There’s an old saying in the visual effects industry that you sometimes forget. When you do, you end up paying a terrible price for it.

“Don’t use the first release of any version of a software.”

In the case of Maya 2009 - I can’t agree more. It’s been a frustrating ordeal. I have a list of five or six things that I extremely dislike about it. But here, I will only complain about two of them. Sadly, these are’t really specific to Maya 2009. They’ve probably been like this since version 4 or something around there.

Not being able to write a custom particle source, whether it be cached or just a node that supplies particles is really annoying. I do realize that you can hack your way around this. While it works, it’s just not ideal. Sometimes you need a level of control over your particles before they land in Maya’s particle system. It would be nice if there was a blessed way of doing so.
The entire particle instanced geometry bit is a little messed up. Maya seems to right out a copy of the geometry on disk when instancing in MI files. I don’t know how Maya is constructed internally - but was it not possible to take advantage of the delay loading or at least just instancing using just the transform node when the geo is the same? Sadly, this is just scratching the surface of the many annoyances Maya has with instancing.

Urgh…it’s not been fun couple of months.

1) Convince yourself that Unicode is 16bit Wide Characters
Joel Spoksky wrote article back in 2003 about Unicode. One of the things he says in the article is: “Some people are under the misconception that Unicode is simply a 16-bit code where each character takes 16 bits and therefore there are 65,536 possible characters. This is not, actually, correct. It is the single most common myth about Unicode, so if you thought that, don’t feel bad.”

He’s right you shouldn’t feel bad about if you thought that…back then. But it’s 2008 - and if you still think that, like I did, feel bad. I spent the better part of the last two day days coming to terms with having to deal with applications that would potentially be in languages other than English. Anyway, I had gotten it in my head - for some reason that wchar_t was Unicode.

Joel is also very correct when he says it isn’t that hard. Just keep the phrase “encoding scheme” in your heart.

2) Following #1 - write some code using converting back and forth between std::string and std::wstring
I realize that something like 80% - at least - of the coders out there have no sense of humor combined with the God complex, so I probably shouldn’t admit that I did that. But I want other people to learn from my mistakes. Anyway, pimply programs aside, I also got it in my head that I had to convert between std::string and std::wstring for “representation purposes” on my terminal. This is also false. But feel free to continue to believe it.

If you’re a C/C++ coder and you’ve been thinking about Unicode but haven’te quite jumped in yet. I suggest you write some code that dumps the literal values of the string as hex integers so you can see the encoding scheme. Sure enough you can take the 1, 2, 3 or 4 byte representation in UTF-8 or UTF-16 and look them up in the code charts. This helped me break away from the “indexing” idea that Joel Spolsky was referring to regarding the 16-bit wide characters.

3) Gloss over, but ignore phrases like “encoding scheme” and “code charts”
Encoding schemes are dumb - who needs encoding schemes if you have 65,536 slots? ASCII for life!

The other dumb thing I did was fail to understand the weird implicit relationship between ASCII and UTF-8. ASCII is implicity UTF-8 by design of UTF-8 - not the other way around. ASCII based strings such as “hello, world!” can be represented in UTF-8 without change. This is because the first 7 bits (127 slots) were left intact. Anything beyond that, you start getting into code pages and tables territory. Wikipedia’s artcile on UTF-8 is a bit terse, but it gets the point across if you read enough times.

4) Not know the differences between UTF-8 and UTF-16…and UTF-32
Why do we need so many schemes? Why can’t everyone just learn English? All the good stuff is in English anyway!

To be honest, I’m not entirely sure why there are different encoding schemes. That does not mean I question their worth. Having been exposed to code since I was in 1st grade - I realize there is generally a reason for every design and line of code written. I just may not be exposed to the circumstances to understand why. With that said, the same string can be represented in UTF-8, UTF-16 or UTF-32. They just look very different. I don’t know know the author of this page about Unicode, but they should be given a gold medal.

If you have a tool on your system that allows you look at the character maps, say like “charmap” - I recommend you run it and take a look at the difference indices (is indexes acceptable these days?). It will give you and idea who this stuff is encoded.

Conclusion
Anyway, if you follow these 4 steps and make some of your own bad habits - you’re well on your way to not understanding anything about Unicode.

One more dumb thign I realized. At some point in time I was familiar with UUencode. If you’re from the old school, you’ll know what I’m talking about. If you aren’t - then look it up. Anyway, I’m certainly not saying that Unicode is the same thing - but UUencode was another encoding scheme. For some reason “encoding scheme” just seemed to escape me until about the afternoon today.

I knew a guy in highschool who claimed that he manually UUdecoded a small JPEG from the newsgroups once. I don’t believe him to this day - but if Japanese needs school kids and use a virtual abacus faster than I can use a caculator - I guess anything is possible

ImageWrangler has been at version 1.0 for a long time. Well, actually it hasn’t. There’s a 2.0 beta that some studios have been testing out. What I’m about to say doesn’t directly apply to that particular build. So lets move on…

Anyway, sometime ago, I decided that I had enough of wxWidgets. I can’t remember the entire list of reason right now, but one of my gripes was the way the tree control functioned. We had derived the tree controller to get the look and feel of our current one. However, if you’ve noticed - there are some annoyances with it. From the programming side - one of the most annoying problems to get around was the right-click/item changed event. This may seem trivial, it can have a severe affect on the user experience. Anyway, it was only one of the problems. I think overall, I just didn’t like how the events were handled in wxWidgets.

…so I decided to move to Qt - fully with the grass is greener on the otherside mentality. It was mostly true. Qt is great. I like almost everything about it. It’s more manageable, the development tools are great. Visual Studio integration is awesome. It works with CMake…the list goes on. Anyway - one of the features of ImageWrangler 2.0 was tabbed browsing. Now, I have to say wxAUI was one of the greatest things to happen to wxWidgets ever. It pretty much brought wxWidgets into the 21st century. One of the coolest parts was the tab controller from wxAUI. Came pretty much ready to use.

…well, that much can’t be said for QTabWidget. Sure, it’s useable and there’s no crazy obvious bugs with it. But, it doesn’t have a close button for the tabs. Grrr…at least not in 4.4.x and earlier. I just read on a wiki and some blog postings that 4.5 is going to save the day. There’s even a pretty nice demonstration video of it.

Anyway, ImageWrangler 2.0 - here we come!

I recently moved back to the City of Angels. Upon moving back, I decided that it was probably time to settle a bit and not move around so much. So I got myself an apartment and a 1080p capable TV. 1080p is amazing. So here’s a bit of my experiences with 1080p, HDMI, and the lessons learned.

XBOX 360
I’ve been told that Microsoft only ships an HDMI cable with the XBOX 360 Elite. However, they ship a component cable that was least capable of doing HD and more importantly 1080p. Thank you Microsoft. Even though you completely ruined Vista - you at least did a lot of things right with the XBOX 360.

PlayStation 3
I went to Circuit City a few days ago to get a PS3. It was a bit of fiasco but I ended up getting one. Sony has HDMI plastered all over the PS3 box. I’ll post pictures later. So I get the PS3 and a couple of Blu-Ray movies. I go home. What do I find? A component cable that’s meant for standard def. I know just enough about video cables to know what goes where. But when I get a Red, White, Yellow component cable - I know it’s probably not going to do 1080p. You know the cable - it’s the red, white for audio and yellow for video cable. Thanks for nothing Sony. I know you’re losing money on every PS3 sold - but at the rate that you’re losing money, couldn’t you throw in an HDMI cable?

So if you’re going ot get a PS3 and an HDTV - make sure you get a bundled package or buy one while you’re getting the PS3.

Time Warner
While I was on the phone with Time Warner, I asked them if the HD cable box comes with an HDMI output. He said yes. Well the box I got has a DVI - and I know it’s almost the same thing, I just want an HDMI port on it. However, they like Microsoft were nice enough to have 1080p cable component cables.

Buying and HDMI Cable
I would recommend NOT going to Best Buy or Circuit City to buy an HDMI cable. A 6ft cable will run you anywhere between $40 to $120. The high end is if you want to get robbed at gun point buying a Monster Cable brand HDMI cable. I paid $40 for mine. Don’t do this either. If you really gotta have it now…Fry’s probably has one for around $15. While HDMI cables are not all created equally - most of the lower priced ones work just fine. If it doesn’t you won’t feel so jipped - you can always take it back. If you’re really patient, you can probably get a 6ft HDMI cable online for around $6 - shipped. Amazon is always a good starting and sometimes good stopping point for this stuff.

Welcome to the world of 1080p.

I failed to find an example to sufficiently make my point. However, please allow me to try to make it anyway.

After 3 or 4 days of using Windows Vista, now I understand why everyone hates it. In the past few years, Microsoft has often been criticized for security vulnerabilities in their various versions of Windows. To rectify this, Microsoft has decided to make it more difficult and annoying for the user to use their computer. An example of this is the UAC - User Annoying Component. If the confirmation prompt for every option isn’t annoying enough, the flashing is going to send me into an epelitic fit. There’s some poor unsuspecting Japanese children who are going to be in for a nasty surprise when they upgrade.

So Microsoft’s idea of security seems to be: make it very difficult for the user to do anything practical. Perhaps, then the user will not use the computer anymore. Users not using computers will result in less security problems reported. Or so Microsoft hopes.

Yes, and I do get it - I know security problems are very hard to solve. No one needs to be told that, myself included. But no one is shelling out $199 to raise their blood pressure every time they sit down at their computer.

I don’t know what was said at the Microsoft meetings about how Vista’s user experience should be. But it makes me wonder what exactly was heard from the customer side and how it got translated into the product we all know as Vista. Microsoft employs a lot of every intelligent people. I cannot allow myself to believe that these intelligent people were the ones that made the final decisions on Vista’s user experience. To try and draw an analogy, in the visual effects and film industry - the reason why things turn to crap is because, often, a person very far from both the work and the audience decides what’s best for both. We’re all familiar with Kevin Smith’s story about Jon Peters and Superman. So until I’m provided with an alternative, I’m going to believe that Vista’s poor user experience comes from the same sort of roots.

However, I have to say - the install process for Vista, for me, was less painful than the typical XP install. I may be the exception to the rule in this respect though.

…I must have missed something somewhere and it’s making my life terribly difficult. Okay, so VS2008 is in beta. I get it. Does that mean I can’t have it on the same machine as VS2005. What if I just wanted to see what it looked like, compile a few projects to test it.

Oh, I can do that?

Well, I can…but do I really want to. The answer is: No. I don’t know what happened - and I will figure it out once I’m done being so frustrated. I just rebuilt one of my applications in VS2005 - after removing…I mean uninstalling…VS2008. When I run an application on a different computer, I get this:

Resolve Partial Assembly failed for Microsoft.VC90.CRT. Reference error message: The referenced assembly is not installed on your system.

There seems to be a bit of a disconnect somewhere. Why would an application I build with 2005 (VC8.0) be referencing assemblies that are used in 2008 (VC9.0)? How do I change it back so that all my VC8.0 code references the correct runtimes? Hmm, the initial hour of Googling shows nothing.

I mean, forget that it’s beta software that you can uninstall. Why in the world did it change the assembly references in an older version of Visual Studio? The only foreseeable way I can fix this is to do a clean install of my OS. Unless I can find something before I decide to do that. What a freaking bloody mess.

I am positive that I missed something somewhere. It just seems like a major oversight to permanently cripple someone’s system just because they installed a beta product.

Oh wait, it’s Microsoft.

Update:
After uninstalling VS2005. Rebooting. Reinstalling. Updating to SP1 - the problem seems to have gone away. There’s a FAQ on the VS2008 site that says that repairing from the VS2005 installation would fix the problem…it doesn’t always.

I think this is a day late - or two. Went to the Eletronic Theater with all the dudes from Method. Some really inspiring stuff there.

I met some of the people from the Massive Software Team after my coworkers Gil and James did a presentation for the Miller spot we did recently. Massive has two female programmers on their team and from the conversation I had with them - they seem brilliant. I guess the idea that girls can’t program doesn’t hold much water. I had a chance to talk to one of them about the dynamics behind Massive. There was a slight language barrier or a me being retarded barrier there - but we got through the conversation.

The fluids course was really interesting - I’m really glad I sat through it. I had to skip out of the real time part to take care of a few things - but I made it back in time to listen to Robert Bridson wrap it up. It was nice to hear someone who has done quite a bit of research address some problems that I’ve ran across in my own work. Fluids is still fairly young and there’s lots to be researched. This makes it an exciting field of research.

These post seem to be getting shorter and shorter.

Today was an interesting day. 10 minutes into both of the lectures I was attending I got called to because our mail server was not responding. Since both my boss and coworker were on the road to SIGGRAPH, I had to connect back to the ranch handle the issue with some assistance from my coworker. But that was fine.

In between these things I did manage to run into Robert Bridson at the A.K. Peters booth. I was able to talk to him for 15 or so minutes about some fluid dynamics stuff. I’ve been reading/following his research with a bit of interest. So it a pretty cool experience getting a chance to talk to him.

It was a bit unfortunate - but I had to miss Robert Bridson’s presentation on curl noise because  I was trying to figure out what was up with the mail server.

This is a short post since I don’t have any real content.

Squish, Bounce, and Collide
The first two papers for this presentation were interesting. The last two didn’t seem relevant to the things that I work on so I didn’t stay. However, the first presentation did make me think if the metaballs need to have some sense of volume preservation.

Point Sets
By far, this one has been the most brilliant lecture I’ve been too. I geek out to math - sometimes. But what these guys were doing was amazing. I especially enjoyed the second presentation describing the use of a method called LOP to do geometric reconstruction. The idea was well presented and it really showed that the speaker had tried several different routes before arriving at his/their conclusion. I’m pretty excited about the entire MLS/LOP phenomenon.

The Point Sets papers presentation was inspiring. There’s a lot of research going on in this area. Several techniques have bee developed to raytrace point sets. One interesting thing you can do with this is write a raytracer where you can stream points to it. With the right methods this allows you to do progressive refinement without having to retessellate the representative geometry. Exciting stuff.
I went to the Autodesk Party and ran into a few people I knew from DNA there. Ran into someone I knew from SESI there as well. It was a pretty crazy event. There were Go-go dancers there. Exciting.

I think I’m too tired to try and type out something meaningful.

I should have written this yesterday - but got caught up doing other things. So I’m sitting in the San Diego Conference Center lobby writing this up now.

Mesh Parameterization
I just sat through my first SIGGRAPH course - ever. It’s actually the first fourth of a full day course. The topic was Mesh Parameterization. The first speaker, Kai, gave a very brief and easy to understand introduction to Differential Geometry. The second speaker, Bruno Levy, gave a talk on some of the applications of methods for texture mapping. Coincidentally, a subset of the problems and techniques they described was used on the Miller job to track the people on the moving mesh across 3D space. A map (texture) was used for look up to maintain the proper UV position within the locality of the triangle as the underlining mesh was changed due to the cloth sim. Since the UV was fixed, a point always mapped proportionately back to the mesh.

Practical Least Squares
The second part of the day was spent in the Practical Least Square Methods course. There were a surprising number of people in this course. Both in the first and second part of the course it was difficult to find a seat. I think SIGGRAPH has some weird no standing policy for the courses. The SIGGRAPH volunteers wouldn’t let people stand in the back of the room. Probably some fire safety thing. The math for the course was a bit hit and miss. There were a few interesting things that I haven’t had read across yet. So at the very least there will be some new reading material. Overall, the course wasn’t as in-depth and practical as I was hoping. However, it reminds me that I should really make a point to understand both the Jacobian and Hessian in Vector Calculus.

I wish I had brought my camera with me. There are some beautiful things to take pictures of. I’ll just have to steal pictures from Spaceman. Speaking of cameras - SIGGRAPH also seems to have a no camera policy as well. The program guide said if you were caught filming or taking pictures of certain things - your credentials would be revoked and you would be asked to leave. I think I understand where they’re coming from.

So today is Monday - so that means that most people who didn’t register for Full Conference are showing up today. I was told to expect everything to be a bit more packed. I haven’t ran into anyone that I worked with at DNA yet. There is the Autodesk User Group Meeting tonight - so hopefully there will be a few familiar faces there.

Also, there’s a surprising number of attractive girls here. By surprising I mean greater than 5. Maybe 10 is a limit though. =)

Nothing like nerd parties.