Interesting article appeared on ArsTechnica today: http://arstechnica.com/information-technology/2013/09/making-the-ultimate-creative-content-os-ubercreate-os-1-0/ Maybe some features requested from pro-users could end up in R2?
that was neat. it was really cool seeing things there haiku does well – scripting, multitasking, metadata, file search from within apps – and there was stuff i’m certain would’ve served him well – stacking and tiling, replicants. the spaces where haiku falls short are not even that daunting – the porting of already open source image and video libraries (i.e. openexr, libquicktime,) and some native apps that better present the amount of control a user has (i.e. the stock mediaplayer is frame precise, but only through a hotkey, only when paused, and there’s no documentation of it anywhere).
what is tablet support like in haiku? i’m thinking about getting one of the monoprice tablets (between them and yiynova, wacom finally has some competition) and it’d be neat to use it with wonderbrush.
Some reaction thoughts to the article, this is all R2 territory.
We need image previews in Tracker, we need a method to do quickly view documents and images like QuickLook, we need many more translators for more exotic image formats including .NEF, .OLY, .CR2 Canon Raw files, .HDR and .EXR files.
We need to add support for color profiles and color calibration ala ColorSync.
We’ve done a lot of work to make applications scale based on font size, but more work will be required to support 4k/retina displays and beyond.
We need up-to-date OpenGL support and drivers, we need a 3d accelerated App Server, we need to flush the OpenGL kit out to export 3d capabilities to applications.
We need to built in support for OpenCL so that processing can be exported the GPU where appropriate.
We need to add support for multitouch, we need to extend Interface Kit to have controls that work with touch, and need to build/modify applications to be multitouch aware.
Luckily we have a very nice file system that supports unlimited attributes, but, we need better support for adding attributes (automatically) and searching them. Basically ArmyKnife on steroids.
Tracker needs many more visual aids for organizing files.
some stuff can be brought in as add-ons and applications for now, yeah? and libraries imported, all of that. looking through the filesystem, libraries definitely exist as a thing on their own, rather than an integral part of the translation kit. there’s bound to be ports of more image libraries as media applications are built and workflows tested.
we need many more translators for more exotic image formats including .NEF, .OLY, .CR2 Canon Raw files, .HDR and .EXR files.[/quote]
Agreed. While we already have EXR, HDR and some outdated RAW images built-in translators, they need some love. I guess if/when someone will port latest openexr and libraw libraries, it would be quick to get that done.
While having 3D hardware support is definitively a must-do, it’s also the most complex task to achieve, which unfortunatly block all others progresses (switching app server to an OpenGL compositor, in particular).
What do you mean by “flushing OpenGL kit out to export 3D to applications”?
Adding multitouch input wont be the hard part of this. What’s hard is to adapt our actual GUI to become touch-friendly, not just touch-aware. Touch input bring new user control paradigms but also new constraints. Small controls that works very fine with small mouse’s cursor and high precision position world don’t work anymore with big opaque finger and lower precision world… And it’s not easy without rethinking the whole GUI from ground. Which means breaking everything based on current GUI…
Beside, so far, Haiku’s target is still desktop, and multi-touch and desktop are still a bit strange use case. When I do own a laptop with multitouch display, I don’t touch the display that much even under Windows 8…
Multitouch input support for R2, indeed, but full multitouch GUI will take a lot longer. And it’s telling considerer we’re not yet (officially, mostly) R1
What do you mean by “flushing OpenGL kit out to export 3D to applications”?
By that I mean adding support for drop-shadows, animations, transparency, transitions, fast bitmap scaling and other fancy 3d effects that are made possible by having hardware accelerated drivers and APIs. These effects, used appropriately, can add to the experience, especially for multimedia type applications.
Adding multitouch input wont be the hard part of this. What’s hard is to adapt our actual GUI to become touch-friendly, not just touch-aware. Touch input bring new user control paradigms but also new constraints… Beside, so far, Haiku’s target is still desktop, and multi-touch and desktop are still a bit strange use case. When I do own a laptop with multitouch display, I don’t touch the display that much even under Windows 8…
Before too long, thanks to Windows 8, it will become difficult to buy a laptop that does not offer a touchscreen display, and touch screen monitors on the desktop, now available, will soon become ubiquitous as well. Therefore I feel confident in saying that multitouch, although currently relegated to the tablet space, will soon infiltrate the traditional Desktop space as well. As a Desktop OS there will be more and more pressure to support multitouch as more and more touch-enabled hardware becomes available.
One approach is to rethink the entire GUI from the ground up like you say, but, that is not really tenuous for us because we’re not even done thinking through the current GUI! A more tenable idea is to supplement the current GUI with touch screen capabilities gradually which I believe will ultimately be the best way to go. Instead of replacing the mouse and keyboard, add the ability to do pinch-to-zoom, panning, flicking through photos to existing apps.
phoudoin wrote: Small controls that works very fine with small mouse's cursor and high precision position world don't work anymore with big opaque finger and lower precision world.. And it's not easy without rethinking the whole GUI from ground. Which means breaking everything based on current GUI...
I don’t see it as an either/or proposition. By keeping the mouse around, we don’t have to rethink the whole GUI all at once and we can build up touch support piece-by-piece. Also, you’ll be able to use touch where it is most appropriate, and ignore it where it’s not. For example, there’s probably not a lot of use for multitouch in Terminal, while a photo management app seems like a pretty natural fit. Imagine for example zooming photos using pinch to zoom in ShowImage with keeping the toolbar zoom buttons available as a fallback. Touch capabilities will naturally be used more by newer apps while older apps will continue to function using the mouse and keyboard.
Multitouch input will be added as an input_server device add-on (or in the current USB HID one) for USB multitouch devices (the most usual these days), plus some “touch gesture” filter add-on.
I own myself a multitouch screen, maybe one day I’ll look at that part.
But the bigger part is, as said above, to make our GUI multitouch friendly, and this could take a longer time.
Regarding TuneTracker Command Control specific case, that could be quicker as it’s a full screen interface with large tiled custom widgets, but it still will need to add gesture support (pull to refresh, pinch to zoom, slide to cancel or whatever make sense there).
as far as i’ve seen, there isn’t much production software out that makes good use of a touchscreen interface. as peripheral, like, say, a multitouch touchpad has plenty use, but it’s far from enough to base a production interface on.
Beos/Haiku already has a commercial application that uses a touchscreen, TuneTracker. As far as I know it is the only commercial application still available and still being developed. I have been using it for years and am thrilled that it runs completely on Haiku now. There should be some way to capitalize on that. There is a media centric commercial application for our favorite OS even before it hits Beta. Now that’s impressive!
I would really like to see touch support in Haiku for TuneTracker. My Thinkpad X220 tablet running Haiku is working great. Does anyone have any suggestions about getting the touchscreen working under Haiku?
a libquicktime port would also be neat, as would some implementation of puredata (it’d murder cortex, especially considering all the patches and externals available for it already).
I support this article.
It will take a lot of work but Microsoft and Apple are sort of helping us by trying to turn the desktop into a dumbed down, oversized smartphone. Not fun for people who just want to get some work done, however.