Notices

justnick,

I'd rather know and try to identify a problem, and this corruption is certain inside that category, it upsets the experience.

As you may have noticed within liqbase I am my worst critic, and I use it and test and run all the time.

When I notice a bug I make a quick note and resolve to fix it when I get infront of an editor.

Its just too early on a Sunday morning to worry

*goes to make bacon butties*

Try not to work too hard on this project and always remember that it took billions of years to mother nature to offer us the world we see today, so you can have some time too to perfect your liqbase

I've thought about it and wrote notes about it many many times over the years about an awful lot of subjects.

However every time I've written something it i eventually have to turn the page and start again.

These pages get lost and thrown away because there simply isnt enough room.
I can never look back over things because of this, and have always wanted a way for me to hold onto what I write.

Its also given me a short term memory

Just because I have started to do this publically now does not mean I am going any faster or slower than before

I simply want to get enough done so I can relax and spend more time being a user.

Actually having a device sitting here and able to mould my thoughts onto is amazing, to be coding in c again feels so natural.

I am not losing anything I write and am building up the background framework to make use of everything I write and draw


Incidentally, there is an incremental update on http://liqbase.net/

0.0.4 I believe.

0.0.5 includes a multitouch test ( http://liqbase.net/ ).

its been rolling around for a while, but last night I managed to describe exactly what was required.
This evening I have put a test in place and want to see if its reasonable.


// use the primary thumb as ground zero
// then keeping the thumb pressed and stationary touch with a secondary finger
// the cursor will jump to roughly the halfway point (depending upon pressure differential)
// this is my chance, I know the original fixed start location
// and I know roughly where the finger is, i know its direction, and something of its position
// if the finger moves closer, its rough position will get closer
// it it moves further away, then so will the rough position
// if it rotates, i will know roughly where it is
// i can use this rough guide to stretch or rotate objects

It can be viewed from options.multitouch test. if it can be refined, maybe we can use it as an input method.

HOLY MOTHER OF INTERNET TABLET!!

It actually WORKED! Well, for about 5 seconds, then it crashed like the original playtest (requiring reboot, you know what I mean lcuk)

Problems: Works well in HiRes, bad in LowRes. Why does the outer box say ext, and the inner say finger, it should be opposite. And, obviously no pinch gesture, but "half-pinch".

How do you plan to differentiate between a normal click and a "multitouch".

Lol, two versions of liqbase in a day. I am overwhelmed.

I find it increasingly incredible how it manages to crash so hard on other machines.
I run enclosed in a wooden crate at performance all day long charging a battery and torturing my poor abused machine with ssh and gcc and liqbase drawing and doodling and running and playing and hacking and I do this on the other machine as well.

And I dont see failure rates anywhere near what appears to happen to you guys.

it genuinely shocks me and the worst part is nothing is its not reproducable :'( its simply failing.


I have extensive logging in the system, if nobody objects I would like to save this log to a file (replaced after a couple of runs) and request that you send it me if things go wrong.

Does this sound reasonable or even would people be ok if i asked on startup "you have an incomplete run log, do you want to send to liquid@.." ?

Wow, simulated multi-touch! That would really rock. But how to tell the difference between this and two taps and a swipe? Maybe if there's 0 time between origin-release (thumb) and secondary-press (finger), you can go into simulated multi-touch mode. If there's even a fraction of a second between release and press, treat it as two taps. Is the touchscreen interface capable of that kind of fine-grained polling?

qole, I already do.
I work on strokes remember

Then, there is the knowledge that a single finger moves in a linear track, if it happens to jump a large distance (bsg.hybrid.jump()) I can say:
"ahhhahhh you broke the laws of physics and an additional mass has pressed on the surface"

Once I have taken that step I limit the equation by asking that you maintain this first thumb in place, this acts as the pivot point.

I just monitor the filtered new results and use that to mean the second finger.

lcuk: perhaps see if the user can use ftp or scp (if they have ssh installed) and upload it to you that way?

qole, theres plenty of scope, i can just make a simple page on the server and push the data to it (or i can upgrade the server and run my own code).

Its no hassle, i'm just thinking of ways to get logs up to me easily.