
Material
You may propose three kinds of contributions :
1) New entries to the Glossary.
2) Examples of real-life applications.
3) Tutorials.
---------
- New entries to the Glossary
- They should indeed be new. Please check out the Glossary before you propose a contribution !
- They should be related to Data Modeling, Statistics, and Probability Theory.
- Text should be no more than 2 screens long.
- Text should be in English or in French. AI ACCESS will translate the text in the "other language" (in case you didn't notice, the site is bilingual.
Well, almost : the French part is OK, the English part is so-so).
- Text should preferably come along with illustrations in a common vector format. Pairs of illustrations than can be profitably dynamically combined are welcome (you'll find many examples in this Glossary). AI ACCESS will reshape the illustrations in its own standard format.
- Examples of real-life applications
- Many of you consultants have used your pet
modeling technique many times for your customers' needs. So your
shelves are certainly loaded with thick reports, and your hard disk
is packed with the same.
This kind of report is exactly what this
site doesn't need. The idea is not to show how hard you worked to get
the final charts and graphs, but rather to convince the visitor of two
things :
*
First, that your pet technique is useful in real life, and generates valuable
results in a reasonable time.
*
Second, that your consulting company is the right choice for anybody who
is serious
about statistics and data modeling, because it is clear that you understand
very well what you are writing about.
This can, and should be accomplished in less than
10 screens. This means some "data compression" work on your part,
but you may also discover that you should have done this very pedagogical
exercise a long time ago.
- Tutorials
Tutorials are a tricky
matter. The temptation is great to simply copy and paste textbooks.
Don't waste even seconds to do that. Remember that this site was first created
to fill in a gap, namely to go as far as possible with explanations only, and
resort to maths only when words won't do anymore. And then, because of the
words, the maths should be easy. I strongly suggest you don't spend any
time on writing a new a tutorial before you produced a couple a crystal clear entries
to the Glossary.
Conditions for a Contribution
- First, what you do not get : I do not offer money for new contributions (the site should be self-supporting, and there is no way the paying section could
generate enough cash buy contributions).
- But I offer something almost as good : your contribution will be acknowledged, and will come with a link to the site and page of your choice. It will look something like :
|

|
|
|
New Entry
|
|
| |
Your text and illustrations are here
|
| |
This entry was contributed by : "Stat-is-Fun" Inc. : www.stat-is-fun.com
|
| |

|
- In addition to the contribution, I ask for a link from your site to my Glossary index page.
It should look something like :
"You may find some additional information
about xxxx by going to http://www.aiaccess.net/e_gm.htm",
although the name of the link does not have to be visible.
- Both my links and your link will stay alive :
- For one (1) year for a contribution to the Glossary.
- For two (2) years for a Tutorial or an account of a real-life application.
- At the end of this period, AI ACCESS will retain ownership of the contribution,
and will be allowed to modify texts and illustrations as it sees fit.
A few tips
- The goal is not to be original or sound savant. The goal is to make it clear to the reader that most of Stats is perfectly understandable if properly explained, and that it is useful.
- A Glossary entry should therefore be mostly "hand-waving Stats", so that concepts are already understood before they are demonstrated (in a later to come Tutorial).
- Focus on "Why was it invented ? To replace what ? For what use ? How does it connect with similar concepts ? What are the alternatives ?". Stay away from the "Definition-Lemma-Theorem-Corollaries" scheme. Ask yourself : "Would a layman understand ?". He should.
- Do not use realistic examples as a means
of introducing a new concept. In my opinion, examples force you to understand
both the example and the concept at the same time, which is bad pedagogy.
"Heads and Tails" are the most realistic example you'll ever find
in this site for introducing a concept. Now, more realistic example are
fine, but they should be deported to a separate section at the end
of the text.
- Contributions do not have to focus on Junior
year topics. Higher level contributions are welcome too: it is my experience
that even researchers or university professors resent the lack of clarity
of many reference texts.
- Avoid "cut-and-paste" contributions, they will be instantly detected and rejected. The very raison d'être of this site is to have a different flavor from your run-of-the-mill textbook or contribution to a technical journal.
- A "real-life application" contribution should do away with all the unnecessary details usually found in such texts. Keep it down to 10
screens at the very most. Focus on why the particular technique was chosen, what were the alternatives, what are the limitations etc...Avoid exasperating short variable names.
Use no real company's or people's names. Better, use no names at all.
- Software screenshots should be avoided, or else carefuly edited. Part of the credibility of the site comes from its independence from any software editor.
What if you are a software editor ? Same conditions apply as to anybody
else.
- Just because you sent a contribution does not mean it will be posted. Some back-and-forth shuttling seems anavoidable anyway.
Ultimately, I reserve the right not to post a proposed contribution.
I hope I'm honest enough not to use a rejected contribution under my own
name, but just to be on the safe side, take some precautions.
- Contrary to what you may think, I'm quite open
to suggestions. I'm seasoned enough to know that I'm not always right. So don't
hesitate to propose new ideas, comment on old ones and, well, even criticize
a bit.
If you're willing to give it a try, there is a small contract ready that will be sent to you upon request.
Show it to your lawyer so he can make it much bigger and incomprehensible (although
in much better English than mine) for a (almost) reasonable fee.
