RapidMiner Wiki

steffen

Hello folks

The RapidMiner Wiki is not working... and the developers have simply too much work to do. Since my last nagging tactic has seemed to work, I will give it another try.

I now think that we need a wiki mainly for writing small articles about questions asked much to often. Honestly: When such a question appears in the forum and a quick search does not help, I often have no will to write the same things again and again .. Please note that I do not complain about the nature of the questions.
Beside this I do not see the focus of the wiki on explaining algorithmns in general ... there are much better references out there, which can together with some RM  specific notes could be linked in an article.

So the question is not, whether we need one (although I added this option to the poll to allow complete disagreement), but WOULD you contribute ?. It is not necessary to write complete masterpieces of literature (as the posts from Mr Haddock). Posting your process for specific tasks (after your questions have been answered) with some explaining would be a good start. This way you can give something back

So ...please vote and add some comments, too

Steffen

mihai

hi,
i voted YES but i have a few things to add: as a beginner,i m sure that RM Wiki would help ma a lot. if i will contribute? well, i would be happy to, but i m not sure that at the moment i am able to. I think people like me have the time but don t have the knowledge and when they have the knowledge they don t have so much time anymore. 

just as a suggestion, if the Wiki project supposes too much work and there is no time, maybe more relevant information posted in other topics can be moved in the "getting started" section (if considered necessary). Any way, i really think that Wiki would be a great idea and this may attract more people into using RapidMiner.

mihai

steffen

Thanks for participating, Mihai

I think people like me have the time but don t have the knowledge and when they have the knowledge they don t have so much time anymore.

Yes, I am afraid that this is correct. But (in the opposite) if the experienced users have the time to ask questions and discuss the answers, they also should have the time to contribute, dont they ?

steffen

push 

To any visitors and non-voters: Noone is checking whether you voted for  "I'd like to but honestly I have no time" and so noone will refuse to help you just because you voted for that option.

no votes = no clear picture = no use

...we all know the difference between absolute and relative numbers

regards,

Steffen

haddock

Hi Steffen,

I voted - bet you can't tell how 

Firstly, you are right about the need for a Wiki, and right also to nag us about it. The problem that I see is the usual Wiki dilemma about open authorship. Some folks, like yourself, patently have good and detailed knowledge about the techniques coded up in RM, others, like myself, do not realise how little they know, but are nevertheless quite prepared to offer their advice, no matter how ludicrous. Like Groucho Marx, I wouldn't read a Wiki that I was allowed to write in. Then there are the questioners, quite often too lazy to read the documentation, let alone a Wiki, and of course the bug hunters, and so on, and so on... Sadly, Steffen, there aren't enough like you to go around. :-\

As RM takes over the known universe this problem will not get smaller, so what is to be done? Personally I think the list based structure of this forum contributes to the problem, and that a tree based structure should be adopted. The only way for a newbie to find an answer is to use the search function, but newbies don't know the vocabulary, so it is easier to type in the question, even if it has already been answered many times.

It is also pretty obvious that the documentation needs a re-vamp, when some operators have no documentation at all it is a bit difficult to reply with some circumlocution for RTFM ( "read the f***ing manual" for those from more genteel backgrounds ). I think what is missing is some form of idiot's guide to finding what isn't there, closely coupled to examples, and a curious smiley, perhaps with an "RTFM" balloon, to make the forum less clogged up.

steffen

/push  >:(

steffen

/last push

TobiasMalbrecht

Hey,

.. there are two more votes, unfortunately not enlarging the number of potential contributors but at least showing that there are more thinking of a Wiki as useful! Nevertheless, thanks to you, Steffen, for nagging us about setting up a Wiki. That is definitely on our todo list. I hope we will have time for that in one of the next couple of weeks. So please be patient .. and keep on nagging!

Cheers,
Tobias

keith

I have no idea if this would work, but the forum discussions often contain the nuggets of what a good wiki entry would be, but needs polishing up to be more readable and concise.  Maybe there would be a way to convert a discussion thread into a basic wiki article.  Then those who have the time but not the expertise could edit it into a readable article, basing the technical content on the expert's explanation in the forum.

steffen

keith wrote:
Maybe there would be a way to convert a discussion thread into a basic wiki article.

Yes, manually . But you are correct, this  should be the primary focus. After a problem has been solved it should result in a new article or modification of an existing article. This is not always possible, but the first step should cover the FAQ - stuff.

Tobias Malbrecht wrote:
.. there are two more votes, unfortunately not enlarging the number of potential contributors but at least showing that there are more thinking of a Wiki as useful!

Thanks. Wohoo, we are getting nearer and nearer to some sort of statistical significance.

haddock wrote:

As RM takes over the known universe this problem will not get smaller, so what is to be done? Personally I think the list based structure of this forum contributes to the problem, and that a tree based structure should be adopted. The only way for a newbie to find an answer is to use the search function, but newbies don't know the vocabulary, so it is easier to type in the question, even if it has already been answered many times.

You are correct. But instead of answering the 1000-th question of how to filter attributes, I will just point to the article and offer the option to come back and ask questions if something is unclear. If the article fails to explain certain things, the interaction with the user will tell us why.

The one and only thing I hate about user questions is the stench von lazyness.
"No, I do not want to look in the manual, explain it to me NOW !"

Yes, I am able to differ between "uuh, I do not know that this is called like that" and "Only stupid people use tutorials, explain it to me now, my time is more valuable than yours"

I think the point is clear...I dont know why, but currently I have a hard time to find the right words...

kind regards,

Steffen

keith

steffen wrote:
Yes, manually .

Well yes, of course.  :-)  My thought was more than anything that can be done to reduce the apparent effort or difficulty involved in creating/editing a Wiki entry would help.  A "publish thread to Wiki" button that creates a new article with the thread title, and includes all the thread text might be simplistic, but it might lower the barrier enough to get some wiki content started.

But you are correct, this  should be the primary focus. After a problem has been solved it should result in a new article or modification of an existing article. This is not always possible, but the first step should cover the FAQ - stuff.

In conjunction with that, having a WIki category for something like "Best of the RM forum discussions", prominently displayed on the forum page, might encourage some self-service research prior to posting, too.

steffen

Well yes, of course.  :-)  My thought was more than anything that can be done to reduce the apparent effort or difficulty involved in creating/editing a Wiki entry would help.  A "publish thread to Wiki" button that creates a new article with the thread title, and includes all the thread text might be simplistic, but it might lower the barrier enough to get some wiki content started.

I am convinced, good idea !

In conjunction with that, having a WIki category for something like "Best of the RM forum discussions", prominently displayed on the forum page, might encourage some self-service research prior to posting, too.

Or a "friendly" remark at the new-topic-dialog-page ;D

kgestrich

Hi,

I like wiki's - they are a great help to store and organise knowledge. - So I am happy to contribute.
What put me off is, that there are several Artikels in the Development section - but no content...  :-\

Any content is to start with is better then non - so I am 100% with your idea to integrate a button to  push a threat into the wiki. Perhapse it would make sense to be able to change the WikiName before creating it or to be able to add it to a n existing article.

A last suggestion - can the sign on from the forum not be used in th e wiki as well - I personally hat too many passwords. That way you would be signed on in the wiki as soon as you log into the forum. You could seamless continue to edit the article in the wiki from the thread of the forum.

Klaus