OSDev.org

Seeing as there a fair amount of newcomers to OSDev (which is good) which keep asking questions that have been answered many times before and are answered on the wiki (which is bad), I would like to suggest (open for discussion, of course) the following: What if there were to be a button next to "Report this post" (or included in that button with a list or something) that said "Didn't Read" or something similar? Members of the forum can then click this button and a moderator and/or admin might accept or deny the notification. If the notification is "confirmed", the moderator may perform an action that automatically inserts a (friendly! not "RTFM" or something) message "kindly" requesting the OP to read a (possibly to be specified) article or link. The topic is then automatically locked.

The goal would be to prevent other members:

• Insulting the OP for his "ignorance" (in some rare cases it may be justified IMHO, but usually newcomers are just starting out and don't know too much about it yet).
• The OP will (should) be satisfied as the link provided by the automatic message should answer his question.
• Nowone wastes his or her time on making (un)constructive posts.
• Improve OSDev friendliness? I know some people that actually don't want to register at this board because they think most of us react too drastically at a newcomer/member which asks a "nooby" question (which should be normal, after all).

The only downside would be that a moderator/admin would have to take a look at the notification every time it is performed. Members should then also be asked to actually use this button instead of either flaming the OP or answering the same question over and over again (the latter is good, but it may still be a waste of time seeing as it is an already-answered question).

I just came up with this and it might come out a bit raw. As I said, I'm always open for suggestions or critiscism. It is after all up to the staff of OSDev to determine whether such a feature is necessary/possible/to be implemented/etc.

It seems like this is equivalent to (and more complicated than) adding a forum rule that everyone should report threads by OPs who didn't read the wiki, and adding a corresponding reason in the report post menu: why not just do that?

NickJohnson wrote:It seems like this is equivalent to (and more complicated than) adding a forum rule that everyone should report threads by OPs who didn't read the wiki, and adding a corresponding reason in the report post menu: why not just do that?

Adding a line to the existing Report Post box seems like a good idea as well. It's basically the same, only it indeed eliminates the need for an extra button and such.

I vote for first three posts needing approval. It's an integrated phpbb feature, troll-proof, and can easily be set.

I also know which of the mods were and are going to argue against that suggestion.

Why go through all the trouble when simply posting "There's a wiki on [insert question], read it. It's called [insert title of wiki page]" IMO I hate reading all the RTFMs, only one person should post that the wiki exists and possible the exact name, sometimes it's just the terminology and not knowing what to search that keeps them(and me) from finding the right page, anyone who feels the question is so bad it doesn't even deserve what I've suggested, should just ignore them. They'll figure out eventually, after getting only one or no replies, that they need to put more effort into finding the answer on their own, and all without polluting osdev with ugly RTFMs.

An Internet forum really serves three purposes. The first is for socialisation between people who share the same hobby/ideals (ie General Ramblings). The second is peer review of ideas and products (OS Theory, Announcements and Test Requests). The third is for people who are doing something for the first time to get advice and information off people who've done it before.

For the first two purposes, efficiency doesn't matter. Human interaction is generally obscenely inefficient (anybody who has ever tried to pick up in a bar has experienced this firsthand). For the third purpose however, it does. The idea is to get the information from the expert to the beginner in the most efficient way possible, both for the expert and the beginner.

Now, if the beginner needs the information, they will have to get it eventually. So at some point, somebody will have to provide them with a link of material to read or answer their question. Otherwise the whole forum goes nowhere. So really it becomes a matter of what is most efficient for the expert(s), whilst making sure the beginner gets the information they need. In my mind, this is simply providing a link (where the question has been asked before), or answering the question and saving the material for reference later (ie editing the wiki). Once a link has been posted or the answer has been given satisfactorily, further replies decrease efficiency of the forum. And we all hate that.

Where we are falling down currently is getting the information from forum replies into wiki articles. I mean, how much of Brendan's or Solar's or Combuster's knowledge on OSDev is currently in the wiki? About 1%, I'd say. This is something we all need to work on more. Even just creating links back from wiki pages to specific forum replies is a great start. Anybody can do that. If you really do know absolutely nothing (like me), you can still help by editing and making things easier to understand. Even if you're no good at English, you can still help by creating links everywhere.

JackScott wrote: ...

Where we are falling down currently is getting the information from forum replies into wiki articles. I mean, how much of Brendan's or Solar's or Combuster's knowledge on OSDev is currently in the wiki? About 1%, I'd say. This is something we all need to work on more. Even just creating links back from wiki pages to specific forum replies is a great start. Anybody can do that. If you really do know absolutely nothing (like me), you can still help by editing and making things easier to understand. Even if you're no good at English, you can still help by creating links everywhere.

I agree that the first post (usually) provides the necessary information that answers the question (like a link to the wiki). It's what's wrapped around this link that bothers most of us I think. Simply dropping the link would be a bit too "uncivilized", so some members apparently feel the need to drop in an extra "Read the F wiki" or "RTFM" next to that. It is true that these members might have searched the information for themselves (there IS a link on top of the forums after all), but is it really necessary to make them feel bad because they are new to the subject?

To reply to JackScott's last paragraph: maybe it would be a good idea to have some sort of initiative (maybe somewhere around the summer holidays, like july - august) where we would try and majorly update the wiki. The downside is that nowone gets paid, lots of people won't feel like it (it's voluntary after all), but IF (and I say if) it were to succeed, we might have a lot of new information in the wiki. This would also be a good time for long-overdue-but-mandatory improvements or updates that have been postponed for ages.

JackScott wrote:Where we are falling down currently is getting the information from forum replies into wiki articles. I mean, how much of Brendan's or Solar's or Combuster's knowledge on OSDev is currently in the wiki? About 1%, I'd say.

In knowledge, a lot of what I know is available on the internet, maybe only for 10% in the wiki, but even more so across the intel and AMD manuals - which I tend to refer to as the bible for good reason. Even my favored subject is covered to more depth than most newbies would need - there's VGA Hardware, the lot of FAQ entries on graphics, plus the bunch of hardware descriptions stored in my user namespace. In the end, the wiki, vgadoc, RBIL and the intel manuals cater for like 75% of what I use in OSdev specific knowledge. Expect me to throw the relevant manual to your head when you forget them. (velocity dependent on the obviousness of the offence)

Most of the actual knowledge is not OS development specific, but general computer science: imperative programming, functional programming, system architecture, data structures, algorithmic complexity, networking, concurrency, program correctness, grammars, and software architecture. And possibly even more important: linear algebra, sequence maths, boolean logic, predicate logic and set theory. Try to find the similarities with the list of required knowledge, much of it overlaps or is a direct consequence of mastering any of those subjects. Expect me to suggest you to get lost in the amazon rainforest when you have holes in this area. (duration dependent on the size of the gap)

And even more important are these skills: precision, sound reasoning and systematic problem solving. And here's a physical problem: while you can train this, part of it is genetically determined in a person's intelligence, and when it's below some level there's no chance that your innate abilities and your training combined can solve problems as complex as kernel design. When its all hopeless, expect me to tell you to get a job as rocket fuel, then delete that message in transit.


IMO the only thing I can and do offer you is my experience of the things any of you could do for yourself. Relative to the 1% you see, is for 90% things that words can't do: skill and experience.

In my case at least, most of what is perceived as "knowledge" is "research skill". I daresay it's similar for most others here.

You read. You understand. You remember - if nothing more than a vague recollection that you read about a subject somewhere, which gives you an idea on where to start looking again (Google keywords, or which book from your bookshelf). You skim the paragraphs again, which revives your memory, and then type the answer.

This is what we are really talking about when we want to make newcomers into OS dever's that can carry their own weight. I have a feeling that 75-90% of the newcomers aren't really interested in learning about the Intel architecture, i.e. reading the manuals to understand what's going on - they want the result, just want their code to run "and please don't bother me with too much detail".

I blame it on the media. Too many casting shows.

The Wiki is a resource for "the rest", the kind of non-referencial information that you can't find in the manuals, but which puts it all together.

I strongly believe that the "old" view of the Wiki ("Expanded View") is better, because it gives you somewhat of a "roadmap" as well as a better overview of what is available from the Wiki. It's more like a book with chapters. This view was what drove development of the Wiki contents over at mega-tokyo.com, which has been frantic for some times, but has fallen into neglect pretty much since we got the new Wiki. I feel that the new "categorized" view doesn't give a good impression of the available information, and doesn't really inspire people to "click around" and/or extend the contents.

Another drawback of the new Wiki structure, or so I feel, is the use of MediaWiki. It's a very good framework, don't get me wrong, but it carries the Wikipedia legacy. One point in case is that the "Recent Changes" page isn't that prominently linked. On Wikipedia, that page is utterly useless. For OSDev.org, it's the second most-important page after the Forum board index: It tells you what is new, or has been changed. Either way, you have a look at it, correct typos, start a discussion, whatever. Looking at "Recent Changes" is what drives a Wiki like ours.

Another, more subtle, effect is that MediaWiki looks so "finished", with its polished looks, discussion page and all, that you feel you shouldn't touch the contents. I very much liked the "hacked" look of PhpWiki (not the framework, which sucked, but the look), because it screamed "construction site, come and join the work".

My suggestions:

• Scrap the Wiki "Short View" or put it into a subpage, and make the "Expanded View" the default again.
• Create a header block that links "Recent Changes", "Orphaned Pages", and "Wanted Pages" as well as the "Wish List".
• Somehow tinker with the theme layouts so the forum header "screams" the Wiki at you. IMNSHO, the Wiki should be the primary focus, with the forum coming a clear second.

Being active in the forums is good. Being active in the Wiki would be even better. That's for both technical information and the more linguistical inclined. The former put the information in there, the latter give it polish and structure (WikiGnomes FTW!).

@Solar's post: +1.

That was a very pleasant read.

Solar wrote:In my case at least, most of what is perceived as "knowledge" is "research skill". I daresay it's similar for most others here.

You read. You understand. You remember - if nothing more than a vague recollection that you read about a subject somewhere, which gives you an idea on where to start looking again (Google keywords, or which book from your bookshelf). You skim the paragraphs again, which revives your memory, and then type the answer.

This is what we are really talking about when we want to make newcomers into OS dever's that can carry their own weight. I have a feeling that 75-90% of the newcomers aren't really interested in learning about the Intel architecture, i.e. reading the manuals to understand what's going on - they want the result, just want their code to run "and please don't bother me with too much detail".

I loved your post, but you bring up a rather interesting point that I'd love to hear your thoughts on: What is your opinion on people who, for whatever reason, hit a concept that they just can't wrap their heads around? Pulling an example out of my rear end, something like paging - you have a person who has read the intel manuals, mined the internet for resources, and just can't seem to understand the concept enough to implement it. When you're talking about turning newcomers into OSDever's who can hold their own, how do you handle people like that? How should the community as a whole?

No shame in not understanding a concept. However, I feel that paging - as a concept - is something you should be familiar with before you set out to write your own OS.

I feel that writing an OS in order to learn about operating systems is going about it the wrong way. There should be a reason beyond simple one-upmanship for you to venture into this field. If you seek knowledge for knowledge's sake, there's enough to be learned in user-space, and more productive, too.

If you go into OS Dev, You should have a good idea of how your current OS of choice works, and preferably how that compares to at least one other system. You should know what's good about it and what's bad, so you can build your own concepts on existing experience. Some background knowledge on OS development theory wouldn't go amiss, too. It's hard to imagine someone collecting this amount of background know-how and not being able to understand paging.

That being said, I have no problem with questions being asked or answers being given. But I think a question should show that the person doing the asking has gone as far as possible on its own, and that this is far enough to show that the person isn't a complete noob. A question like "Couldn't understand paging, how does it work?" means that person has either does not have the necessary technical capabilities yet, or hasn't made a sincere effort. Wait, that's the same...

As to what I would answer:

We have worded our "rules" and "introductions" more and more harshly over the years, to no avail. Suggestions I've read here in recent days continue in that vein. I start to believe that we should reverse that trend. Rewrite the Introduction / Beginner Mistakes / ... part of the Wiki as a generic text, as friendly as can be, that details that OS development is simply the wrong place to learn about a programming language, a processor architecture, or technical documentation. List a number of user-space projects for people to try instead. Make it perfectly clear that it isn't meant to be an unfriendly reply, but a sincere constructive criticism. I'd gladly help with said text, but I can't promise to come up with something within the week.

Then, when the next noobish question arises, link to that document, a second link to a discussion of said document, and then close the thread immediately. Better yet, make that option open to users, not only operators. StackOverflow has something like that: If a number of people "vote" for a question to be closed that way (e.g. through the "Report this post" dialog), it gets a canned reply and is closed automatically.

Solar wrote:My suggestions:

• Scrap the Wiki "Short View" or put it into a subpage, and make the "Expanded View" the default again.
• Create a header block that links "Recent Changes", "Orphaned Pages", and "Wanted Pages" as well as the "Wish List".
• Somehow tinker with the theme layouts so the forum header "screams" the Wiki at you. IMNSHO, the Wiki should be the primary focus, with the forum coming a clear second.

Being active in the forums is good. Being active in the Wiki would be even better. That's for both technical information and the more linguistical inclined. The former put the information in there, the latter give it polish and structure (WikiGnomes FTW!).

I agree with all your points. The only remark I have is that "Wish List" already points to the other pages, but IMHO it doesn't really matter since that would only emphasize the "help us out" part of it all. The last part of your post is also why I suggested we'd have some sort of "let's revise it" undertaking somewhere in the summer holidays. Maybe we could start revising page by page and update them if necessary. Maybe it would even be a good idea to (if it would be done) to have some sort of banner "To Be Revised" just like we have "This article is a stub" and such. This way the wiki's "old" content may be revised and some new content may be added.

Solar wrote:We have worded our "rules" and "introductions" more and more harshly over the years, to no avail. Suggestions I've read here in recent days continue in that vein. I start to believe that we should reverse that trend. Rewrite the Introduction / Beginner Mistakes / ... part of the Wiki as a generic text, as friendly as can be, that details that OS development is simply the wrong place to learn about a programming language, a processor architecture, or technical documentation. List a number of user-space projects for people to try instead. Make it perfectly clear that it isn't meant to be an unfriendly reply, but a sincere constructive criticism. I'd gladly help with said text, but I can't promise to come up with something within the week.

Then, when the next noobish question arises, link to that document, a second link to a discussion of said document, and then close the thread immediately. Better yet, make that option open to users, not only operators. StackOverflow has something like that: If a number of people "vote" for a question to be closed that way (e.g. through the "Report this post" dialog), it gets a canned reply and is closed automatically.

+1: Sounds like a damn sound plan to me.

Maintaining a forum and separate wiki is not working, but there is a better solution (or will be soon):

http://blog.stackexchange.com/post/5184 ... change-2-0

This is IMO a much better way to share knowledge. It presents everything in a question-and-answer style, but also allows for collaborative wiki-style editing of the questions and answers. It also just about moderates itself, keeping the signal-to-noise ratio high.

All I have to say is that you better begin to become more friendly, because your so called community is dieing, now faster than ever. The first thing I would change is your nonsensical slogan. "The Place to Start for Operating System Developers" is impossible, because if someone was already an OS Developer, then why would he or she need a place to start, they must have already began developing, or surely they could not already call themselves an Operating System Developer, now could they?

Secondly, I would remove such users as Combuster who use this forum as a way to compensate and project. Rather than helping others, members like him are driving people away. He and others like him are smug and do not do well for a communities growth. Members with his attitude are worthless on many levels, surpassing all growth deprivation he and others like him cause. I could personally do without him completely on these boards and I think you could too.

Thirdly, Solar is no different than Combuster or any of the other handful of members like them. Solar you have written in the past (recently) about how people without much knowledge should just leave and never come back (paraphrasing). Anything you say in opposite of that now is hypocritical. We know you share the same aspect as Combuster, what's with all this Mother Theresa crap you're posting about now? Too late.

Lastly, I am glad this forum is being ran into the ground. I think it would be better to have it completely go away, so that the information provided in the wiki might become more of a focus and could perhaps one day be better written and with more content.