OSDev.org

JAAman wrote: my point was that we have lost too many of our better people and the problem isnt with the quality of the questions (given the chance, most people will improve over time), but with the quality of the responses

and not specifically with the things mentioned in that conversation, as i dont really have a problem with expecting people to do some of there own research, (although i think some questions could be answered better), but with the attitudes of the responding people -- like my dad always told me, if you dont have something nice to say, then just dont say anything... (which is actually much easier to do online then in person...) -- i never have a problem with the 'younger' (not necessarily age-wise) peoples questions, as they are often inexperienced, but when the regulars respond, i expect more maturity -- if you dont want to answer the question, or if you think the persons name, or use of capitalization, or puctuation proves that they arent serious and dont deserve an answer? nobody is making you reply...

i think a lot of things could be improved if people would just reread the post before posting it -- a lot of times i write a very lengthy reply to someone, and then press preview, then walk away, and come back in a couple minutes, and read what i wrote -- most of the time i will just close the tab and continue reading other threads -- realising that my post was angry or would start a flame war

I would personally recommend that those who do not wish to redirect newbies or help them, that they simply ignore those posts and let the moderators/admin deal with the redundant spam-like posts.

Regular members need to understand that administrative actions are not their right, privilege or responsibility. We do not need a mob mentality or vigilante "n00b hunters" on this forum.

I am personally wondering why the poster 'a' in that quote hasn't read the bloody manual yet - He doesn't allude to it in his text, only that he has read tutorials. Maybe he should?

As in any "what's your opinion" thread, I do my first reply before reading what all the others said, so I give my opinion, not the reaction to other's opinions.

Three points, really:

• New wiki not as "accessible" as the old MT one,
• Too few "RTFM / redirect" replies,
• Overall quality of discussion.

New wiki not as "accessible" as the old MT one

The old Wiki (OS FAQ @ MT), on the start page, listed a comprehensive overview of what was in there. While some pages had misleading or too-long titles, it was an impressive ressource at first glance. Like the index of a good technical book.

The new Wiki, while being better organised, is too "hierarchical" for my tastes: It doesn't give you an index of what's in the book, it gives you a number of books to chose from, so to speak. It also does no longer push "FAQ" into your face, which was A Good Thing IMHO.

So, while the new wiki retains its position as a valuable resource, it lost its appeal as a "beginner's starting point". Together with the rather defensively worded "sticky thread" (Read this before you post...), many newcomers asked their already-covered questions before even realizing there is a Wiki, and how good its contents are.

Too few "RTFM / redirect" replies

The new [ wiki ] markup tag makes it really easy to redirect into the Wiki (which is where redirects should point IMHO, not into old forum threads). However, I get the impression that too few people actually use the Wiki at all. Instead, questions that have been discussed hundreds of times before (like, DJGPP vs. Cygwin, whether to build a cross-compiler, how to set up your memory management) are getting discussed again every time they pop up.

My impression is that the focus of the whole board has shifted from "getting questions answered (effectively)" (something that made MT a great place IMHO) to "discussing" (i.e., killing time / showing off).

To maintain a good signal/noise ratio, it is necessary to cut off the 146th thread on "how do I make a flat binary", post a RTFM / [ wiki ] reply, and leave it at that, so we can focus on questions that aren't sufficiently answered by the wiki yet.

Overall quality of discussion.

Because of this "very friendly umpteenth discussion on basics" problem, several excellent developers have left the vicinity, or post only rarely, because - to be quite blunt - you cannot expect cracks to hang out where people discuss how to use a compiler. Those who remain, are all too often lured into uttering "RTFM" / "STFW" without a link (a Google metalink suffices for the really dumb questions), which only deepens the problem.



Suggestion

• Create a "Wiki index page", call it a sitemap if you will, and make that the Main Page. Offer the categorized overview as a link, or alternative, but newcomers should see at one glance the breadth of wisdom contained there.
• Make the wiki link much more prominent on the page header. Do call it a "FAQ". Make a locked sticky thread "READ ME BEFORE YOU POST!", and threaten with grieveous bodily harm any fool who posts without visiting the Wiki and / or search page (before offering them a tea and a cookie, because we're a friendly bunch here).
• Everyone applying some self-control not to reiterate old holy wars, especially when feeling strongly about an issue, but rather pointing to the relevant Wiki page (where the holy war is done in a focussed way on the discussion page).

Now I'll read what the others wrote.

Solar wrote: To maintain a good signal/noise ratio, it is necessary to cut off the 146th thread on "how do I make a flat binary", post a RTFM / [ wiki ] reply, and leave it at that, so we can focus on questions that aren't sufficiently answered by the wiki yet.

I'll add one thing to this. If it does turn into a holy war and a 5-page discussion, please please please add it to the wiki so that the information is in there. What's the good of having a wiki if it doesn't have any information?

Seriously, don't even worry if it's formatted badly or whatever. Somebody will fix it later (yes, that's me volunteering).

Solar wrote:Create a "Wiki index page", call it a sitemap if you will, and make that the Main Page. Offer the categorized overview as a link, or alternative, but newcomers should see at one glance the breadth of wisdom contained there.

What I recall from the MT index was that it was a bloody mess and with respect to that the current front page is a huge improvement. I can't easily think of a system that would keep the layout clean while having all the information on the front page.
I did find a way to cheat on the system (which can be a serious improvement for the FAQ page) that allows you to arbitrarily link a page under a different name - which allows the proper title to be retained in the topic while keeping it accessible from the FAQ category. Another solution would be to change the category to a manually maintained page.

Solar wrote:Make the wiki link much more prominent on the page header. Do call it a "FAQ". Make a locked sticky thread "READ ME BEFORE YOU POST!", and threaten with grieveous bodily harm any fool who posts without visiting the Wiki and / or search page (before offering them a tea and a cookie, because we're a friendly bunch here).

I just edited the top sticky post in general programming to reflect that, if some admin can be coerced into removing the trailing junk and lock the thread we can achieve that idea.

Everyone applying some self-control not to reiterate old holy wars, especially when feeling strongly about an issue, but rather pointing to the relevant Wiki page (where the holy war is done in a focussed way on the discussion page).

...I'm applying self-control by not pointing at people ...

Just a reminder that this posts contains images and as far as I know you must be logged in to view images.

Statistics

From a little scavenging I was able to produce a moderately accurate plot of the activity of these forums from 2001 to late 2007. The time values are not perfect since my algorithm for converting months and years into days was not perfect. It gave each month 30 days, and each year 365 days. It always subtracted 1995 from 2007. This together produced the odd time values used to reference points in history.

The major point is that a high slope means high activity while a low slope or no slope means low activity or no activity. Activity is computed by the creation or bumping of a thread/topic (not posts). A problem with this is it can steal from previous points in history if someone bumps an old thread/topic from a year ago to current day. This will cause one unit of activity to be shown for today instead of a year ago. So although not perfect it still gives a rough look at activity increase over the years.

Another pretty interesting part is where the low thread identification numbers come from, and are these from the threads that got transfered from the Mega-Tokyo forums.



This is another pretty interesting one which shows the number of views per thread. The last post time dates the thread on the plot. So if a old thread was bumped then it will show up not when it was created, but when the bump was made. A bump -- meaning someone posted a reply to the thread.

It shows a very high amount of views early around 2001 which could be from the lack of participants and the majority of viewers. So this chart might convey a little about the ratio between viewers and posters.

• When the span is small and the middle is higher it could mean more viewers than posters and when the span is large it could mean viewers.
• The span middle should relate to the ratio where a span having a high center means more viewers than posters, and a span with a low center means more posters than viewers.
• If the span sort of covers both the low and high it should relate to there being a more balanced ratio between viewers and posters.

Also it could show that an increased exposure such as many more people visiting the site to perhaps just search the forums for answers has happened late in the time line where the original trend in 2001 is starting to appear again, but this time the number of views per thread has risen noticeably.

Mega-Tokyo
I agree with Solar about the Mega-Tokyo issue. I really think he is right on target with it. I like these forums, but there are things missing that were part of the Mega-Tokyo forum and some of those things are what Solar has pointed out. Also about the FAQ. It does make a lot of sense.

Also just for the record what ever Candy, Solar, and Pype.Clicker says about the forum is most likely right or so close that a little ironing would fix how ever it is off. I would also add Brendan in saying this, because he simply holds a really large amount of professionalism at times in which even some of the strongest people would just break.

About Original IRC Quote
Just to re-illiterate I want to say that the two guys quoted could be completely wrong, and as a number of people have pointed out that they think the RTFM scene is justified. Also as some have said there are at least a significant number of cases where things are way too RTFM or no were near an RTFM or answer. Either of the later are bad, and of course myself being guilty of things like this.

OK, I had the time to read through the thread, so here's my second take.

We had two rather strong communities coming together, and as can be seen here there still seems to be some one-upmanship involved. This is very bad, and in more than one way. I can judge only from the MT perspective, so bear with me.

While MT had its share of noobs and newbie-questions, at an increasing rate, I'd say it is unfair to say that it was the MT merger that brought the noob problem here. Every technical board has this problem, partially because experienced people leave, partially because popularity rises. If both boards had never merged, both would have faced the same problem.

Quoting relative member numbers or activity rates to show who brought more to the table is pretty stupid (sorry), because we are one forum and one Wiki now, end of story. Let's focus on tomorrow instead of yesterday.

However, the problem does lie in this yesterday. The merger brought much insecurity, new people, new structures. As a moderator, I had to think twice before taking any action, because I was not sure if my "MT way of modding" would be OK with the OSDev part of the community. The FAQ was in transition, and still is. None of this would bother newcomers, but it reduced the effectiveness of the old farts, made them a bit uneasy, made the forum feel a bit less like "digital home". Some left, some reduced their forum time. Me, I resigned from moderator status in no small part because of what I wrote in this paragraph.

This loss of experience could have killed the community outright, but we're here and talking, so there's hope.

I remember Dex calling MT's mods "heavy-handed" on his homepage. I was quite OK with that. Moderator activity makes it clear that this particular forum is not a free-for-all discussion place, but that the technical sub-forums are goal-oriented. Much of the insecurity mentioned above made itself felt in the moderators being less active, less consistent, the forum being less "moderated" overall. (Those who remember Usenet News will know the difference between a moderated and an unmoderated list all too well...)

Another point for experienced people turning away could have been the changes in the hardware field. A few years ago, a hobbyist OS aiming at single-core 386 hardware could be pretty great, because it ran on an Athlon XP, and who had a hyperthreading PIV anyway. Today, you have to think about SMP and long mode right from the start if you still want "the real thing" - the bar was raised by a considerable margin, and for me, it was the end of any plans for a continuation of my OS project.

To do

Moving the FAQ Wiki more into focus is IMHO the pivot point. Basically, everything posted in "OS Development" can be one of two things: A violation of "READ ME FIRST", or a feature request / bug report for the Wiki. The forum should be seen as something akin to a "osfaq-devel" mailing list, the occassional time-killing discussion or humorous banter nonwithstanding.

Yes, the MT Wiki front page was a mess that was constantly being worked upon, reorganized and reviewed. It was like a desk cluttered with reference printouts, tutorials, and generally good advice.

The new Wiki, IMHO, tries too much to be a book, with a nice shiny cover; an effect perhaps compounded by the use of the MediaWiki software of Wikipedia fame. We are not an encyclopedia, nor should we ever try to be. Things move too fast in the OS world to try this.

I vote for a return to an "all-in-one" frontpage.

i never meant to imply that MT brought the 'noobs', im sorry if it sounded like i did -- really the only reason i mentioned it at all, was that this is not the first time i have heard people mention that the MT board 'changed its name' to OSdev.org... i believe a lot of people dont know that this board was here before the merger

the only thing i didnt like about it, was the fact that overnight i went from the most-frequent poster here, to the least-frequent poster, but thats not really a problem, as i tend to be quite shy and try to stay out of sight anyway

i certainly dont think any of these problems were caused by the merger, and it wasnt actually till a few months later that i started seeing the real problems...

actually, if you want to know the truth, the worst problems seemed to start around the time you resigned... i was really sorry to see that -- i logged in and saw you were no longer on the list, i was really sad about that, since i really respect your opinion on programming topics, and have always felt you were the best of the mods -- important to me, since i have never really trusted myself...

JAAman wrote:i never meant to imply that MT brought the 'noobs', im sorry if it sounded like i did -- really the only reason i mentioned it at all, was that this is not the first time i have heard people mention that the MT board 'changed its name' to OSdev.org... i believe a lot of people dont know that this board was here before the merger

A little OT, but I wanted to say that this is yet more evidence of just how painless the transition away from MT was. Major kudos to chase and all who helped!

yes, you are right, it was quite seamless

though i must retract one thing:

the only real problem is, as an osdev mod i would have expected to be (at least) informed of the proposed merger before it happened... not that i would have had a problem with it, but...

only way i found out that it even happened is i looked at MT and noticed that the OSdev board was gone... and a message said it had moved here... i was thinking 'what?? i thought it was moving to osdever...' and 'why is df the one telling me, and not chase...' and i quickly came back to osdev to look and see if there were any messages from chase in my box or on the mod board... but there was nothing -- there wasnt anything at all from chase until he decided to add the old MT mods to the mod staff, he did ask about that (which i had no problem with, though i did appreciate being asked)

JAAman wrote:actually, if you want to know the truth...

High praise indeed. I would love to follow up on this with a promise of return, but I cannot really say as I am changing jobs and relocating, so everything from a complete vanish to a return to regular attendance is possible.

Method
Place a time delay on posts. Such than when posting it becomes delayed by say six to twelve hours.

Why
1. It would cause people to not look to the forum for a quick answer, instead people would only ask questions when they have pretty much exhausted their resources. After exhausting their resources then they post the question. You should notice that most people -- the newer ones -- expect an answer with in a few hours. This appears to be the trend on forums in today's world.

2. To reduce casual conversation in the technical question answering sub forums. Some people actually use the forums to have casual conversations which I like, and I think is fine. But, the only problem of allowing casual conversation sub forums is that questions may be posted there in hopes of getting a quick answer leading to moderators having to constantly move threads into the waiting que (six to twelve hours).

A Improved Method
Use a slow stepping of the delay. So that it starts out at a minute and moves upward over the weeks. As it moves up slowly we could take note as how the forum starts to change. Carefully, watch the activity to determine if users are disgruntled about the new method and are just not wanting to spend time posting about it. See if the thread topics change in context to something more effective.

Most users are going to check the site once or twice a day, or when they become notified of a new post. This delay would give them time to read back over their waiting posts which would should be equal to the "walking away and coming back later and reading your post". This might increase the value of each post as the user notices problems in their posts and goes back to correct them.

A lot of people end up posting _too_ _quickly_, including myself. This leads to errors and typos which then leads to lots of edits or other problems. If we post and then take time to think about it we usually change our minds or realize we need to add something more.

Actually letting the post display, but distinguished from actual posted posts would allow the user to click the thread and see their post there which would give them an idea of what it looks like (after the question or most recent post) which is an important factor here. Then once the delay is up the distinguishing mark could be removed and the posts becomes posted.

I know posting a post sounds weird, but my intention was just the explain the idea not actually refine it to a usable point.

I don't believe that delaying the posting time is the answer. Your idea may have merit on a forum where theory is mainly discussed, and where long posts are more often than not, but here, quick answers are often useful.

I believe the problem is simply people not reading and doing proper research. We have a notice. We have a wiki. We have google. Moreover, we have the whole bloody internet at our disposal. For the most part, if people did research, the entire 'OS Development' forum would be much smaller.

As for the regulars, I hope that this thread wakes us up. I haven't always been the most 'appropriate' poster, and as such, I am trying to watch my posts more carefully and to follow other such complaints on this thread. I hope other do the same.

On a lighter note, how can we force new members to do research?

On a lighter note, how can we force new members to do research?

Are you going to:

1. Wait six hours for an answer.
2. See if you can find the answer quickly in the FAQ, Web, Wiki, or search existing questions in the forums -- then retract your question if you already found the answer.

The answer is obvious.
Not the answer is obvious meaning my idea was right, but that is does qualify as a reason.

I don't believe that delaying the posting time is the answer. Your idea may have merit on a forum where theory is mainly discussed, and where long posts are more often than not, but here, quick answers are often useful.

Also the above may be the problem. We want quick answers. So get a quick answer you have to ask a quick question and quick questions do not have any relation to doing any research or at least no significant amount of research.

I seriously don't like the idea of retarding the forums just to stop a few newbies complaining about RTFM replies. This seems to be getting out of proportion. (I don't want to wait 6 hours for a reply to a question, and I do research!)