Fixate or Glue -- Which is best?

[Xray]

Algodoo gives us two ways to stick opbjects to each other, and I want to know what the differences are between the two. I personally prefer to glue stuff because it doesn't leave those unsightly fixates that I see stuck all over some people's scenes. But are there any advantages to using fixates over glue?

Thanks!

[Conundrumer]

sometimes, when you glue complex layers of objects, when you remove an object, everything still stays glued to each other, or sometimes everything magically gets glued to the background adn you ave to redo glueing everything properly. fixate helps with clarity, in that you can actually see what's attached to see

[Kilinich]

Both way is acceptable, just don't try to mix it - that could be real mess if you delete something from complex body and it can be separated unpredictable way.

[Xray]

Thanks guys! Your replies were very helpful!

[jon_joy_1999]

I'm not sure about the mechanics under the hood of algodoo, but each fixate uses about 400 bytes of data in the .phn file, and I would imagine that each fixate would add a negligible, but non-zero-amount of overhead as it's rendered on screen, so you may have better performance if you glue things together if you're planning on using a ton of fixjoints

[faytree]

glue:
glue on geometryActions works like a fixate but it doesn't show any fixate on the material.

fixate:
fixate is very easier than glue and it saves 0.50 sec for me.But it shows there is a fixate on the material.Kinda annoying right?.
But for me is 60% annoying and 40% not annoying (fixate).
70% not annoying and 30% annoying (glue).

-jon_joy1999
i agree with you

-Kilinich
i was about to say that .

[electronicboy]

You are better off using fixates. Ps, you can select to hide fixates while running, not sure where that option is, but its there to be used!

Glueing objects is known to create issues with deleting objects and such!
(In fact, at one point in algodoo's history, glueing created a alpha quality release! Causing some scenes to be unloadable!)

Also, I haven't seen any performance difference between fixates or glueing.

Also, all a phz file is, is a bunch of pictures and a thyme script to respawn all the objects.
You can't really link the size of that file to the overall performance!
The number of fixates it would take to actually slow rendering down is probably astronomical!
Also, you have to remember, does the physics engine have any performance loss between the two?

[jon_joy_1999]

You are better off using fixates. Ps, you can select to hide fixates while running, not sure where that option is, but its there to be used!

it's under "Options -> Rendering -> Draw axels/fixjoints when rendering

Glueing objects is known to create issues with deleting objects and such!
(In fact, at one point in algodoo's history, glueing created a alpha quality release! Causing some scenes to be unloadable!)

Also, I haven't seen any performance difference between fixates or glueing.

neither have I [ed: yet], but still, the process of drawing a fixjoint does require some non-zero amount of processing power [ed: which we will see in a bit]

Also, all a phz file is, is a bunch of pictures and a thyme script to respawn all the objects.
You can't really link the size of that file to the overall performance!

I was looking at the raw file size of the PHN, not the actual PHZ. my fixate PHZ file is quite a bit larger than that, I just crunched some numbers, and got this:

Code: Select all

                       Fixate            Glue               Difference
fix test NN.phz   |   21,800            18,867         |   2933
                  |                                    |   
checksums.txt     |   42                42             |   0
scene.phn         |   12,964            12,611         |   353
thumb.png         |   20,395            17,517         |   2,878
Total             |   33,401            30,170         |   3,231
Ratio             |   0.65267507        0.625356314    |   0.027318755

The number of fixates it would take to actually slow rendering down is probably astronomical!

I tested this theory as well. on my ASUS G73SW-3DE I started to see a performance drop at around 1000 fixates. at nearly 10,000 (the most I could fit on my screen) fixjoints my refresh rate was 14 FPS, and my simulation time was around 40%. additionally,
.phz file size of 9984 fixjoints is 682,531
.phn file size same scene is 3,844,634

Also, you have to remember, does the physics engine have any performance loss between the two?

I doubt there is any difference between the two WRT the physics engine. glue and fixates seem to be defined by the Body variable. the Glued variable, or a fixjoint with proper Geom1 and Geom2 variables, defines if the two geometries are attached together. all the other fixjoint variables seem to be for graphical representation alone.

based on this information, I'm guessing it takes Algodoo 0.01% of a timestep to process a fixjoint
I don't have any information regarding how long it takes Algodoo to process Glues. I don't know how I would go about gathering that information, but I am open to suggestions and ideas

[Xray]

One very important difference between glue and fixates that I just recerntly realized is that you can glue objects together that are not physically touching each other! In fact, they can be separated by a very large distance. You cannot do that with fixates. There may be some unusual situations where you would want to have that capability, and, in those cases, glueing is your only option.

[electronicboy]

If you get the identity IDs of the 2 objects, you can spawn a fixate with the fixate spawn code.
Under the hood, fixates and glues are fairly similar, with glues allowing more then one object with one glue, fixates giving a visual, less buggy method.

[Xray]

If you get the identity IDs of the 2 objects, you can spawn a fixate with the fixate spawn code.
Under the hood, fixates and glues are fairly similar, with glues allowing more then one object with one glue, fixates giving a visual, less buggy method.

That's good to know. Thanks!

BTW - How do you know what's "under the hood"? Did you figure it out some how on your own, or do you have special access to information about Algodoo that most other people do not? Just curioius.

[electronicboy]

It's just something I've figured out mainly from logic and looking at how stuff works.
I have no access to any resources that anyone on here doesn't have access to, other then a few contacts whom wouldn't release any info about how the game works! (not even emil will drop the smallest of hints!)

[Xray]

(not even emil will drop the smallest of hints!)

Why not? He is the program developer, and so he, more than anyone else, should be willing to disseminate as much information about it as possible! Why keep it secret???

[electronicboy]

Why give away how your program works, and open yourself up for competition?

Algodoo is one of the MANY physics simulators out there, yet, algodoo is ahead of the competition because of its performance and feature benefits, as well as it's on-going support!

I seam to also find that the Algodoo community may be a lot more quiet then it used to be, but if someone on the board asks a question, its usually replied to within 12 hours, where its possible for the community to answer.
If not, I can usually pass the topic link to a algoryx member of staff, who look at it around 7am GMT (Sweden is +1, and my main contact from algoryx is usually online around that time, who is more then willing to deal with the problem).

[Xray]

Good points! Thanks

By the way, I wasn't suggesting that he give away secrets of "how" the program works under the hood. He would be shooting himself in the foot if he did that! I just hope that he keeps the Algodoo community updated on enhancements, bug fixes, and new features. I'm looking forward to the day when (if) Algoryx comes out with a complete User manual on Algodoo, and especially on Thyme! I would be one of the first users to purchase a copy.