Vis / Space: lwjgl non game application

Started by dronus, June 19, 2008, 21:56:28

Previous topic - Next topic

dronus

Hello,

I like to invite for beta testing..   

Vis / Space is a 3d browser environment, useful as an  artitistic, educational, scientific or just-for-fun tool.

Check it out here: http://vis.hfbk.net

It comes as an client-server application. The client is done as Java application using lwjgl and some other binary additions for 3d, video and input device support. It is build for Linux 32/64 bit, Windows and Mac OS X (Mac lacks some features due to reduced access to testing platforms..).
Besides no real game, lacking shadows, reflections and all other nifty effects, it can push modern GPU to the limit, showing wild structures of increasing complexity as you surf the web.


I would like to see some comments and abitrary harsh critique :-)

greets

Paul

bobjob


irongiant

It took me a while to get used to the mouse control, I think it should be more responsive. Note that I ran the app on a quite old laptop, so I'm not entirely sure about the responsiveness. Keys control looks fine.

I did a search and it wasn't easy to browse the results and find what I actually wanted. Plus, the whole bunch of floating words got me confused about what they are.

Overall, I really like the idea but it isn't obvious how to use it to browse the Web.

dronus

Yeah, the whole institution was down due to network problems it seems. Sorry for that.

"Overall, I really like the idea but it isn't obvious how to use it to browse the Web."
I totally aggre and disaggre :-) 

Browsing web sites for information is simply done best in the meant way: by the ordinary 2d browser. web pages are made for, and thousands of developers try to make it easy and attractive so we don't compare in this way.
Actually sitting on the sofa, putting the image to a wall by a beam; drinking some beer and just traveling around is more satisfying with vis / space in my opinion.. and thats its meant to. If you are addicted by some image or text you see, you can "escape" the visual world by hitting ctrl+b to open the hovered item in good ol' browser...   but later the evening you won't :-) .

wolf_m

I like the idea, but a lot of things annoy me.

- The navigation doesn't work well for me, it's not intuitive (you need to really work hard on that)
- I can't resize text fields
- I can't drag and drop stuff around freely due to the navigation limitations when d&d is enabled
- A lot of text doesn't fit into the text fields, they should grow vertically or you need a scroll bar
- I can't adjust the font size
- Pictures either don't work at all or I can't figure it out
- Same for video / sound
- There are no traditional forms (buttons, input fields, ...) to generate POST messages
- javascript, XHTML, CSS etc all thrown away
- The presentation is rather crude, it looks like a hacker project
- Too much stuff on the screen, can't be helped, I suppose... Maybe use an LoD algorithm to not show a percentage of keywords when a cloud is far off?

dronus

Oh vey, ..  as i stated in last post, it doesn't replace a browser and never will to.

All topics except "The navigation doesn't work well for me, it's not intuitive"  and "I can't drag and drop stuff around freely" are right as they are. Pictures can be sourced by the several picture sources (mixImages for example). maybe we found a nice way to 3d d&d. Also full six-degree-of-fredom navigation is possible with a SpaceMouse or SpaceNavigator (and some improvised data glove gear but thats another story).

It is the intention of vis space to get an abstracted view of web content, not a styled one. So there are NO sites. Data ist gathered from several soruces and mixed together. In this data space, layout of web pages doesn't matter and is totally ignored (except styling elements that sneaks into google images search for example).
All apple comforts are absolutely no-go, there won't be reflections or soft touch ever. this seems to be called hacker look ...

Much stuff on screen is what we like. On an mid class gamer machine one could surf around a whole evening, thus gathering about 10000 of keywords and images and still walk through fluently. So the "clear" hotkey is not mentioned in the manual :-)

wolf_m

Your way of navigating this is simply not smart. If I were you, I'd give this to 60 subjects. They'll confirm this.

I can see that the abstracted view of web content is intentional. There's just no point to it.

Even calling it art is stretching it. It's just a data view. I don't have the need for a 3D data view of textual web content, especially since the text is in a lot of cases unreadable.

I wasn't talking about "apple comforts" regarding the look. I meant the color palette, the symbols, the UI layout. It all yells "I AM UNFRIENDLY".

My advice to you would be to get some designers on the team. This is all programmer's art, and it's very obvious.

wolf_m

Sorry for just posting negative comments, that's not nice. I just realized that. And I have some positive comments on it.

- It never crashed here
- Making it stutter seems to be very hard
- I like the overhead map
- Your data trees align nicely in 3D space, always manageable
- zoom-to-keyword looks cool when there's a lot of stuff on the screen
- mixImages / googleImages / flickr / ... works pretty well, I didn't realize that before

A faster way of navigating could be implemented pretty easily by letting me use the overhead map for it. Do it. The map also needs a zoom feature (mousewheel?).

wolf_m

I'm also quite surprised that this is coming from the Hochschule für bildende Künste. I'm not seeing the art aspect of this. It's 'just' engineering.
Is it part of a bigger installation and thus deserves the art label through meta themes? Please elaborate on that.

Von mir aus gern auch auf deutsch via PM. Gruß aus Hamburg nach Hamburg :)

dronus

Well about "art":  I think it is between art and science (like universities should do...).  This is not an artwork itself, but a tool to. From an artists view, the computers often lack the tools to toddle about with like a brush and some paint.. The intersection of programmers and artists is often very small, so I think it is no problem (for an project that isn't an individiual artist) to choose the "engineering" way and build a tool. It's obviosly not as simple, universal and strong as a brush is. But we hope some people get used to it and use it for whatever they like.. 

We did some exhibitions with the predecessor software (eg. http://www.hirnsohle.de/photo.php?tag=imagespace ) It was connected with several other parts (image recognition body controllers, sensors, touchscreens etc. ) where vis / space was the basic framework.

vis / space on its own has some "readymade support", building up exhibitions of web found content just by user interaction. On a longer session, often several nice views occur where keywords and images form a more expressive combination, by random or by hidden proximities of topics. Sometimes borders vanish, as vis / space doesn't transport context information like a website do, so the delivery channels of information hides away (eg. political debates mashed up by porn, you wouldn't  bother if you accidently open a porn site just by an off target google click).

thanks for any comments so far

Paul