hello.

I just had one of those wonderful moments where a bunch of ideas that had been floating around in my head for a number of years came together and made sense, thanks to a section of Alva Noë’s book Out of Our Heads: Why You Are Not Your Brain, and Other Lessons from the Biology of Consciousness. In Chapter 4, he challenges the common metaphor for the brain as the “Mission Control”  of the body — the place where all stimulation comes in and is noted, processed, and responded to. Instead, he says, our perception, and even our reaction, is distributed throughout our body and even through our environment.  To counter this, he offers the example of a snail’s response to being touched. At first touch, the snail will recoil, but with repeated touches, the snail becomes habituated to the touch, and doesn’t recoil. The sensory neurons in the snail’s nervous system are linked to the motor neurons, and the response to the initial touch is to cue the motor neurons to move the snail away.  As repeated touches occur, the snail’s nervous system learns the pattern as “normal” and the connection between the motor neurons and the sensory stimulus is lessened over time.  There’s no central brain managing this — the change is a result of the connection between the neurons and the patterns of action in the environment in which the snail is embedded, argues Noë. It’s not just about the changing in the coupling between the sensory neurons and the motor neurons, because that change would not occur without the repeated pattern of touch that the snail encounters.  It all happens without a “mission control” brain to process it.

Continue Reading »

Make: Electronics, Charles Platt.  © 2009 Make Books, Sebastapol, CA; 1st edition ISBN: 0596153740

Just go buy it. It’s the best introductory book I’ve read on electronics.

To start with, the book is gorgeous.  Maybe you can’t judge a book by its cover, but you sure can by browsing its pages.  This book is filled with clear photos and color illustrations (makes it so much easier to read the color-coded wires!) and big, clear circuit drawings and diagrams.  It’s a pleasure to look at, so you want to read more.

Charles Platt writes in a tone, and with a philosophy that I thoroughly agree with:  learn by doing it.  I love the fact that he not only gives exercises, but gives some that he knows are going to fail, and tells you so.  He shows you what can go wrong, and makes you do it, so you’ve already experienced the failure and don’t fear it. Exercises like licking a 9V battery, or measuring the resistance of your tongue seem scary at first, but are safer than they seem, and valuable learning exercises.

Platt doesn’t hide his mistakes, either. He uses them as stories to illustrate his lessons. Reading  – and seeing in pictures — how he blew up a capacitor, for example, is fascinating, and lets you know that when you make mistakes, you’ll survive too.  The stories of his mistakes are very reassuring.

There is plenty of electrical theory in this book, but you don’t feel like it’s being shoved down your throat.  Platt explains conversationally in examples, pictures, and short biographical sketches of some of the big names in electrical history. By the end of each chapter, you’ve absorbed a lot of material, without the feeling of exhaustion that comes from reading most textbooks.

Platt includes shopping lists of all the parts you need, suggests sources, and tells you what acceptable substitutes will do the job.  He makes it as easy as possible for you to get what you need. Even better, he includes directions for cleanup and recycling of your parts too, letting you know what can go in the trash and what needs to be recycled responsibly.  It’s great to see an author treating the whole life cycle of a project as a matter of habit you should learn.  I’ll be taking that away as a lesson for my own writing in the future.

The whole production team on this book deserve praise for this one.  It’s well written, well edited and well designed. It’s a great learning guide, and will be a staple on my shelf.

The chances are good that if you’re reading this, you probably know how to ride a bicycle.  You probably learned so many years ago that you don’t even remember how you learned.  When you get on a bike now, it’s so natural to you that you don’t even think about it.  In fact, if you did think about it, you’d probably fall off. Bike riding is so commonplace that we use it as a cliche for learning and retaining new skills:  “It’s just like riding a bike,” we say, “you never forget.”

That’s easy for you to say.  You know how to ride a bike.  I don’t.  I never learned as a child.  I’ve tried a few times as an adult, and let me remind you, it’s not easy to learn.  You fall off, you scrape your knees, you get frustrated.  Eventually (they tell me), you master it, and then it becomes “intuitive.” Yeah, sure.

Designers and engineers (and users) often praise computer interfaces as  intuitive or natural. It’s commonly held up as the ideal for user interfaces.  But I think the terms are misused. Interfaces in general, and computer interfaces in particular, exist to connect people with technological systems or tools that require significant training to master.  They abstract complexity to make it easier to use. But in order to use an interface, any interface, you need to learn a little something.  You need to be given an idea of what it controls. You need to learn what you can manipulate and what you can’t.  And you need to learn what actions map to what results.  I’ll spare you the Don Norman-esque rant here, because Gibson, Norman, et. al. have already explained all of this just fine.  If you don’t know Norman’s stuff, just think about the bike. It abstracts a lot of physics and mechanics for you, takes advantage of your physical abilities (like your sense of balance and direction), and it amplifies your body’s ability to do something — once you learn how to use it. Many would argue that it’s a good interface.  As a non-bike rider, I might argue the case, but I know I’m in the minority.

When you refer to an interface as natural or intuitive, you assume it doesn’t need to be taught, that “you can just figure it out.”  The example in vogue is the multi-touch interface, particularly Apple’s pinch-and-zoom gesture.  It gets a lot of press as being intuitive, but the truth of the matter is, if Apple hadn’t done a fairly brilliant series of tutorials disguised as commercial ads when they launched the iPhone, there’d be a lot more iPhone users struggling with their phones.  I’ve seen a lot of people pick it up pretty quickly, sure; but I’ve seen equally as many struggle with it the first time.  It has to be learned.  It may only take a few seconds, but it’s still learning.

What bothers me about all this talk of natural interfaces or intuitive interactions is that it makes designers and engineers less symathetic to end users.  If the interface is natural, we reason, anyone can use it. We get impatient with users who don’t understand it.  Not understanding the interface is not natural! They must not have any intuition! Stupid users!

In contrast, if you see all interaction as something that has to be learned, you know that you have a responsibility to teach people to use the things you build. You have motivation to make it as easy to learn as possible, both for your own convenience and theirs.  My friend Bill Braine, paraphrasing any one of a number of cognitive psychologists (including Norman), put it nicely (and I’m paraphrasing him a bit here):

Intuitive interfaces draw heavily on earlier learned behavior, while unintuitive ones require distinct new skills or metaphorical connections. That’s why so many metaphors (cut & paste, gold stars, file folders) made their way into digital interfaces. An interface’s ‘intuitiveness’ or ‘naturalness’ is a measure of  the extent to which draws on existing learned behavior.

In other words, an interface’s ease-of-use isn’t a binary quality.  It’s not natural or unnatural, intuitive or unintuitive.  There’s a sliding scale, and the scale depends not only on the cognitive load the designer assumes is there, but also on the user’s experience. If you must refer to an interface as intuitive, ask how intuitive is it? Better yet, praise the interface as easy-to-learn.

You can never eliminate the learning curve, no matter how much you flatten it out. So next time you’re frustrated by users struggling with your intuitive interface, remember  learning to ride a bike.

I’m on the jury for SIGGRAPH Asia’s Emerging Tech this year, thanks to Lars Erik Holmquist. Looking forward to seeing some great submissions, so submit! Here’s the CfP:

Call for participation:

EMERGING TECHNOLOGIES: ADAPTATION

SIGGRAPHA ASIA 2009
Pacifico Yokohama, Yokohama, Japan
16-19 December 2009
/www.siggraph.org/asia2009/

SIGGRAPH Asia 2009 welcomes contributors and volunteers to participate in the 2nd edition of the annual SIGGRAPH Asia Conference and Exhibition on Computer Graphics and Interactive Techniques.

ACM SIGGRAPH is now presenting two conferences each year.  The first SIGGRAPH Asia, in December in Singapore, quickly established SIGGRAPH’s new event in Asia as one of the world’s two leading annual showcases for the next wave of interactive technologies.  In Yokohama, SIGGRAPH Asia 2009 will present technological innovations that will change the way we work, live, and play. Let your work be experienced by thousands of attendees from throughout the world!

Emerging Technologies shares an overall theme with the Art Gallery: Adaptation. For SIGGRAPH Asia 2009, we are seeking works that show how digital technology adapts, or can help us adapt, to a world in flux. Adaptation can have many meanings, from machines that repair themselves under stress to technologies that help humans survive in adverse conditions. Adaptations can also be unexpected or unwanted. We are particularly interested in seeing examples of how computer graphics and interactive techniques are evolving to adapt to new conditions: technical, social, environmental, or something else entirely.

We are looking for creative, innovative technologies and applications in many fields, including:

• Audio
• Biotech and biologically inspired technology
• Computer graphics and animation
• Context-aware applications
• Design
• Displays
• Games
• High-speed networking
• Interaction technology
• Mobile technology
• Physical computing
• Robots
• Ubiquitous computing
• Virtual and mixed reality

Feel free to submit even if your work is not covered by the above list! If it is new, cool, and exciting, we want to see it at SIGGRAPH Asia 2009!

The main form of presentation at Emerging Technologies is live, working, hands-on demonstrations. It is essential that your work can be effectively demonstrated by the time the conference opens in December. To ensure this, the submission process requires all submitters to provide extensive technical, logistical, and practical documentation. Running an Emerging Technologies demo is hard work, but it is also very rewarding. Your work will be experienced by thousands of people, and thousands more will learn about it in international media reports.

We invite submissions from academic institutions, corporations, and individuals. All submissions will be reviewed by experts in relevant fields, and the final selections will be made by a jury. In addition to submitted works, a portion of the exhibition will be curated; please contact the Emerging Technologies Chair if you have suggestions for curated works! Juried and curated works will be clearly distinguished in the conference program.

The members of the SIGGRAPH Asia 2009 Emerging Technologies Jury are:

• Mark Billinghurst, HIT Lab New Zealand
• Cynthia Breazeal, MIT Media Lab
• Paul Debevec, USC Institute for Creative Technologies
• Tom Igoe, New York University
• Jun Rekimoto, The University of Tokyo
• Kimiko Ryokai, University of California, Berkeley
• Albrecht Schmidt, University of Duisburg-Essen
• Annika Waern, Interactive Institute

Accepted Emerging Technologies works will be presented as on-going demonstrations throughout SIGGRAPH Asia 2009 in Yokohama. Technical support will be provided to make the most of each demo. If your work is accepted, you will have the opportunity to give a technical presentation, advertised in the conference program. Press tours and other media outreach will be organized to further increase awareness and publicity. A promotional video will be widely distributed, and a description of each exhibit will be published online and in a printed catalog. Additionally, we will seek to offer selected works the opportunity to be published as peer-reviewed papers in an edited special issue of an academic journal.

The submission deadline is 5 June, 23:59 UTC/GMT. The jury’s decisions will be announced in early August.
Full submission details can be found at: /www.siggraph.org/asia2009/for_presenters/etech/index.php

We look forward to seeing your submission for SIGGRAPH Asia 2009 Emerging Technologies!

Lars Erik Holmquist
Emerging Technologies Chair
Swedish Institute of Computer Science and Södertörn University

A few people asked for the story of my visit to Tony Di Fiore’s research site in chronological order.  Here are all the posts in order:

• Part 1
• Part 2
• Part 3
• Part 4
• part 5
• part 6
• part 7
• Part 8
• Part 9
• Part 10
• coda

I’m home now, after a couple days of traveling back from Tiputini.  The trip home was calm, but not without a few interesting moments.

The trip home was basically a repeat of the trip downriver, but with a lot more people.  In addition to Tony and I, two of the student groups were leaving, as well as Diego, the station manager, who was going back to Quito for a break.

Leaving at 7:30 AM, we took a boat two hours and change upriver to another research station, where we waited for a truck.  The truck took us an hour or so down a road cut by the oil company through the forest to a checkpoint at the edge their territory.  There, we got stuck for a couple of hours waiting for another boat because of the rain.  The second boat was delayed because it was bringing another group of undergrads to Tiputini station, and their flight was delayed landing in Coca.  We sat around in a garage at the oil company checkpoint waiting, while the students chatted, flirted, grouped off to play cards, show off, and so forth.  Watching it all, Tony turned to me and said with a smile, “You really don’t need to go to the forest to see primate behavior, it’s all right here.”

The boat finally came, and we headed up the last stretch of the river to Coca.  We got stuck on a sandbar at one point, and the boat driver had to do some fancy motor work and poling to get us back in the channel.  By that time we were almost late for the plane. Fortunately, they held it for us (since our group was the majority of the people on that particular flight) and we made it back to Quito safe and sound around 7PM. 12 hours from station to hotel.

Photos from the trip are finally online on my Flickr site, for those interested.  All in all, a great trip, and one I hope to repeat in the future.

18 Jan 2009

Fourth day no big generator. The morning is overcast, and a bit muggy.

Summarizing my project notes from the whole trip so far:

Notes on gear use

Each researcher carries four main pieces of electronic equipment in the field: GPS unit (on lanyard in dry bag attached to backpack, usually); Palm (on lanyard in dry bag around neck or shoulder for easy access); voice recorder (haven’t seen anyone use them, presumably in pack) and telemetry receiver (around neck). None of the devices communicate with each other currently, though it would be useful if each note in the Palm were tagged with GPS coordinates. Likewise, a telemetry unit that logged its readings and geotagged them would be useful.

Again, any communications between devices would have to be wireless, but would also have to avoid interference with the telemetry signals. Worth noting how my Nikon D70 camera generated considerable RF interference for the telemetry unit.

Summary of possible useful items for the lab

• coaxial cable repair kit, along with coax connectors and cable
• Uninterruptible power supply
• Solar panel on roof/battery storage unit in lab
• dry box improvements (?):
  • all incandescent lamps for greater heat
  • humidity sensor to trigger on/off instead of timer?

Conclusions on radio tests

2.4GHz radios work fine out to about 60m, though there is significant attenuation due to interference from vegetation. Our proximity tests didn’t conclusively show a difference between 1m, 5m, 10m, though the tests were not very rigorous. There was a lot of noise. The data from the tests is here, in Excel format:
(post excel files here)

Things to consider for next tests

I think it’s worth investigating 900MHz solutions for the longer range, which might mean more coherent signal at shorter range. Is the Atmel AT86RF212 worth considering? It’s their only 800/900MHz offering that I could find. Also, look at the Digi XBee Pro DigiMesh 900 radios.

I’d also like to do successful accelerometer tests.

Future height signal strength tests should use tower 2 or the canopy walkways to avoid signal interference of the metal scaffolding of tower 1.

Every device to be tested should be thoroughly sealed. Even on a pleasant day, there’s enough humidity to cause potential interference. Put it in a box, seal the outside, controls on the outside with rubber washers and gaskets.

Antennas should ideally be woven into the collar, or perhaps even be the collar.

RFID and tracking

Early on, I speculated on the idea of using RFID in tracking. After a little reading, I’m skeptical about it. First, because passive RFID is already in use in animal tracking, written up in many papers as “Passive Integrated Transponders,” or embedded RFID tags. For our purposes, passive tags would make no sense, too short range. Active RFID tags would make more sense, and many are even within the frequency range of current telemetry,but there’s no such thing as a reader-tag, which is more or less what we need. The amount of customization we’d need to do is great enough that it makes more sense just to get generic transceivers and build the units that way, or to use something existing like the Sirtrack proximity collars.

GPS tests

The micro mini was a failure, most likely due to too small an antenna. Aaron at Sparkfun notes that they had their best results in testing with no interference, so the canopy here is probably the issue. Worth investigating the EM-406 or 408, maybe, with patch antenna.

The Nokia E71 got signal while on the river, but not under the canopy. Worth trying an iPhone out here, perhaps, or better yet, an iPod Touch with a GPS backpack. Tony mentioned the thought of switching to them one day in the forest.

Camera traps

Camera traps look easy to build; just a camera, microcontroller speaking the camera’s control protocol, PIR sensor, and batteries in a Pelican case. Usually a photocell as well to make adjustments for day/night, and sometimes an IR array for night light (not necessary with the monkeys).

Possible camera trap improvements:

• make your own, for cheaper (publish an instructable)
• Use the same power source for camera as for microcontroller board
• Use more energy-dense power source
• Investigate CHDK, Canon Hack Developers’ Kit, for possible use in camera traps. There is a motion detection script for it.

Things we didn’t get to

• coming up with a way to Access as a database tool with an online DB
• coming up with an automatic upload scheme so that field researchers’ data is automatically available to Tony and Eduardo on a daily basis
• investigating other handheld database solutions

Worth noting that the team here is largely Mac-based, but the software is mostly Windows-based. They do most of their work in Parallels.

Diego, the station manager, just came in with a giant toad, about the size of his palm, brown with a tan back. It was in the bathroom.

Last walk in the forest was good, just a short walk to the canopy walkways (less than 1km out there). They’re rope walkways from a tower like the other two to a couple of large trees in the area. Probably great for watching the canopy, but no way am I getting up there.

Tomorrow, a full boat back to Coca with the undergrads from Bethel University, the ones from Kalamazoo, Tony, and I.

17 Jan 09

As to how the new undergrads are: loud.  I know all about the dating lives of the guys in the next cabin now.

It’s another sunny day, hot for walking, but very pleasant otherwise.

Not much to report today, it was a pretty relaxed day for me.  Tony went out with his assistants to identify plants and check on how they’re flowering as a way to be able to estimate monkey diets.  I stayed in, and with no net and no radio work to do, spent time working on Arduino examples. A little pedestrian for the location, perhaps, but it was a nice break.

About 11 or so, he came back in, and shortly thereafter we took a walk up to canopy tower 2. It’s further away, about 2600 meters, which seems like nothing at home, but here, for me, it was a long hike.  Worth it, though, because the view from there is fantastic.  It’s a taller tower than the other one, I think, and larger. It’s wooden instead of metal scaffolding like tower 1, and surrounds the tree for more stability.  At the top there are three levels, a veritable tree fort.  It’s on the top of a ridge overlooking a lagoon, and as I said, has a great view.  I can see why the bird people love it.

On the way there, we ran across some leafcutter ants.  Pretty cool to see a long trail of leaf bits marching off into the forest. I got a few pix to be posted later.

While we at the tower, Heidi, who’s researching the photosynthesis of palm trees, came out to collect some samples.  In order to do this, she had to climb a palm, drag a saw up after her, and saw off a leaf.  No mean feat since the palm is 15-20m high and has no branches except at the top.  What she used was a pair of leg braces that clamp on to the tree and allow you to climb it one step at a time. The braces are attached to the tree by a loop of wire, and when you push your foot down on the brace, it tightens the wire loop and secures the brace whereever it is on the tree.  Lift your foot and you loosen the wire, allowing you to slide your foot up the tree. We had a good time watching her and catching the palm leaves.  Then we went down to the lagoon, had a quick paddle, and came back.  The shower I had when I got back was second only to last Saturday’s shower in its greatness.

Didn’t see any monkeys while we were out there, but Tony picked up one of the wooly groups, L group, on the telemetry unit, and was mighty happy. No one’s picked them up in a few days.

The main generator is still out, so all A/C is gone, and only the lab and the kitchen have power. There’s the occasional breeze, and a fan, so it’s plenty cool in the lab.

When many people are using the internet here, loading a web page is a bit like watching grass grow. I’ve been editing the same wiki page for the last hour and have almost made one change.

I’ve had crazy dreams since I’ve been here. Not sure why, but they’re all over the map. Some anxious, some amusing, most just plain weird.

16 Jan 09

Big generator still out today so we’re still operating on the small one. Conserving power, and no internet this morning. Running on battery power as long as I can.

Proxmity Tracking paper

Reading the proximity paper I mentioned a few days ago. A few interesting things stick out:

The collars had a weight of 125g. That’s too high for us. Are they less now?

The error rates look high to me. In the field, 34% exhibited problems they could not correct in the field. Apparently this is similar to the rates for GPS collars. Given my own experience with small GPS units, I wonder if that’s due to the size of the GPS collar. In any event, a 34% failure rate for proximity collars seems high.

The number of erroneous contacts is worth noting. Below a 15-second sample rate, they saw many continuous contacts show up as multiple contacts. That’s not so useful, though they do note that as the sample rate is taken higher, the number of erroneous multiple contacts decreases.

The number of “phantom radio detections” they had seems high too. I wonder if they are all analog radios, which is what’s causing the error?

They state that “because this was the first field deployment of these collars, such problems were not unexpected and determining their causes should enhance design modifications to limit similar problems.” In other words, they might be functioning much better now. This paper is from 2006, meaning there’s likely been another revision or two of the collars.

Their tests were all for the 1 – 1.5m range, though the collars can supposedly go out to 100m. I wonder if any research has been done on the longer ranges.

Further research: “Passive integrated transponder (PIT) tags have been used successfully to monitor individuals’ proximity to structures or use of certain areas (Prentice et al. 1990, Harper and Batzli 1996, Boarman et al. 1998, Kunz 2001, Gibbons and Andrews 2004) but have not been used to document individual-to-individual contacts. To our knowledge, the only previous attempt at proximity detection has been the MateID system (Sirtrack Ltd.; Ji et al. 1999, 2005)….. This system is limited in that contact can be estimated between, but not within, the 2 groups (e.g., genders) and only at very close ranges.”

“One-way systems may be available in the future, utilizing fixed-location receivers and animals carrying only transmitters, which would be similar to some current PIT tag applications (Gibbons and Andrews 2004). Such an arrangement would have positive implications for power consumption and collar size.” It would also have positive implications for our relay idea.

The rest of the day

In other news, it was a fairly quiet day. After a morning reading up on animal telemetry, we had a nice small lunch — only two tables out of about twelve were filled, because it’s Friday, and the boat took the last large group of undergrads back up the river after breakfast this morning. We had a leisurely lunch, filled with stories from Rudy the ornithologist about his travels around Bolivia, the hallucinogenic effects of ayahuasca and San Pedro cactus, and so forth.

After lunch, I finished off a little code and some email, then went for a walk on my own. It was a gorgeous sunny day, and the view from the canopy tower was spectacular. Lots of butterflies, but sadly, no monkeys. Then, on the way back, I ran across a group of wooly monkeys and got the chance to take some more pictures. Finally got a half decent one of a monkey staring down at me while he wrestles with another monkey. Those little bastards are hard to get pictures of!

Came back to camp and coded a little more, then did a quick run into the forest with Tony and Margaret looking for a wooly group she had been tracking earlier in the day. We didn’t find them, unfortunately. There is some confusion at the moment as to which monkeys are in which groups, because both Margaret and Peter have run across at least two signals at the same time. Tony’s not sure what it means, but he’s intrigued.

A new batch of undergrads have arrived this afternoon. We’ll see what they’re like at dinner.

15 Jan 2009

There’s a group of howler monkeys voicing their opinions nearby as I type.  I can hear them through the window. It’s pretty quiet, so they’re probably several hundred meters away in the forest, I’m guessing.

Saw tamarind monkeys today on the way back from breakfast.  There’s a group that hangs out in the camp and I had never noticed them before, but Tony pointed them out.  It was a group of about 9.  They’re tiny, maybe the size of a large rodent.  They ran all through the trees from the stream by the dining hall to beyond the lab.

It’s a frustrating technology morning for Tony.  First, he downloaded nearly six days of data from his GPS only to find that the memory holds three days, and defaults to wrapping around. This means that the most recent three days overwrote the previous three days.  That sucked.  Then, we had planned to walk out to one of the camera traps by the river, about 2km, and replace the batteries and memory card on the camera.  We hiked out there, longer than we thought because one area was flooded out and we had to skirt around it, and when we got there, he realized the batteries were not in Tony’s pocket.  A wasted trip for him. But hey, I enjoyed the walk, and the spot is really beautiful, right by the river.  I didn’t get any shots though, too much foliage.

Update on José: we ran into him in the forest today, taking down his traps.  He told us it turns out there’s no room on the Friday plane from Coca (where the boat lands) to Quito is full, so he’ll have to stay a day in Coca before he gets to Quito.

Did another test with the GPS unit, this time on the dock of the river.  Results:  nothing.  Sigh. It was pleasant to sit out by the river while it collected data though.  Got through a couple more chapters of Moby Dick while butterflies landed on my legs.

The big generator went out this afternoon, so they’ve got the lab on a small generator.  Now we have the fragrance of gasoline exhaust wafting in the window.

The food here is pretty good. It’s simple, rice with most meals, and all the bread is the same sandwich buns (hamburger buns, basically). This morning we had French toast made with the buns, and tonight we had broccoli and cauliflower over noodles in a cheese sauce, with garlic bread from the same buns.  Lunch was a to-go pack for the forest, two ham and cheese sandwiches on, you guessed it, the same buns.  The folks cooking are pretty creative with a limited set of staples to work with.  We’ve also seen an amazing variety of uses of bananas.  Tonight they were poached in a sugar sauce, I think. Sometimes there are banana chips, or grilled bananas.  There are always mangoes and some other fresh fruits available, usually apples and passionfruit.

Tonight for fun, I explained how GPS works, and Tony played human evolution raps from his class. Now there’s a more serious conversation about how to identify individual monkeys from the coloration of their genitalia.