OT:Theory and Hypothesis

rmack350 wrote on 1/10/2005, 9:02 AM
There was a brief exchange about the value of theory on this forum. I had all of this typed up but the thread dissappeared before I could post. Not wanting to waste my time, I'm posting anyway.

Disclaimer: The following examples relate entirely to Vegas. I try to explain the difference between two terms and how it relates to solving problems in Vegas. Of course, the ideas relate to all sorts of other things as well.

We use hypothesis and theory all the time in this forum. When someone has a problem, we offer ideas as to what causes the problem. Maybe it's one thing, maybe it's another. This is hypothesis.

Sometimes we take that hypothesis and we test it. In testing, we gather some evidence and then we can say that the evidence supports a conclusion. At that point it's a theory.

From that point we can take that supported conclusion and make a prediction that if the person with the problem does XYZ, the problem will go away. If the prediction pans out then there's further support for the theory.

And then sometimes we test a hypothesis, gather evidence, and form a theory, but we've misinterpreted the evidence. Later on we try a few more tests or make predictions and these contradict the theory. So the theory is wrong.

Here's an example that comes to mind. For a long time many of us had problems with stalling playback from 1394 drives. One hypothesis was that this was caused by the controller chip used in the bridge card in the drive case. The further hypothesis was made that this bridge card should be using a TI controller chip rather than any other. The test was to replace the enclosure with one that used the preferred chip and it seemed that this worked for everyone who tried it. So, with lots of samples pointing to the TI chip we came to the theory that the TI chip was the best choice for bridge boards.

Later on we got some reports that some people with TI based bridge boards were still having the problem. The theory was disproved because the tests were misinterpreted. Yes, switching to this bridge board helped many people but not all. The new theory must be that the controller chip was not the only factor contributing to playback stalls. Perhaps it was a combination of Oxford911 controller in the computer and TI chip in the enclosure?

And so on. The problem went on and on until eventually people replaced their hardware with new units that worked better. The problem eventually went away on it's own.

Okay, that was longish but I hope you get the point about hypothesis and theory. Theory is supported by facts and testing. It is also subject to being proved false. However, it is definitely not an opinion, nor is it a guess, and it is definitely not simply anecdotal evidence.

Rob Mack

Comments

PeterWright wrote on 1/10/2005, 6:52 PM
Well put Rob.
John_Cline wrote on 1/10/2005, 7:36 PM
One small correction; the TI chip is used in the controller card, the Oxford chip is used in the enclosure. Nevertheless, good post, Rob.

John
apit34356 wrote on 1/10/2005, 8:05 PM
rmack350, I assume that the following statement, "However, it is definitely not an opinion, nor is it a guess, and it is definitely not simply anecdotal evidence." is responsing to my comment.
A theory is an opinion on a series of deductions and facts, no more or less. Just attend a convention on astrophysics concerning magnetic fields and solar radiation. Watch the lecturers start punching each other out during lunch over who's right and who is entitled to research funds. .
rmack350 wrote on 1/10/2005, 8:28 PM
Doh!

John, I should really know that! My head has been so filled up with work details that I can't keep stuff straight.

Yep. I stand red-faced and corrected.

Rob Mack
rmack350 wrote on 1/10/2005, 9:14 PM
I dissagree. It's more than an opinion on a series of deductions and facts. It's a researched and tested conclusion that will stand up to review and further testing. It is a conclusion from which you can make reliable predictions.

It's very true that even well supported and researched theory is actively challenged and defended. Especially new ideas. And especially in a field like astrophysics where there probably aren't so many opportunities for hands-on testing.

People are people. We can very quickly become narrow minded where emotions like pride and self interest are concerned, and that's pretty much what you're describing.

I stand by my claim that theory isn't simply an opinion. Perhaps we can agree that it is an informed opinion. The point being that one must gather data, make tests, and be able to make predictions based on the theory.

Good researchers should be willing to admit when they're wrong and adjust their ideas, of course. They should also be able to tell bogus arguments from good arguments. And they need the humility to recognize that it's nearly impossible to prove anything beyond any possible doubt, but fairly easy to disprove things.

Bringing it back to film and video production, my favorite DPs (and favorite friends in the biz, not coincidentally) have always been the ones who wanted to do tests and comparisons. They were good craftsmen and also very smart and studious. And they were always willing to change their mind and look at things in new ways.

Vegas users can always do a lot of testing. I think that this is a great way to make use of this forum. Read posts, follow up with other sources, and try things out for yourself. Come to conclusions and then test your conclusions. I'm quite happy to try out other people's problems when I have the time.

Certainly don't take ideas as fact simply because someone tells you they are. Do your own testing. Otherwise you'll end up thinking that SD video is 4:3.

Rob Mack
Grazie wrote on 1/10/2005, 9:25 PM
Excellent and timely thread Rob. Try get to read Sir Peter Brain Medawar's book "The Art of the Soluble (1967)".

Thank you.

Grazie
farss wrote on 1/10/2005, 11:22 PM
All very well put and Vegas is probably the best tool for running experiments.
Things like generated media can be reliably used to simulate real world situations and evaluated impartially. One thing I would really like to see in Vegas is a zoomable preview window for doing these kinds of things.
One thing I do find curious, video in general and digital video even more so is far more science than art than film is yet the film mob seem much more scientific in their approach. Maybe it's because they're burning dollars at such a rate or because you've got to do a lot more hard yards to join the inner circles.
mark2929 wrote on 1/10/2005, 11:58 PM
"It's a researched and tested conclusion that will stand up to review and further testing. It is a conclusion from which you can make reliable predictions."

And yet in the theory of Evolution there is not one example or Shred of evidence of One Species changing into a other... A Dog no matter what Breed is still a Dog....:)
Coursedesign wrote on 1/11/2005, 12:07 AM
Some people argue that evolution shouldn't be taught in the schools, because they "KNOW it is wrong."

Seems reasonable enough that evolution is taught together with all other scientific findings (facts or sometimes beliefs), and presented as "the latest" or "the best that is currently known."

It should always be known to all students of science that there are no absolute truths in science. We are all conditioned to look at (and study) the world in a certain way, based on our past experience and learning.

Suddenly somebody comes along and thinks differently, and the truth changes. Just look at quantum mechanics.

Like the Hitchhiker's Guide to the Galaxy says, "There is a belief that as soon as we begin to understand the world, it changes into something infinitely more complicated. Some people say that this has already happened." :O)

Perhaps somebody else remembers the wording better, it's been a while since I read this classical very independent thinker.
apit34356 wrote on 1/11/2005, 2:18 AM
rmack350, the problem with "informed opinion", is that way too many times a theory becomes "viewed as an actual fact" by the public and sometimes by the "academic" community when they teach at campus. Most abuse occurs in lower sciences and social programs, where statistics is freely used to create "informed opinions", which leads to the public and the court system treating them as "facts". Over the last 2 years, the federal court system has been slowly reducing the use of statistical analysis as facts, too many experts using same data with different conclusions. Science and public would be better serviced if the word "theory" had a modifier like "24%theory" that would "rate" the information used and percentile of information used in the analysis, ie, there is known 100 examples of mountains over 10k alt., 3 mountains with 10.1k alt was used to determine all plant life at 10k alt.s and above.... so the proposed theory would be rate 3%limited-theory, helping the public understand. In the old days, I would have argued that "theories" were "very informed opinions", but I have watched the sound bite became king in the public and way too influential in most undergraduate programs.
rmack350 wrote on 1/11/2005, 8:26 AM
Not that I want to duck the topic (because I'd love to get into it) but the thread isn't about that issue. Nor had that issue been mentioned.

Theory vs Hypothesis is more an issue of language and arriving at a common meaning for the terms. On a secondary level, we're talking about the difference between unsupported conjecture and reasonably well supported ideas that can be used to actually get something done.

The example I'm trying to use is troubleshooting problems in a program like Vegas. For example, maybe you think that random frames in a render are caused by "xyz". So you test some scenarios to see if "xyz" consistently produces the problem. You also look for other explanations and you spend some time looking for ways to disprove the hypothesis.

A trap to this is that you may discover that "xyz" really does produce the problem but you you ignore that "ABC" also produces the problem. So all you've demonstrated is that xyz produces the problem. It's helpful to be honest with yourself about how far your evidence really takes you. The fact that "xyz" produces the problem doesn't mean that "xyz" is the only source of the problem.

Rob Mack
rmack350 wrote on 1/11/2005, 8:36 AM
I agree that people need to be more rigorous in their ability to discern the differences between fact, data (collections of facts?), hypothesis, and theory.

Some of the problem is human nature. We all want to have our ideas set in stone so that we can have a bedrock of solid ground to rely on. It's really unfortunate that we take so many things for immutable truth when they just aren't.

Troubleshooting software and cars is a lot more satisfying than troubleshooting dogma.

Rob Mack
Grazie wrote on 1/11/2005, 8:40 AM
"It's really unfortunate that we take so many things for immutable truth when they just aren't." . . ABSOLUTLEY! No question about it . . set in stone . . NO shadow of a doubt . . .

. .. oh yes . .. oh yes . .. :)

I used to be a skeptic .. but now I'm not so sure . . .
apit34356 wrote on 1/11/2005, 9:05 AM
A trap to this is that you may discover that "xyz" really does produce the problem but you you ignore that "ABC" also produces the problem. So all you've demonstrated is that xyz produces the problem. It's helpful to be honest with yourself about how far your evidence really takes you. The fact that "xyz" produces the problem doesn't mean that "xyz" is the only source of the problem.

Yes,rmack350, your example points out the problem of limited analysis of the information at hand. I detect an engineer's view from your example, its nice to have another technical view of vegas in operation. This is why I find interesting reading your's, farss, bj_m posts and other's posts and about your work experience. Usually theory and actually application of the theory are two different worlds.
mark2929 wrote on 1/11/2005, 9:55 AM
A Theory is way to look at how to make sense and create a model of a possible Solution Thus giving a framework ...

SOMETIMES A Theory can be a complete red herring, and looking at a Problem completely fresh can yield new Theories ...Or even an Answer ..The only real answer is a Provable one and even then there can be more than one answer.. Sometimes we think we have enough Evidence to say something is Provable when actually it isnt......

We all try to make sense of the world and our lack of understanding can be replaced by Theories.. This can help fill in the Gaps and help our minds see Reality as Understandable/Definable.... We are all here and something is sure making it possible..

Perhaps someone should come up with a new Theory that Unites all Religon into a Single Most likely "Real" Theory..Using Logic and evidence Like for instance Birth Control is neccesary to stop the world becoming Unsustainable ect OR Equality for everyone....

Ooops this episode has been done many times in star trek.. Perhaps we should just celebrate Diversity and Just teach Tolerance ..OH yes thats about where we are... OK then Life is an Adventure So whatever path You take Best to you !