Wednesday, November 25, 2009

Rickys talk

FORM FACTOR

SUGGESTED Benefits of tabletops

  • Enhanced collaboration - face to face vs side,by side
  • More equitable participation
  • social 'flattening'
  • - subtle social benefits.
  • supports eye contacts, casual interaction with device,
  • Natural interaction
  • Mimics the physics of real world - so simpler to use
  • Can support hightly 'fluid' interaction
  • Bimanual interaction - higher interaction bandwidth

Territoriality
- people use orientation and proximity to identify personal zones
- centre space is public/shared
Moving items
- momentum
- magnified control
- [ river along the edge ]
- spinning table
Problems
- bad lag
- what causes lag in a vision system ?
- lag proportional to number of people in system?
- fidelity of recognition
- fingers are quite big ( 1-cm sqaure )
- text orientation

What are tables good for?
Ordering in restaurant

Augmented table tops
- tracking real objects ( background subtraction... )
- arms /hands ...
- IR pens..
- Pen's
- when pressed down..

Is the ecology of devices about designing for multiple screen sizes
- how to prioritise what is important to display
- how to respond to near by devices.


What are my three inspirations/applications
- 1. huge card sorting
- 2. calendar with in-situ expansion of days.
- 3. zooming catalogue of anything.

Tuesday, November 24, 2009

On the 'if I had to build a carpet again' pile

http://www.glowsafe.com/e-shop/index.php?cPath=39_41

For that Tron effect.

works mostly on clothing. I wonder what voltage ?

+44 (0) 870 199 8524

Monday, November 23, 2009

more bits about the cloud

from meeting with Richie

He's still on about getting it perfect and not needing external PC. I'm more concerned about building something that moves.

I'm going to order those Motor-shields and get them built by the guys in electronics.

lab port - Ctrl/2/42

shields.

http://www.adafruit.com/index.php?main_page=product_info&cPath=17_21&products_id=83&zenid=0ebc36853cd859761469bdde6097b26a
16:55
http://www.lantronix.com/device-networking/embedded-device-servers/xport-direct-plus.html
16:58
http://www.ladyada.net/make/eshield/

Friday, November 20, 2009

What a day.

OK I got in this morning ( after getting rained on via the way in) - no problem.

I was trying to sort out the wireing on the board when I manage to blow one of the H-bridges on the board - opps. Ian then comes over and tries to figure out the wiring electrically. and in the process manages to zap the board and over the course of the day 4 more h-bridges start smoking.

I get panicked by this and start pulling together plan b. This as 2 parts
part one get 4 motor stepper shield [ here or here ] or [ here for two ] the down side is it looks like you can't by then prebuilt (boo).
The second part is to balance one ball (orange) against another ( blue), this means we can run 2 balls from one motor meaning 12 motors in all or 2 per row, so I just need 3 separate arduinios with 3 separate power supplies.

Plan C involves replacing the balls with lighter ones.

So by the end of the day Ian managed to get the up cut off switches working. I'm hoping to do a software fix to make the right ones work. The problem is I can't work on it on Monday/Tuesday.

Over lunch we talked about forgetting the whole thing. I pointed out we have made the vidieo already.

Tried, but while I was working on the cloud someone came in who wanted to use the cloud in a photo shoot about signing a memorandum of understanding. Life is too weird.

Wednesday, November 18, 2009

Not bad day

Got all the sensors on line.
Only have to change floor mat on sensor 2 to make sure getting reliable counts.
i might work on network to get it to be more reliable.

Stats

I have revived an email. Looking for a suitible stats program to do some analysis with. This time the datais in my SQL format.
To be sure there is no stats program with a good interface I could recogmend. Personally I am temptered to get my old copy of statview working.
Strange

Sensor net back on line

All the arduinos are now talking to the server!

Pressure pad on unit 2 (central stairs ) might have worn down. We need re-calibration.

Tuesday, November 17, 2009

State of the sensor network

  1. Sensor 202 ( entrance ) - network off line. I think the machine its self is OK. Contact networking again.
  2. Sensor 201 ( central stair case ) - one line and working but I think pressure pad is down replace.
  3. Sensor 203 (far staircase) -network ok. won't work from external power supply, Need to check pads. Fix put in 10K resistor to RX ping to ground. - Might get USB power supply to.

I think I can get 2 up and running by end of week. Possibly 3.

Monday, November 16, 2009

Friday the 13th not bad...

In many ways it is quite a good day for me.

I did an outline application with Helen for the in the wild EPSRC bit.
Yvonne resubmitted a bit in to In the Wild.
The EPSRC bit with Ruth is going well.
Now Ann A. invited me on the small public engagement bid.

More importantly my IPhone app was cleared so it appeared on ITunes connect!

This was all so pleasing that even having a flat tyre on the bike on the way home hasn't put me down at all.

Curiosu new language of the day

It's called Go from google.

Compiled, odd notion of classes but most intreastingly of all appears to have multiple returns. I'm glad to see the return of truely compiled languages.

Thursday, November 12, 2009

remember the experiment from hell well..

So I have the chi reviews.

I can't say much except I'm disappointed. The reviews sound very positive but they gave abysmally poor scores (2s) but no indication as to what was wrong with the paper. The only area for improvement someone gave was a couple of missing papers ( well we cutout far more than 2 papers to make it fit into a note).
Like many people I'm not convinced that the reviewers knew what they where talking about but manage to express this in such a vague way it was hard to pull them up on something specific which might have made a difference to the overall review. One good review might have been a good start.
To be honest I thought it had a reasonable chance of getting in and a better chance that some of the other papers that got better marks.
I'm starting to think that the thing that threw them was that we rejected the original hypothesis and then failed to come up with a convincing and ardent explication for what was going on. If the paper had just confirmed space syntax that would have worked. If we had said flatly this is because of X then I think they might have gone with it if it seemed plausible.

I think the thing that also went wrong was that we used fairly obscure syntax ( isovists ) and combined it with HCI/psychology style experiments. All of the reviewers didn't seem t understand the power of the winthin subjects 2*2*2 test and one critiqued the full reporting we did of all ANOVAs.

I don't know. I'm feeling like I've been defrauded by CHI. revolutionised
At least I'm not alone. May other researches think the same about their papers.

Wednesday, November 11, 2009

Box physics

http://www.box2d.org/

hmm - I like the version of this in processing but the lack of documentation was always a drawback.

very useful for doing more real interfaces for the table.

Monday, November 9, 2009

My new favorate link of the day

Links to HCIbib lovely browsing experience.

Microsft reinvent ZUI and lots of other things from 20 years ago.

Came across this link to a Microsoft demo.

  • Naturally there are bits I like and bits I don't.
  • So pen input - fairly normal even the mac has stroke to text.
  • The zoomable interface I like a lot but I would say that wouldn't I?
  • I like the notion of zoom able searching ( get more back)
  • The selection through eye tracking is again fairly old ( cute how the guy writing the blog is fairly excited by this). The big problem is that the eye cascades over the interface quite quickly. If something on the screen moves ( like the animated icon opening ) then you get the eye moving toward it. It must take a lot of blinkered self control to do this part of the demo.Given your motor vision and the eye movement are separate why not just let it grow under the finger/mouse.
  • Speech recognition -again fairly old - I saw this in 1994? this has the same problem as any command line interface, you have to remember the command.
  • Now you could use eye tracking as part of the emotion interface (the machine sees what are you looking at - why not start to load plain links before you look at them?) . Its also useful for the computer to know if you are looking at it ( use a audible alert to get your attention if your not).
  • I do like the flexible interfaces and the glass interface.
  • the gesturing looks fu. Again your don't want to do this more than occasionally with out getting gorilla arms ( remember I did VR research )

Friday, November 6, 2009

Cloud All new motors installed

but 8 are not working, 14 are and we are missing 2 wires.

I spent most of the day working on the new grant proposal with Ruth.

The building is dead

And it's getting dark.

Not a soul around

Thursday, November 5, 2009

Chi reviews

Just got the chi reviews back. No that hot but not clear what they were objecting to either. I'm going to give it a day to calm down then have a constructive look at them.


Wednesday, November 4, 2009

Cloud Plan B.

Got all but 4 motors not attached.

Found this link to a prebuilt Adafruit motor sheild for Ardunio.

Can run 4 bidirectiona DC motors. That is one per row or one per two rows if we can put the same voltage across two motors ( assuming current it is OK). So need 6 of these or possibly three if luck ( we never are).

Now wondering what to do with 22 DC motors...

Tuesday, November 3, 2009

Did I mention we are on Itunes U.

The labs project videos are now on the web - so much more lush than the one I made ( see you tube).

I was away when they were being done so I'm not in them.
Note these will open up in your copy of Itunes via ItunesU ready for down loading



Monday, November 2, 2009

JLB - meeting

We had a long meeting about the failure of the building for the academics. Not for research I must admit but people are staying at home to work in droves. Many people think that the OU is step up for remote working - it isn't. Teaching students at a distance has huge support but there is nothing to support academics working on building courses together short of emailing word documents around.

The meeting was looking for suggestions. I liked the talk of dividing the building up between mostly monastery - people looking to work in absolute silence, no phones the slight hint of a conversation would be met by people walking out into an open space. And the cafe - people tolerate noise/conversation and use head phones to exclude the outside world when they need to.

I suggested work rooms like IET - places to go and work when you need silence/exclusion. Other suggestions included monastery/cafe times - assumes a lot of social cohesion.

I thought security was being ignored so I spoke up for the use of rolltop desks big enough to cover your computer. That way leaving your desk to was a simple task of covering the whole thing with one key. That way you can leave things on your desk and come back to it undisturbed. That went down better than the IET style work places.

Lovely conversation about reintrodiction tea time - taking time to have tea with each other once a day. Generally thought to be sensible to put groups together again.

After woods I suggested using floor tiles + giving out an interruption meter.

Not so hot day for cloud

Ok to day I had one tutorial. One big meeting but I managed to get 4 spheres installed. The down side is that the second row gets about half the voltage of the first row. This results in no movement up and only some down. Not cool. I am not sure I want to loose a couple of weeks with electronics guys looking at ciurcit again.