidden by default) a detailed explanation on what Minor Change is all about.
Quote Selected Text

IMO the question, if asked at all, should be asked after the edit session. I don't always know beforehand how big my changes/additions will be. Here is an idea, if you don't like the bump idea:

when editing, only provide a "preview" and a "cancel" button, nothing else.
upon PREVIEW, show the preview page, which has four buttons: "back to editor", "publish as minor change", "publish as major change", "cancel". And make one of the publish buttons much bigger than the other, which one depends on the automatic change evaluation. (Do NOT let the buttons switch places!!!)

Advantages: author/editor is forced to preview and to think about minor/major change.
One can do more than leading the horse. Laugh | :laugh:

Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]

I only read code that is properly indented, and rendered in a non-proportional font; hint: use PRE tags in forum messages

Thiru Thirunavukarasu22-Dec-09 11:35

22-Dec-09 11:35

Luc Pattyn wrote:
1. An editor can make mistakes, if he changes a lot and I have visited before, I'll want and read it again, even if he intended not to change the meaning.
2. I will skip badly formatted questions (e.g. code without PRE tags), and may want notification when the formatting gets improved, while not being interested in minor changes.

Excellent points! I didn't think of it that way; you've convinced me. Smile | :)

We'll try your original suggestions (when we get to this task that is) and see how that works out. You're latter options are great as well. We can always make changes later on if there's some unforeseen issues with the first approach.

Thanks!

Luc Pattyn22-Dec-09 11:45

22-Dec-09 11:45

Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]

I only read code that is properly indented, and rendered in a non-proportional font; hint: use PRE tags in forum messages

Nelek22-Dec-09 22:27

22-Dec-09 22:27

Nice points.

I wouldn't mind to be forced to preview and choose before getting my edit posted.

Regards.
--------
M.D.V. Wink | ;)

Nelek22-Dec-09 8:20

22-Dec-09 8:20

People that do "real" changes are, most of the times, users with advanced knowledge and usual users/readers/contributors of this site. So these people know what to do and how to do it. So when one of that users do a major change, they will check / uncheck the tag, because it is good for their reputation points and so on (if I understood right the FAQs). I guess not so many will forget it.

Taking a look into your answer below:
"When we have subscriptions in place (or when sending out emails) we'll hide minor change edits. We also hide them on the version history page.--In both cases to reduce noise."

Having minor changes as default will automatically help you with this purpose.

When I ask a question, I keep track on the thread/andwers and so on during a few days (until I consider that probably no more answers will come). If "standard" users don't keep an eye on his own questions, is not your problem, it is his/her problem, but it will help you with less data to handle for your history and feedbacks.

Anyways... it was just an opinion.

Regards.
--------
M.D.V. Wink | ;)

Thiru Thirunavukarasu22-Dec-09 11:15

22-Dec-09 11:15

Thanks for your thoughts on this but I have to be honest. I don't think setting the default to Minor Change is the way to go. We either need some very clever way to determine this programmatically or have the editor make an explicit choice. Sniff | :^)

Nelek22-Dec-09 22:29

22-Dec-09 22:29

You are right. I just thought from the point of view of an average user. I didn't have the perspective of expert users. I think luc's ideas are much better (if not the best aproach) to make it.

Regards.
--------
M.D.V. Wink | ;)

Thiru Thirunavukarasu23-Dec-09 4:33

23-Dec-09 4:33

Great! Glad we're all agreed Wink | ;)

About voting system

Nelek21-Dec-09 22:38

21-Dec-09 22:38

Hi,

I have had for a time the issue that I can not vote messages in forums when I am at work, because of proxy/firewall of the firm. But in the new "Quick Answers" I haven't got this problem.

I am just curious, why is that? what is different?

Regards.
--------
M.D.V. Wink | ;)

Re: About voting system

Rajesh R Subramanian21-Dec-09 23:01

Rajesh R Subramanian

21-Dec-09 23:01

With quick answers, each query works like if it were an article. With our regular message boards, you'll be voting on a post. (I may be wrong though...)

“Follow your bliss.” – Joseph Campbell

Re: About voting system

Chris Maunder22-Dec-09 5:19

Chris Maunder

22-Dec-09 5:19

Are you seeing an error when you try to vote? (either javascript error or site error?)

cheers,
Chris Maunder

The Code Project Co-founder
Microsoft C++ MVP

Re: About voting system

Nelek22-Dec-09 8:24

22-Dec-09 8:24

Now that I am at home, I can see all and I can vote (5 to your message is mine), but at work... I don't get any explicit error, I can't just see the buttons to vote. I only get a generic error at the bottom-left of the explorer with "Error on the site" (or however it is said on the english version, mine is german and says "Fehler auf der Seite")

I have just noticed as well that, when I am surfing between messages in the forum, at work it is much slower than at home. Here I don't practically see anything when I go from one message to another. At work it reloads the forum-page when changing between messages. I guess I can't do anything, it very probably depends on Proxy/Firewall settings in the company, not in my computer.

I am using Internet Explorer 7, Windows XP SP2.

Thanks anyways.

Regards.
--------
M.D.V. Wink | ;)

The quick answer thing

Rajesh R Subramanian21-Dec-09 21:36

Rajesh R Subramanian

21-Dec-09 21:36

The quick answers faq says that the old style technical forums will be eventually "closed". Why on earth would this need to be done?!

I think that the Quick answers thing is a good initiative and I've put efforts from my side (whatever little may it be) and have been giving suggestions, reporting bugs, participating in the thing, etc., hoping to make it better. But not did I have the faintest idea of how this is going to "replace" the threaded style forums.

If I should give a very honest opinion, I think that the quick answers thing *still* sucks massive hairy balls and would take may be umm... a long time to become a replacement for the threaded style programming forums, if at at all, ever. The threaded style forums has evolved over a long period of time, is time tested and is proven to work just right.

There are several issues with the Quick Answers thing (What's it? call it a "thing"?):

The enquirers themselves post their further queries as "Answers". There's no way you could control this or make each and every one of them understand how exactly is this thing is supposed to be used. No, they won't read an FAQ. Not all of them at least. And there are too many of them. This will never happen with the existing system.
There are multiple issues with tagging. There's no way, for example, to add a new tag or edit an existing one to make it better.
I cannot remove a query (duplicates, urgnz pls code sndz, clueless, etc.,).
If I ask a query to the enquirer (even if I ask it in the message board at the bottom), the enquirer posts the clarification to my query as an "answer". I'm sick of this, because I'm not being notified. OK, I know you guys will possibly let me subscribe to the thread, but I want to be notified ONLY if the OP provides clarifications to my query, please. Not if someone adds another answer, or if the query itself was modified to add more information. The existing system is absolutely the best when looked at how it deals with this issue! The possibility of anyone screwing it up is very less.
I see people with absolutley no clue of C++ will "answer" C++ queries (like, "Hey, I'm not going to do your homework"). Well, if CG says it, I'll buy it because he actually can *do* that C++ homework if he desires to. However, I've seen numerous people, who've never visited the C++ forum, will answer a c++ query, ONLY to tell that they won't do homework (I'll mention some of those queries were actually NOT homework, but people take it up on themselves). Add clutter to the query, and add 10 to their reputation! Awesomeness! This won't happen with the forums, because they won't dare visiting the C++ forum. This is not exclusive to the C++ queries alone. I'm not saying they are wrong or right, but this is how this new "system" works.
This one is VERY important - with the threaded style, I can just go on answering queries one after another, without having to navigate away from a page. Unless I've reached the end of one page. With quick answers, each question is an "article", no matter how sh*tty, rubbish or clueless it is. I've to visit that particular page and answer it. With the forum style, even the most clueless person will know how to give me clarification (by "reply"ing to my post), when I ask for it!
I see this new system has taken a large amount of inspiration from stackoverflow. Good, but there's no need to make this place as an exact replica of SO. I love the threaded style forums at CP, and that's why I'm here. If this becomes another copy of SO, I might just go to SO as well. I'm (and many more are, I guess) at CP because of the threaded style UI, which makes navigating through queries a breeze. I think that there's no mandate to "improve" something that is already pretty good, unless you're making it better. Unless you're going to follow the footsteps of MS and do what they did to XP, "improving" it to Vista.

*We* are the people using it, so I'd doubt if you should take any drastic steps like shutting down the threaded style forums without taking in our opinions.

I haven't given any solid feedback that you people can use to make the Q/A better, or even the existing forums better. But I will post you guys, as and when I get ideas and/or encounter bugs. But my idea of creating this thread is to let you guys know of what I think of the new system as a replacement to the existing one.

There are several improvements that can be done to the threaded style forums, and I (and others too) may give their feedback and opinion on this. But I do not wish to waste a minute of my time if it's been decided to be shut down.

I'm not discouraging the new system. But I'm saying that we can have both systems side by side, at least for now. We can talk! Say what? A poll, perhaps? Or a sticky right here with a link to the discussion on this? I'm sure the new system is no match for the threaded forums.

“Follow your bliss.” – Joseph Campbell

Richard MacCutchan21-Dec-09 22:14

Richard MacCutchan

21-Dec-09 22:14

I agree 100% with Rajesh's comments. I find QA far less easy to work with than the existing forums, and also far easier to abuse by posters who may not understand the rules and, as Rajesh points out, simply don't bother to read the FAQ.

Nelek21-Dec-09 22:22

21-Dec-09 22:22

+ 1

Regards.
--------
M.D.V. Wink | ;)

Nagy Vilmos22-Dec-09 0:17

Nagy Vilmos

22-Dec-09 0:17

I'll agree with the sentiment that once we the users are happy then it is time to move on, but the new Q&A is still flawed.

I have to say I like the idea, and I am tryuing to be invloved and use it. I generally look at C# and java and my filters are thus set. It's cool from that point of view. Also anyone with enoughn reputation can go in and edit the contributions making them clearer and correcting tipping errors. From that point of view it is better then the old forums.

My view would be to keep a limited number of discussion forums for just that. It's not a question, but if I want to discuss some new language thang then it's the lounge and a flaming or Q&A where it doesn't fit. Can we have a way to add an addendum to prior answers especially for adding details when a new answer is not needed but you don't want to [or can't] change the original post.

Panic, Chaos, Destruction.
My work here is done.