Helpful Information
 
 
Category: Web Design Help
Can a PC do design work by itself?

hi there!

i just had this idea of putting "creative guidelines" into program code. i got flamed badly by our graphics artist/art director that this is not possible. but i think, there must be kind an algorithm for doing "good layout". (he has no idea about programming / neuronal networks etc.)

do yous guys think it is possible to make a program do layouting and be creative too?

you can make a program to make layouts based on specific rules and guidelines .. ya AI experts would use different terms but this is basically what is happening .. as AI is right now, no you cannot make a program that will *create* as in what I think you mean ..
Nico

30-40 years till machines are equal to humans.
But if electronic music is any indication we are all in
trouble.

Donald Knuth's TEX idea was more or less what you are talking about. With addon libraries like LATEX you can produce professional output just knowing WHAT you want and without knowing HOW it should look.
TEX of course is very text- and documentä-oriented, and does not fit into the the link/multimedia/... oriented web-community.
I guess it's a good place to start. It's free, well documented, much previewed and probalby bug-free code.
XML lends itself as implementation base.

i know tex and latex.
tex is a bad point to start explaining from since it is only about text and formatting afaik.

I thought about an AI-program designing eg. a website. that is you feed it with some input about what it should look like (eg. "fancy, old-style, colourful..."), tell it some border-parameters (page size, menu items etc.) and then it will make some layouts that follow guidelines it gains by watching other layouts or web-pages or whatever.

do you think this is possible as such? i doubt there is endless ways to make layouts (just calc the permutation of the objects on the paper relative to a resolution, say in milimeters, and you have all ways.)

So: can one make a program that chooses the ones from this pool that are "good"?

and what is "good" design anyway?

In the CD that they are giving to people registering to the <a href="http://www.freshconfernce.com" target="_blank">Fresh Conference </a> a Design Agency is presenting a project in which they developed an application that "creates" COOL layouts. But they themselves consider it a joke, in the sense that THEY had to tell the app what GOOD layouts were, and how to do them.

So in the end, computers will never be "creative" ... humans are always behind them telling them what to do.

*even if programmed to do random stuff, it's always ME who tells it to do it.

yeah, it´s always you who tells it what to do. i talk about telling it to think! ;)


But they themselves consider it a joke, in the sense that THEY had to tell the app what GOOD layouts were, and how to do them.

what´s the joke about this? i think this is considered "learning", what did you think this is?

i think the human brain works big time like a program. just sometimes some outside-influences that make it disbehave...

i´ll have a look at your link later (you misspelled ;) )

I guess this would be more accurate :
you asked "do yous guys think it is possible to make a program do layouting and be creative too?" .... yes, do layouts yes...
no... be creative no.

for I think that concious creativity in a human atribute. but doing layouts sure, you can program a computer to do so.

When refering to TEX, I first of all meant its philosophy. The user says what he wants the programm knows how to do it. The user doesn't simply throw icons pictures and text-bits to the programm but clearly organizes the info.

If you don't like TEX, XML would be the ideal way to organize the data; stylesheets will organize the display.
So all you have to do is:
- define a sensible set of XML-tags
- find a set of rules (for each stylesheet) to display the data for each tag.

I thinks thats possible but certainly not easy. After all, TEX was only possible because decades/centuries of typesetting experience had been accumulated and was available in writen form. (And TEX is somewhat limited to that good but conservative style).
Deriving stylesheets automatically from existing web pages looks difficult to me, since the programm may have access to the layout, it can't easily evaluate the data (unless the data is organized in a XML-type fashion).
So you have to painstakingly find try to find out e.g. which link is the page-owners email address and which is an interesting page the author is referencing (certainly these types of data should be treated differently, and mixing them up in a reference page will surely mess up the style sheet).

I think it all comes down to: "can you define GOOD?". If you can express, programmitcally, the qualities of a good layout, you could sic a program on generating random layouts and only keeping the ones that fit its definition of "good".
To define "good", you could have another program that started with a very general description of "good", then showed layouts from the first program to a panel of users to evaluate. It would then look for correlations in layouts that were deemed "good" by the panel of users.
Complicated, maybe, but certainly easier than writing software to book airline tickets for a person over the phone (done at MIT) or play a really mean game of chess.

Originally posted by bricker42
I think it all comes down to: "can you define GOOD?". If you can express, programmitcally, the qualities of a good layout, you could sic a program on generating random layouts and only keeping the ones that fit its definition of "good".
To define "good", you could have another program that started with a very general description of "good", then showed layouts from the first program to a panel of users to evaluate. It would then look for correlations in layouts that were deemed "good" by the panel of users.
Complicated, maybe, but certainly easier than writing software to book airline tickets for a person over the phone (done at MIT) or play a really mean game of chess.

I think there's something else to it: "good" doesn't mean the same thing to different people (nor creative; we can see that from one of the previous posts where author has not much respect for "electronic" music, though some consider it creative).

We have examples of "creative" in history of arts where "bull****" (in the sense of previous art; consider "modern" art and some of the completely blank paintings which resulted from it -- of course, really creative) turned to be a better form of creativity.

Currently, I believe the only problem with "computer creativity" to be the lack of real randomness (most, if not all, RNG's in computers are pseudo-random number generators).

When it comes to rest, it's all a matter of pattern and template we're used to. Someone mentioned that there couldn't be too much layouts which are deemed acceptable, and rightly so. It's known that "mathematics" is quite "creative" (consider fractals, and lots of other functions), and maths is something what can (in a large number of cases) easily be implemented on a computer.

Next, we must ask if there really exists a creativity like most want to think of it? Is what we (I proudly believe I am one of them :)) humans actually do any sort of incognitive (sorry if this is not a correct negation of word cognitive; english is not my native language) process, or actually a mere selection of choices from pre-learned "truths" (like, black goes nice with red, blue with green etc).

I believe it to be the latter, and don't recognize any creativity whatsoever (this is a BIG lie, there is creativity involved in any process humans do, except mechanical stuff -- the question is could it be delimited like it's commonly done?).

I hope to raise some interesting points for all of you, and not to bore you severely.

Best regards.

PS. If not anything else, my english will probably be amusing :)

TeX (Knuth in his "The TeXbook" suggests writing it like this where the "correct" way is impossible) is a software package made with publishing in mind. It's made by a programmer and computer scientist (who got his degree in mathematics), so it's pretty much extensible.

If you work with bare TeX (TeX primitives -- commands at the lowest level of TeX language), you probably won't find a lot of "intuitiveness" in there (though, when looked from the point of implementor, most of the design decisions are clear).

The next level of TeX language: primitives + Plain TeX format (a collection of simple macros) adds just a few of simple layout related commands (including most of the symbol mappings for maths most people use, like \alpha, \int, etc).

What is here being talked about is LaTeX -- a package for structured TeX documents. The key in here is "structured". This means that there is a structure in those documents, and that usefull information can be obtained from the structure itself. Humans are fond of dealing with "structure", because it gives them a better insight into anything.

In LaTeX, structure is used to make it possible to distinguish chapter from sections, pictures from tables, and formulas from English words. It doesn't imply any sort of creativity on the LaTeX side -- it's still the duty of the document writer to recognize the structure of the document (s)he is writing.

The production of "beautiful" documents which LaTeX and TeX amount to, does involve creativity, but on the human side. TeX is a masterpiece of software (some commercial packages still, after 20 years, don't have some options it has), but not of that kind, and let's not give it credit it does not deserve or need.

As with any structured documents, XML documents have structure, but it needs to be recognized before it can be input.

The question here is not about producing the same documents if given same circumstances (that's what structured documents parred with their document processors do), but rather of producing different "acceptable" (whatever this may mean) documents in the same circumstances.

How this is accomplished "in" human brain is (I believe) through experience, or a large number of learned patterns.

But, now I have definitely bored you all, so, logging off if there's no interest in this discussion...

Greets

Trying to get robots to take over my job as Artistic Director are you? :p There is a difference between being creative and simply following orders. I think you'd need to create a machine that could feel (not just emulate, but actually feel hurt, happy, mad, be self aware etc..).

Creativity is drawn from our past experiences and our emotions. Could our very being (souls) be recreated with complex algorithems? Would you be playing God in trying to create such a being? What reprecautions would we face if we were to create a machine with immaterial properties?

You know thinking of these types of things scare me, I'd rather see our intellectual superiors work on planetary defence systems, new anti-pollution devices for our vehicles...

Again it is early and I'm in rant mode :D

gee, that would suck, if you're not the best. There would be no middle class designers. (good motivation for me)

Then progs would substitude programmers and so on untill The Matrix will be in non fiction section of the library.

But as long as people don't let that happen, and the quality of work of those programs won't be as good as a person's. I'm not gonna form an organization against A.I.
---------------

I think it's possible, if you could simulate what the human brain goes through in a lifetime, and to be creative it has to be able to experience things.


that's scary.



But you all are too smart for me, i'm pretty sure i'll regret i posted this. lol










privacy (GDPR)