New to maya Api

[mona777]

Hello Rob,

I'm so very new to all this and started reading about Maya API plugins last week. I downloaded your Maya plugins folder but can't run any of the programs. I've also tried running programs from http://download.autodesk.com/us/maya/2009h...pp-example.html but couldn't. The problem seems to be the Maya libraries.

I have found all the libraries such as maya/MIOStream.h and maya/MString.h in C:\Program Files\Autodesk\Maya2008\include\maya. In visual studio (2005) I went to "projects->Properties->C/C++->Additional Include Directories" and put "C:\Program Files\Autodesk\Maya2008\include" and "C:\Program Files\Autodesk\Maya2008\include\Maya" which didn't make any difference then I went to "projects->properties->Linker->General->Additional Library Directories" and again put "C:\Program Files\Autodesk\Maya2008\include\maya" which again didn't make any difference and I get compiler/linking errors such as fatal error C1083: Cannot open include file: 'maya/MIOStream.h': No such file or directory. Will you please guide me as to what I am doing wrong?

Many thanks,
Mona

[rob-bateman]

QUOTE(Mona777 @ 02/23/09, 08:42 AM) [snapback]301485[/snapback]
I'm so very new to all this and started reading about Maya API plugins last week. I downloaded your Maya plugins folder but can't run any of the programs.

I haven't touched that since 2004 - so yes, they won't work until you re-compile them.

QUOTE
have found all the libraries such as maya/MIOStream.h and maya/MString.h

To be honest, the Maya API is not somewhere you want to be learning the basics of C++ in. Learn C++ first, then come back to the API. Jumping into the API without knowledge of C++ is a complete waste of your time.

QUOTE
In visual studio (2005) I went to "projects->Properties->C/C++->Additional Include Directories" and put "C:\Program Files\Autodesk\Maya2008\include" and "C:\Program Files\Autodesk\Maya2008\include\Maya" which didn't make any difference

Sorry, but again - this is a trivial C++ problem. How can it find maya/MIOStream.h if you've removed the maya/ from the search path....? i.e. it should be:

"C:\Program Files\Autodesk\Maya2008\include"

QUOTE
then I went to "projects->properties->Linker->General->Additional Library Directories" and again put "C:\Program Files\Autodesk\Maya2008\include\maya" which again didn't make any difference and I get compiler/linking errors such as fatal error C1083: Cannot open include file: 'maya/MIOStream.h': No such file or directory. Will you please guide me as to what I am doing wrong?

And again, this is a trivial C++ problem. Now, I try to be as polite and as helpful as possible, but if you don't know the difference between a library and a header file, there really isn't much I can do to help you I'm afraid. (I'll give you a hint, one ends with .h, the other ends with .lib). Learn C++. It's not something you can do when learning the API. It's something you have to do beforehand.

I'm sorry if this reply is a little bit abrupt, and not as helpful as you were hoping.... but unfortunately, it's brutally honest.

[mona777]

I had done what you said but it didn't work: How can it find maya/MIOStream.h if you've removed the maya/ from the search path....? i.e. it should be: "C:\Program Files\Autodesk\Maya2008\include"

I appreciate your honesty, I am self taught and have gaps in my knowledge in C++. However I am in a situation that I really have to do this and am putting all the hours that I can into it. A teacher would want to teach someone who is willing to learn. I will learn this regardless but I had assumed that forums where for this kind of thing.

Thank you for your time

[rob-bateman]

QUOTE(Mona777 @ 02/23/09, 09:35 AM) [snapback]301492[/snapback]
I had done what you said but it didn't work: How can it find maya/MIOStream.h if you've removed the maya/ from the search path....? i.e. it should be: "C:\Program Files\Autodesk\Maya2008\include"

There are only 4 possibilities.

1. You don't have Maya installed, and the path "C:\Program Files\Autodesk\Maya2008\include" doesn't exist.
2. You installed Maya, but didn't install the SDK (in which case "C:\Program Files\Autodesk\Maya2008\include" doesn't exist)
3. You have 64bit windows, and have installed the 32bit version of maya and infact should have directed it to the 0x86 program files folder.
4. You haven't set it in VC++ correctly.

If i had to guess, I'd say the problem is 3 or 4.

QUOTE(Mona777 @ 02/23/09, 09:35 AM) [snapback]301492[/snapback]
I appreciate your honesty, I am self taught and have gaps in my knowledge in C++. However I am in a situation that I really have to do this and am putting all the hours that I can into it.

Work smarter, not harder. The first gaps you have to fill are how to link libraries, how to set include paths, and how to write DLL projects. If you don't spend time learning that, then it doesn't matter how hard you work - you're not going to get any further.

QUOTE(Mona777 @ 02/23/09, 09:35 AM) [snapback]301492[/snapback]
A teacher would want to teach someone who is willing to learn.

A teacher is someone who wants to teach. Most people don't want to teach. Q.E.D. most people are not teachers.

However - most people are genuinely nice people, and will lend a hand once in a while. But don't confuse lending a hand with teaching, because they are 2 very different things indeed.

QUOTE(Mona777 @ 02/23/09, 09:35 AM) [snapback]301492[/snapback]
I will learn this regardless but I had assumed that forums where for this kind of thing.

No. These forums are to provide help and assistance to those who are stuck, from those who wish to help. If you learn something in that process that's an entirely unrelated matter. If tuition is that you are after, you're not likely to find it here..... (because help != tuition).

[mona777]

I thought you were a teacher at Bournemouth, and being "nice" does not mean being rude! Anyhow your comments were helpful thank you.

[joojaa]

im splitting the topic off from top.

As for me i think rob was very polite. People tend to become no nonsense after while,because despite us being nice we cant be nice for everybody so it counts. Theres not just enough time.

[videohead]

Based on your level of experience, I'd suggest that you use the wizard to get started (I recall that were having problems with it), and later try to understand the project settings, so that you can handcraft a project or even compile a Maya plugin from the command-line outside of visual studio should you wish to do so. Alternatively, you could start from an existing project (.vcproj) or solution (.sln) in Maya's sample plugin directory (C:\Program Files\Alias\Maya7.0\devkit\plug-ins on my machine).

However, if you do not understand C++ you wont get very far. And even if you understand C++, you still need to understand the philosophy of the Maya API. Only then will you find these forums to be of help. Its hard to give any other guidance other than "go read the manual", when it appears that there is so much you still need to know. I therefore recommend that in the future, should you decide to post questions, show that you understand how the API works with respect to the problem you're solving, so that it will be easier for the folks here to guide or correct your thinking.

All the best.

[mona777]

Thank you all. Despite all of this, I learnt from you all. I was looking for "new to API" and found this thread. I'll read the Complete Maya Programming and before asking any more questions. I do understand C++ but mainly with openGL but as you have said rightly, I haven't really done a lot of work with the project settings in VS maybe it's because I couldn't find a site on that on the net.

also I had tried to run a program from c:\Program Files\Autodesk\Maya2008\devkit\plug-ins

but got this:

animCubeNode : warning PRJ0009 : Build log could not be opened for writing.
Make sure that the file is not open by another process and is not write-protected.
?animCubeNode : error PRJ0007 : Could not create output directory 'c:\program files\autodesk\maya2008\devkit\plug-ins.debug'.
animCubeNode : error PRJ0006 : Could not open the temporary file 'c:\Program Files\Autodesk\Maya2008\devkit\plug-ins\Win32\Debug\RSP00000121682456.rsp'. Make sure the file exists and that the directory is not

As I said thank you all and I won't be asking stuff until I know more

[joojaa]

Well i dot think reading complete maya programming is best use of your time. The book in general is not very good and reads more like the refebnrces of maya documentation with little sugar on top. But worst it doesn't really tell you very much about what and how you should do your designs. And some of its examples are just downright poisonous to learning maya.

Its just kindiof assuming that you know not just the syntax oc c++ but also how to desiugn architecture, how to implement things wich is NOT the same as just knowing how to type code. If coding was about thet then it would be easy. But unfortunately it is not.

[videohead]

QUOTE(Joojaa @ 02/24/09, 10:45 AM) [snapback]301575[/snapback]
Well i dot think reading complete maya programming is best use of your time. The book in general is not very good and reads more like the refebnrces of maya documentation with little sugar on top. But worst it doesn't really tell you very much about what and how you should do your designs. And some of its examples are just downright poisonous to learning maya

To add to that: I learned from the API doc. You may wish to go the same route that I did i.e. print the first 7 chapters of the API guide -- everything up to commands and nodes, minus the chapters on shapes and meshes for now, cos they are really hard (I'm only just getting understanding them, besides you want to be as productive as soon as possible). The forum may be of some help with the knottier aspects of the document.

(When I first started to learn it took me two days to realize that the reason why the "Hello Word" wouldn't print was that Maya required fflush statement, else the output would be buffered; besides I wasn't quite sure where to look for the "Hello World", and for a while I kept starring at the MEL window and wondering "why!?" ... ).

BTW, I found the first "complete maya programming" book annoying, and the second to be an even greater disappointment, because it read more like a maths text. That said, it may of help to you. Give it a try, and don't be afraid to turf it, if you find its not helping.

[joojaa]

QUOTE(videohead @ 02/24/09, 04:51 PM) [snapback]301590[/snapback]
BTW, I found the first "complete maya programming" book annoying, and the second to be an even greater disappointment, because it read more like a maths text.

The supprising thing about the book is that it sells at all. I mean its quite a consensus that the book isn't really worth it. Ok so there may be 14 pages in the book that are worth anything. But finding them is pain. And offcousre if you dont know the subject it appears valid. It kind of sidesteps things so to make a fuzzy feeling.

[rob-bateman]

QUOTE(Mona777 @ 02/23/09, 11:33 AM) [snapback]301512[/snapback]
I thought you were a teacher at Bournemouth

That was a very very long time ago.

QUOTE(Mona777 @ 02/23/09, 11:33 AM) [snapback]301512[/snapback]
and being "nice" does not mean being rude!

I'm never rude, but I am brutally honest and don't have that much spare time to answer questions anymore. Your question could (and would) have been answered by anyone viewing these forums. In an average week, I get a large number of e-mails, PM's on forums etc, all asking questions about the API and programming in general. These range from industry professionals, all the way through to students trying to grasp first principles. I try to respond to as many as possible, but due to restrictions on my time (i.e. my job and life in general) I can't answer all of those questions to the full satisfaction of those who ask. As a result, my answers are normally very direct and to the point.

QUOTE
Anyhow your comments were helpful thank you.

And they always will be helpful, but don't expect my answers be in the form of tuition.

[flaiver]

QUOTE(Joojaa @ 02/24/09, 07:33 AM) [snapback]301593[/snapback]
The supprising thing about the book is that it sells at all. I mean its quite a consensus that the book isn't really worth it.

mhmm.... I am not your opinion. Currently, I am using the maya api for something like 5 years and I think, that these books can be a good starting point to the maya api if you:

a ) have allready good experience using Maya in gernell (using nodes is the key)
b ) have good knowkedge in C++
c ) willing getting further information about the maya Api (like this forum)

I think, both gould books are good for user, who have little experience with design pattern (and aren't interested to learn it ). Besides this, they show WHERE to start - in my opinion, that's the only fact what's messing in the maya docs (even all the rest is just great documented). But you are right: these books can let people think, MPxCommand is the way to go, whats wrong. But I think, these people aren't really willing to learn the maya api. Just my five cents

flaiver

[joojaa]

QUOTE(flaiver @ 03/14/09, 03:00 PM) [snapback]302999[/snapback]
I think, both gould books are good for user, who have little experience with design pattern (and aren't interested to learn it ).

Yes but also the lack of this si why they have so horrible problems going against the grain.
I mean often you see people asking best way but then it turns out they dont want to. Explaining design up font saves a lot of issues. One can waste years of your life in the pursuit of avoiding this but sure as hell if you keep developping YOU WILL ENCOUNTER THIS.
QUOTE
Besides this, they show WHERE to start - in my opinion, that's the only fact what's messing in the maya docs (even all the rest is just great documented).

Ok, even so dont you think its a bit overkill for where to start. II mean half of the examples in the book are actually doing more harm than good. Anyway i dont really think it points out wher to start wery good. IT neglects pointing out you need to KNOW maya.

But i actually think the book is very bad for where to start.

I think you could probably squeze all this knowlege in one article. My point is that the book is BAD because. It tells you what you want to hear, wich is allmost correct but SOOO incorrect in the end. The correct answer is sometimes painful. Painful or not right is right.

[noto]

I'm starting out with the API also.

I have seen these kind of comments on Gould's book(s) before, but they are diffuse in their criticism.

It'd be great if the wise API folk here could detail these criticisms, perhaps done here as a final place that could be pointed at for future reference?

Write the criticisms technically and explicitly, not in a lay form (as for a API beginner) - I think this is more valuable than watered down versions. The meaning of these techincal criticism may not be immediately apparent to people like myself, but they will become clearer and of greater help with experience.

Many thanks.

[joojaa]

QUOTE(noto @ 04/03/09, 08:27 AM) [snapback]304790[/snapback]
The meaning of these techincal criticism may not be immediately apparent to people like myself, but they will become clearer and of greater help with experience.

If you mean specially code, then theres nothing worng with the code, just that nobody would wirte that code most of the time.

(mostly on the 1 book sincce i own it but dont ahve it at hand, flipped trough the second for a few days and said utter crap never bothered to buy it)

1. Gould book wastes a lot of pages on being a reference guide. We dont really need a reference we allready have one thats more technically accurate then the one in the book will ever be. The manual is for this purpose, its freely available and its the one you should follow.

Technically this means your being fooled by a lot of text into thinking thet theres something valuable there. Text that you allready have. Sometimes this may be a good thing it enables you to check the clarification of things, except most of the time the book is doing things in way that straight forward and easy to conceptually understand. Im not sitting next to the book so i cant refeer to it but it quite often shows stuff done flat out direct loop. While nobody would actually do so. They would inject a few nodes to do the same thing.

1. Gould books tell you how to do the straightforward things. It asumes people know what to do. Allmost all of the examples are watered down to something you'd be expected to know anyway. Well the real problem is that while you need theese simple examples to check that your progression works. But come on, how many of you would actually go out and make a node for rolling things along one dimension? So the book needs some parctically useful examples. THe closest one to even a remotely practical thing in the first book is the melting deformer. (not muvch better things in the second)

Technically this mean he teaches all you need to know EXCEPT the hardest part. Namely how you come to the concluison that this is where youd want to be. He does illuminating a path of development in the different posts examples. But the thing is why would i bother to do this? Maya allready has this functionality built in, as a combination of 2 nodes. Anyway the path he takes is totaly retarded. He never actually draws any attention to the fact that nobody in their sane mind does commands like this.

So what he should be showing is somekind of test driven development cycle instead. What i would do is id show how to make a node that implements a different metric for extrusion (while i don't believe mayas extrude is broken i know it can be thought out differently with different results), where a modeller has supplied a number of cases that need to work out if possible , and you develop against that.

1. Gould books doest seem tell you anything about how to maintain your nodes. And how to progress out of the development phase into use and recompile upgrade without breaking everything up. While true its not necessarily in the scope of the book it would be nice if he could mention the few rudimentary tricks may has up its sleeve.

2. The books doest explain how to write good AE templates and how to tie up your code to the maya GUI. This is off course what youd want to do most times.

3. It would be NICE if he could spend some time on making shading nodes that would be mental ray compatible. Off course the book is old but still. This si also something most people would WANT to do. More on shading nodes also.

4. The book doesn't clarify how maya users work. This might not be a big thing for some people. But id you'd ask me to bet d say that AT LEAST 1/2 of all developers who would do api work would be ordinary programmers with no maya training. And they would be much more likely to read a book of this kind. So at MINIMUM there should be a explanation for them on how its organized. They Do understand all sorts of programming things so for them 2 pages of very highlevel mumbojumbo would suffice, it would be told up front saving hours of confusion.

5. The book should contain some examples of how maya users pervert the use of nodes to do something never anticipated. This would underline the need for you to be extra careful when designing something, as you might not get rid of it later because a lot of people would be relying son something you didn't think of. It would also enable you to work out strategies that are more useful for maya users. And teach you something valuable of mayas use at the same time, as well as how mayas ecosystem works.

6. How to design something that works. He quite literary does not touch it. Im not certain that one could deeply do this at all. But atleast one should try to start illuminating it somehow.

7. A a better brief clarification on what the api does and what it does NOT do. The symbiotic relationship between mayas scriptig enviroment and the api ad how they do not overlap.

8. Rapid prototyping needs to be explained. Thing is the book is allmost all rapid prototype but its not said it is. Theres usually a wide gap from here to finished product. Sure i understand all your internal code would be of this kind at first but they still need finalization if you work with more than 10 people.

9. The book needs a new revision. Its too old. Has no way o helping you when it comes to constraints etc...

10. There should be more on advanced topics which cause a lot of trouble, form declaring more complex data into. How you build in a connection to existing nodal structures based on the node reference, and what the caveats at that is. So you can maximize your and users benefit form your nodes.

[rob-bateman]

Only got volume 2 to hand....

1. The first 7 chapters are elementary math notes. That's 120 pages of math. I'm sorry, but I can't imagine too many people who want to write plug-ins for a 3D app, do not know what a point is. Your average reader is going to skip every single one of the first 7 chapters. That leaves you with 4 chapters, Polygons, Nurbs, Subdivs, and selection contexts.

2. The Nurbs chapter. Once upon a time, about 10 years ago, these were all the rage. These days they are not that important. He spends 30 pages explaining the maths behind nurbs, and then the rest of the chapter doesn't provide you with any more information than the docs for MFnNurbsSurface and MFnNurbsCurve. He then proceeds to demonstrate how to make a curve in mel, and how to do it the API. Its 150 pages that you'll never find a use for.

3. The subdivs chapter. Same problems as with the Nurbs chapter.

4. The Polygon chapter. Same problems as with the Nurbs chapter.

5. There is literally no mention of error checking in either book.

6. MPxCommand this, MPxCommand that, and let's throw in another MPxCommand. The molecule example (8.3.3 in Vol2) implements a plug-in based poly modifier as a command, which is utterly retarded. He then spends the next 145 pages rectifying the source code through 5 different revisions until he finally provides a completely different example that does it properly. You have to read 100 pages of repeated crap until he finally mentions construction history at all (which should have been on the first page!). It's an incredibly confused way of explaining one of the most core aspects of the Maya API.

7. What really gets me is that the examples are lengthy, and largely pointless. What he ends up doing, is writing lots of pages of code about things that really have nothing to do with Maya at all. There are literally hundreds of pages of code that assumes the reader is a fucking idiot: (and they are - they were stupid enough to buy the book after all)

CODE
// this is an array of X coordinates.
float $xcoords[];

// this is an array of Y coordinates.
float $ycoords[];

// Use a for loop for 360 values....
// $i here, will start at zero and move all the way to 360 incrementing by 1 each time.
// a for loop is a programming thing that allows you to repeat stuff.
for($i=0;$i<360;++$i)
{

  // Now use the sine of $i.
  // Did you know that the sine($i) is a command?
  // And it's also really useful?
  float $sa = sin($i);

  // And now, here i call cosine.
  // That's also a mel command you know.
  // and it's also really useful.
  // it's a bit like sine, but not quite.
  float $ca = cos($i);

  // here I shall use the sine value we computed previously,
  // and I shall assign it to an element in the x coord array
  $xcoords[$i] = $sa;

  // here I shall use the cosine value we computed previously,
  // and I shall assign it to an element in the y coord array
  $ycoords[$i] = $ca;
}

I'm sorry, but those code examples are just full of guff, and have very little to do with the Maya API or mel. They are just insulting to the level of intelligence of the people he thinks are going to be reading the book.

1. The contexts chapter. This is another slightly insulting aspect. The chapter is lifted almost word for word from his very first book. The source code examples are exactly the same as the ones in his very first book.

2. By far the biggest problem is not so much what he's included, but what he has failed to include. Particles? Deformations? Exporters? MR Shaders? Fluids? Cloth? etc etc.

btw: If you do have both of the gould books, they do make quite a good stand for a laptop....

[joojaa]

Deformations are in the first book. But not that the example is very exciting tough it is the best of them. One point on nurbs, they are used but not so much in maya, and anyone expecting to use maya. The biiggest problem taht even with theese 2 books you couldnt fill one volume up with the info.

[noto]

Excellent, thankyou both very much.

If you feel you could field a couple of other questions, I'd be greatful:

• Do you believe the Alias/Autodesk dev kit examples are good illustrations of sound API useage? (you'd hope so..., but you never know who is writing the docs and their relationship with dev).

• I will be thoroughly investigating 1. Docs (inc. white paper), 2. Here (inc. the forum sticky, revisting it again and again), and 3. Rob's site (many thanks). Are there any other resources you'd recommend? (given the lack of [reliable] published material).

• If you were starting out again, what would you have wished you knew? (there is a nice quote by someone along the lines of: 'if you knew you were going to have amnesia next week, what would you write now in three paragraphs to best capture the knowledge you'd want to pass on to yourself learning from scratch' ...if you could do this for explicit amnesia of your API knowledge?

Much appreciation.

[joojaa]

QUOTE(noto @ 04/04/09, 08:04 AM) [snapback]304915[/snapback]
'if you knew you were going to have amnesia next week, what would you write now in three paragraphs to best capture the knowledge you'd want to pass on to yourself learning from scratch' ...if you could do this for explicit amnesia of your API knowledge?

The most important thing is not that the code works, the most important thing is that it maintains and extends easily. You can count on somebody coming and saying 2 days before the deadline that's cool now can you make it do this? If you cant its seen as utter crap. Incidentally people will ask for a lot of changes they will never ever use. But if you just make something that works buts hard to maintain youll find it eats up all your time from other things.

Quite interestingly once you roll out something its never used as you intended. Users can always find drawbacks in your systems that were incomprehensible when you started out. On the other hand they find your system is great because of the flaws you have in it. So you cant fix your flaws easily without braking a sizable populations greatest tool ever.

Make it simple. If it does one thing well ist far more usefull than otherwise. For some reason you can think of many things one could extend it to, but its better if you instead strip it to bare minimum. No gui no nothing just soemthing that magically does one thing thats it. People don't seem to want choice, they want something that you can explain they should want. Irrespectively of the fact that its useful or not, on that note make it have shiny buttons.

If you find theese in contradiction with previous paragraphs then think again.*

*take maya python integration it was hailed by most as the best thing on this earth because now you don't need to do all kinds of support functions to do simple stuff. So whats the first thing they do? Write even more support functions for it to be easier. Must have missed something. But aha! It does go in line with point 1 and 2