a SoR project for all

[bietz]

Hey guys,

I've noticed that there seems to be a lot of interest in software and algorithms with this project, so I was thinking that Microsoft Robotics Studio (MSRS) wouldn't be a bad idea as a possible platform.
I know it's M$ and all that, but it would allow us to do any kind of advanced algorithms with the processing power of a computer, without caring about the low power of an MCU. And the best part of it would be that anyone could run his own programs in simulation, and then give the same exact code to the one who has the real robot, so that (without any changes) it could work with the real robot.
We could do our own vision algorithms and AI algorithms and test them out in simulation, before doing it on the real robot, and since everything works on the internet with MSRS, when the code executes on the robot, everybody could see its webcam images, and a whole UI on the web (even control it).

It's just an idea, and maybe it's not that close to what SoR usually does, but who knows...

[Asellith]

If code examples are given with the algorithms then they should be in c for both the AVR and the Pic that will cover most the the projects people do here. Those are the primary controllers used. Others exist but the people who use those can easily convert from C. So to benefit the most people I think C examples should be used.

[TrickyNekro]

If code examples are given with the algorithms then they should be in c for both the AVR and the Pic that will cover most the the projects people do here. Those are the primary controllers used. Others exist but the people who use those can easily convert from C. So to benefit the most people I think C examples should be used.

And I someone doesn't know C... I'm in for the project but I only know Basic that far...

I think that pseudo code is the best solution... Someone can program his micro to do whatever, give the starting program but also a pseudo code in order to enable others who don't know that specific language BOTH compiler to work with the whole concept of the book, project, whatever call it...


That's my point of view... but not only because I don't know C...

Lefteris

[bietz]

yeah, and actually MSRS would be kind of painful, because it has a pretty steep learning curve, and its architecture is very different from usual things (it's all about services, concurrency and coordination, which is what robotics really needs).
But MSRS also has a Visual Programming Language (VPL), which allows you to create programs by just dragging and dropping blocks (again, both for simulation and the real robot). So that should be pretty easy to understand, even for those who can't program. (not that we have to use MSRS as the main platform. It could actually also be just a choice)

[airman00]

it will be microcontroller based not computer based
I was thinking of actually printing the book out in paperback. It wouldn't be that expensive , because it would be all black and white and I see a pretty big market for it .

We might post a free pdf version online and print a  paperback version , because many people ( including me ) would prefer to have  a hard copy  rather than an online version.

As for who gets paid , I'm not sure how that will work. So for now :anyone that submits their article is agreeing that they will not be upset and will not argue if they don't get any money from it. The reason i want to do this project is because it would be pretty cool to have a robotics book out there with an article that I wrote in there. I'm pretty sure that's the reason for the rest of you guys writing an article.  So remember , whoever submits an article , you aren't doing it for the money.

Deadline is the first of August for all articles to be submitted.

Here is how an article should be written:
1. Short summary of what the algorithm does, Purpose of the algorithm
2. How the algorithm works in regular english
3. Pseudo code of the algorithm
4. Sample circuit schematic to support algorithm on  both a PIC and on an AVR
5. Actual code in BASIC and in C

We will need to choose the standard PIC and AVR that will be used in the sample circuit schematic.

Also, we should include both  BASIC and C code. For those who don't know C or for those who don't know BASIC , just post the code you do know and then someone here on the forum will translate your BASIC code or maybe just your pseudo code into C.

[benji]

superchicku,, i cant do both image processing and pathfinding thats too much
ill just do A* pathfinding and yea dikstra's algorithm is a part of A* ,it will be included in my article.

i think that algorithms could be mcu based or computer based,as the algorithm can be extended and that would need more processing power.
i think a psuedo code is GREAT
i dont support writing a C code or BASIC code as i told before, implementing can be done in different ways.

a psuedo code is enough, i think,but adding some C or basic code wont harm.

about MSRS i dont think thats a good idea as long as this is not widely used, we want a book that every beginner can benefit from.

about making money out of it ,true airman, everyone should be doin this not for money,free book is better.
 also , i dont think its appropriate to let other people make money out of it while we are the ones who did it,,right?

[bietz]

it will be microcontroller based not computer based

Well, MSRS doesn't need computer-based robots: most robots used with MSRS have a microcontroller and basically have a program that continuously sends sensor data to the computer and receives back motor data (through bluetooth or wifi, for instance).

But I agree that MSRS is certainly a bad idea for the book, since it's not the purpose, but the robot could also have an msrs program for the microcontroller which would allow to use MSRS instead. Having MSRS programs for algorithms would allow really anybody to try them out, even if they can't buy real hardware (simulation is there), but again, this is probably not the purpose of all this. It could just be an addition.

[Kohanbash]

i dont think its appropriate to let other people make money out of it while we are the ones who did it,,right?

It takes a lot of work to go from the writings to the book...
I think an agreement that if the book makes more than $4000 (or some other number)  profit than we split profits between authors would be reasonable.

[airman00]

heres what I plan to do:
review everyones article
Those that are accepted will make their way into the book

It will first be a free ebook and then I will try to arrange it to become a paperback book.
Please note that if I invest money into making it a paperback book I will keep the profits ( which I don't expect to be a lot ) If it becomes a lot of money in profits I will then give some money to those who wrote articles.
If someone else wants to join me and invest into making this a book ( that means putting in money and a ton of time to get it published) contact me . I feel that only those who invest time and money into publishing the book deserve to keep the small profit I think will emerge. This paperback will come out after the free ebook.

I ( and hopefully other people will join)will do a ton of work and invest the money that is needed to go from writings to book. If the book is a success and there is a lot of profits than maybe some money will go to the authors. . Authors may or may not get paid in the end , but those who invest money into it have a greater chance of getting paid.

Is everyone cool with that?

[TrickyNekro]

heres what I plan to do:
review everyones article
Those that are accepted will make their way into the book

It will first be a free ebook and then I will try to arrange it to become a paperback book.
Please note that if I invest money into making it a paperback book I will keep the profits ( which I don't expect to be a lot ) If it becomes a lot of money in profits I will then give some money to those who wrote articles.
If someone else wants to join me and invest into making this a book ( that means putting in money and a ton of time to get it published) contact me . I feel that only those who invest time and money into publishing the book deserve to keep the small profit I think will emerge. This paperback will come out after the free ebook.

I ( and hopefully other people will join)will do a ton of work and invest the money that is needed to go from writings to book. If the book is a success and there is a lot of profits than maybe some money will go to the authors. . Authors may or may not get paid in the end , but those who invest money into it have a greater chance of getting paid.

Is everyone cool with that?

Well, to tell the truth... nope...
I think that it's a book about beginners right, then the lowest the price the more people to have the book...
First of all, those who have written at least one article of this book get a hard copy for free (well if postal costs is a problem then we will arrange this...)
Secondly, I know you will run all the efforts about having the book printed and such, but I still like having no profits after all...
You do it because you want to, not to make any profits... Still, the only thing I want is my name to be heard....
And, most importantly... who determines that the profits are small and so... problems even bad hearts will occur....
So...
I think that any price we place should just cover the printing costs, and nothing more...
But If someone insists on placing a bigger price, let that money cover our needs in electronics material...

Also, I feel like all of us must place our work on solid ground, meaning that we FIRST give out information about
how to build the whole project, the cads, the schematics, some other pictures and so on and then the algorithms...
That way the reader knows what he reads... and why is that so... And most tutorials on the forum are based on the same concept like
the 50$ robot...

Thirdly... I know it's much more difficult to accomplish, but it should be nice for all of us how are building the projects on the book
base our works on the same electronics... Cases may and should vary in order to give ideas to readers... but microcontrollers  and
on board equipment no...
So... we give the basic schematics of our Main Processing Unit and we are to go....
If PICs are used then we should give details... PICs are on the untouchable for many, most of them require programmers and
JDM ones don't worl all the time fine...
If we are to use Atmel due to their nature and probable use of ATmega series we should also give schematics... and how we program them...
We may also include some JDM programmers as well...
And given all we have access to a certain microcontroller we all base our projects there...
I vote for ATMEGA8535...
If other than basic 8bit RISC Atmel controllers are used we also give schematics...
And like that... if someone got what I said...


We can't base our book only in algorithms... readers need feedback...
Think as a noob reader after all...


All these require much effort, but at least I'm in...

Lefteris

[Asellith]

I thought this was an algorithm reference manual sorta thing not a newbie 50 projects to do with robotics book? I thought this was going to be more like a standard programming language reference manual. Just page after page of algorithm examples and math calculations with short descriptions. That way when I need to implement A* into my robot I go to the A* section look over the math and pseudo code. Then use that to start programing the module I want for my robot. I don't want to know how you tested it or what electronics you used. (maybe a link to a tutorial on the robot you built to test it) But the book should just be about the algorithms and implementation on those. Or at least that was my understanding

[TrickyNekro]

I thought this was an algorithm reference manual sorta thing not a newbie 50 projects to do with robotics book? I thought this was going to be more like a standard programming language reference manual. Just page after page of algorithm examples and math calculations with short descriptions. That way when I need to implement A* into my robot I go to the A* section look over the math and pseudo code. Then use that to start programing the module I want for my robot. I don't want to know how you tested it or what electronics you used. (maybe a link to a tutorial on the robot you built to test it) But the book should just be about the algorithms and implementation on those. Or at least that was my understanding

It should be both.... or at least this is my part... you can't do precise gripping without a gripper...

[airman00]

It will be an algorithms manual but like TrickyNekro said , you need to explain a little about grippers if your article deals with them

The whole profits issue can be solved easily
I and whoever else wants to join, will invest a lot of money and time printing the book into paperback.

The money invested by me would total a couple hundred dollars.
Then we would charge a slight profit on each book , so that the profits would pay back the investors and maybe some money for each author. 
Each author would receive a copy of the book for free ( included in the money that I invest)

If we see that there ends up being a lot of profits we can make a Robot Fund. Every month or two a certain project tutorial ( which is chosen by vote) will be funded by that extra profit. Or maybe funding could be given in steps for multiple projects that require money ( flying robots , anyone?)

Everyones names will be included in the book. I'm not doing this for the money, just for the glory.

The standard robot reference for each article is genius ! The robot does not necessarily have to have a step by step tutorial for it , or even have to exist. All it has to be is an abstract robot platform  . Then each article would use a certain CAD and just modify it for the articles needs ( add sensors , etc.)

I could design such a robot , but I would use sketchup since it is free. We just need to decide on specs  ( the shape , dimensions , etc. )
 
I would make multiple versions of it  ( using servos , using DC motor , using legs, etc.) Each author would merely have to add on sensors to it.

[airman00]

also the algorithms need to be geared to what I beginner can use ( not heavy stuff like FFT     )

A* is a bit complex but it has applications for beginners

Dictionary of Algorithms - http://www.nist.gov/dads/  ( thanks pomprocker)

[Asellith]

Umm I know we are all cool here and I am not at this time planing to contribute anything but my 2 cents so here it goes. With the author/publisher arrangements being as fluid as they seem to be now a legal document of some sort should be created and signed by all parties to save everyone hassle later. Particularly you Airman because if your running things and this thing takes off (you never know what will happen) and you make a lot of money then the authors need to know in specific terms what they are entitled to. I would recommend that even between friends and what we are talking about here is as it stands now a loss collaboration between people who have and most likely will never meet in person. I just don't want this to come back and bite anyone and prevent something good from happening when proper communication and some terms laid out on paper with a signature will make things go so much smoother.

Also check out www.cafepress.com

this will let you print a book as you want. You can compile the book and set the markup to $0 so it just costs the buyer the cost of production. Plus the bit that cafepress keeps. so for 0 money down you can make the book available. They come in different sizes and bindings. Could also produce a companion CD to go with it for more money that would have a custom logo and all the source code/pictures and other data already on the disk.

[superchiku]

1st of all please dont make it commercial at the  moment it seems nobody reads my post as i said earlier all the beginner wants is to learn and having 0 knowledge of anything he starts out and what does he find , there are so many books out there costing him $$$ but he doesnt know what to choose so if we keep it free for the time being, more beginners will take it as their 1st choice coz its FREE and then when the book bcomes popular we can attach our trademarks and make it buyable .

As for the algorithms and all i would like to mention something so many ppl here talking abt them but essentially this book is going to be for beginners so what use are these algorithms for the beginners if they dont know how to write embedded C for the specific microcontrollers they use, so i suggest that i submit some tutorials abt AVR programming and someone else on PIC since these are the 2 most widely used controllers and these be included on the beginning of the book so that the begiinner knows how and where to start , then we can arrange the algorithms in increasing order of their difficulty and their respective pseudocodes .

If my idea is ok then lets work on that as it will be more beginner friendly and ill be more than happy writing some basic avr programming tutorials for the book which i have been wanting to do so far..

[airman00]

there will be a free ebook version online  and a paperback version

a legal document is needed so we dont have any confusion , who will prepare it ?

I like my idea of charging maybe ~ 1 dollar extra per book and then maybe ~50 cents of that will be split among authors based on how many articles they write and the other 50 cents goes into a tutorial contest fund. Basically the 200 dollar robot tutorial but like every month or two

a small tutorial in the beginning of the book for using AVR and PIC is not really needed, there are sooo many online  ( one of which I even wrote  )

[TrickyNekro]

I totally agree...

The matter fact is that Internet is not a good place for such arrangements anyway...
Things are fluid yes.. every fluid... so I suggest writing the tutorials just here in the members area...
And then If we all agree make a book...

For example...
You see admin here encourage us to write
tutorials by contests and I'm joining every time for the fun and maybe the glory...
But there are advertisements here on the forum... meaning that the more people we attract by
our projects the more money he get... But of course the forum is well maintained so I don't mind
that at all....


My point is that noone is doing anything for the soul of his very one mother...
We all pursue something with our job and time... For one is glory, for others
personal satisfaction of being thanked, for others................ etc etc etc.....


So if we are to publish a book, we should make several bank accounts with all
earnings being shared to any author.... Legalizing maybe a problem cause
we are separated far each other...




But first things first, let's get down writing good tutorials...
And as soon as we are finished we can see farer...

For now... let's get down to work...

Lefteris

[paulstreats]

My points of view:

1) the book should be a book of algoythms and not complete pojects (were not interested in a gripper, just the algorythms that are used. Do they have a name? surely if you have a good algorythm and progress it, it can be used for more things than just a gripper. Controlling a gripper is a program and not an algorythm so work out what algorythms are working in the program then give them a name say what they do and submit them).

2)If somebody wnts to print a book then fair play. Keep the profits. As long as people get credit for their work.

3)My whole thoughts are a book that is continuously added to meaning there is no submission date. But if I come up with an algorythm in the future that can compare 2 images from a camera and extract rich complex object data then how do I share it if the cookbook submission date has pssed and no more entries get added to it?

4)Rather than specifying microcontroller types etc.. whay not just have a standard naming convention so if an algorythm on a PIC is comparing the value on PORTAbits.A0. Write it as analog_port_0 so if you are using an avr, you can easily understand what it is meaning (make them generic rather than pltform aimed). You would probably find that provided they are done in ANSI compliant C, then it isnt necessarry to aim them at a particular processor, The algorythm should work on all (look at included algorythms with the C language to make clear what a good algorythm is. Take the ITOA algorythm for example its included in nearly all distros of the C language. It doesnt change in structure or content but it works on all compilers and for all platforms)

5)Its nice to see people wanting to do high level stuff but also remember that basic algorythms also want to be included such as basic trig. maybe a data sorting algorythm (once youve stored your sensor data, you need an algorythm to sort the data and search it for future comparing)

6)Also the question over pc based/ microcontroller based targetting was suggested. Its important to realise that it was only 15years ago(at the most) that pc's were running on 20mhz processors running 3d computer games etc... Microcontrollers re slow compared to modern computers but are you really expecting them to do everything that your pc does now? (think about how much processing power is used by the operating system alone not to mention how many threads they must be running)

[Asellith]

The "legal document" doesn't have to be much. It is just a paper detailing out the terms and conditions of submitting an article for publication. Anyone could write it just something that can be mailed or emailed and faxed back that has a signature on it before an article is used. Its just a formality that covers everyone from misunderstanding.

[airman00]

The "legal document" doesn't have to be much. It is just a paper detailing out the terms and conditions of submitting an article for publication. Anyone could write it just something that can be mailed or emailed and faxed back that has a signature on it before an article is used. Its just a formality that covers everyone from misunderstanding.

yep

who wants to write it up?

[TrickyNekro]

My points of view:

1) the book should be a book of algoythms and not complete pojects (were not interested in a gripper, just the algorythms that are used. Do they have a name? surely if you have a good algorythm and progress it, it can be used for more things than just a gripper. Controlling a gripper is a program and not an algorythm so work out what algorythms are working in the program then give them a name say what they do and submit them).

Man we want to help noobs or just to show off with our algorithms...
No... we want to help people...
For instance... What is a Microcontroller is a e - book from parallax...
It's highly impossible to read it and not learn at least something helpful or even advanced...
The thing is we want to help... and that's my point for everyone... everyone...

Not just algorithms and math types... Pictures, complete projects and so on...
We can find algorithms on the net... complex ones, about almost everything and such...
But it's even impossible to find good completed projects that cleanly describe what the micro is doing
and help the reader learn...

It's the duty of the most advanced people like us to help the others who don't know...
And we don't help noobs with some lines... we help with pictures, we help with many
examples, we help with our troubles reversed... and so on...

At least, that's what I'm gonna do... My style of writing is this... I won't oblige anyone to write this way...

Hope I get as much support as I can't for this point of view...

And as I said, first things first... Let us write our tutorials here on the forum and then
let us speak about the book....

Lefteris

[airman00]

But it's even impossible to find good completed projects that cleanly describe what the micro is doing
and help the reader learn...

Exactly, thats the whole point of this project.

but to do this we need to adopt my idea of a standard robot platform  - I can make a 3D one on sketchup but I'll do it next week ( crazy busy now)  what do you guys think of this universal abstract robot platform idea?

uploading it to the members tutorial first is also fine

TrikyNekro understands what this is all about!

[benji]

1/about this legal paper,,
we are from different nationalities and countries,
 man i dont think that paper means anything,,or how is it the things go by?

2/i think this is algorithms book, not a book for a total noob,,

3/do include as many languages as you can, but a pseudo code is a must.

4/internet is fullllllllllll of noob tutors, we need somthing alittle more advanced

5/this book can be updated each year, he first of august so people can add to their algorithms, open new sections,
maybe the book can be extended to more than just algorithms book

6/airman00 , i guess your the person to manage all this, make a post like the 200$ robot contest post to let only the people who wants to write in the book post there with their article sunject

7/about money,,,, i think let the book comes free pdf , the printed version can be sold
but if the profit comes more than a specific number then the authors should be paid too.

[benji]

again people,, for noobs there are plenty of stuff online and they wont pay a dime for a book, or even bother to download it
the internet is a huge source of information for noobs so the book would be just another book
my idea was to make somthing BIG,, not another noob book
noobs have SoR to learn from, actually noobs would like to ask questions in the forum more than reading a book.

i say let the book be a little more advanced than for noobs,, lets get it into a higher level.

the tutorial in this website are for noobs, so let the book be somthing more advanced

[airman00]

the book should be aimed for a person with some background ( the knowledge level of someone who built the equivalent of the 50 dollar robot )

as for the post with all the articles ,ill make that tonight'

anyone want to write a demo article?

[benji]

great , so the points that are set till now are

1/ Algorithms book (for people with backgrounds ,at least read the 50$ tutorial)

2/first of august is the last date to send your articles to airman00

3/psuedo code is a MUST


the things we should agree about:

1/the whole money thing

2/the file extention to send the article (office word?, wordpad?)

Any other stuff i forgot?

[airman00]

upload your pages to the members tutorials  , so then we have universal format. then I'll download it from there and compile it all into .pdf format.

I spoke with Admin and he said that we should post all articles in the members tutorials under the Topic " SoR Book "
from the extra revenue that Admin gets from more hits to SoR for that ebook pages , Admin will hold more robot contests which will end up including articles as entries

do you all agree about the whole standard abstract robot idea?


the money issue we'll worry about later , after everyone submits their article

I'll add in all these guidelines into the main page for SoR at the members pages once we are sure about all the points

[benji]

come on folks lets knock it hard,, we want a bomb book....

by the way,, i have 2 things to do now for SoR,,

1/the 200$ contest (hacking optical mouse)

2/Algorithm book  (A*)

yeeeeeeeeeeeeehaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

[airman00]

yep everyone aim to submit their articles by August 1st

I will be emailing reminders to those who pledged articles every month to remind them to write it

How many articles are we aiming for in total?