And we can go to the last presentation of this morning, that's experimental mobile wireless GRASS-based GIS for handheld computers running Django Linux. The paper is by Jan Stankovic, Marcus Neiteler and Roberto Flor, and it is presented

by Jan Stankovic.

As Marcus said in his keynote speech, GRASS goes mobile. Let me give you some introduction why to go mobile with GRASS or with any geographical

information system. In last year there was a dynamic development of communication technologies and it brings such, we can say, new dimension to GIS at all, and that new dimension is mobility.

It makes much easier if we want to do some field work, we can take our devices with us, and what is fantastic on that, we can be with the devices that have been in our office because what we are doing in the field, we can immediately in real time have another

computer far away. And mobility brought, we can say, a new branch to informational technologies and we can call those technologies mobile informational technologies. So now let us see what it is. Well, we can characterize it as informational technology that is targeted to processing

of spatial data that are actively using wireless communication technologies to transfer data between mobile clients or between mobile clients and servers and control center. And mobility information technologies use an advantage of knowledge of their actual

position of the client. So what are the main building stones of applications of mobility information technologies? As far as, of course, GIS, then mobile, personal computers, last year we have many

notebooks, a lot of powerful PDAs and smartphones, which are a combination of PDA and mobile phone. Other thing we need is wireless communication technology, then we need positioning system to determine position. They can be global or local. Global, we have GPS and local, we can determine position from VLAN.

And of course, applications are very specific. So from time to time, we need some other technology. Well, now let us go through particular building stones and I will tell you something more.

Well, mostly used wireless technologies in recent times are technologies based on GSM and operating networks. In this position, we have second generation of GSM running everywhere for one or maybe some longer time than one year. We have second and half generation, which brought us HSCSD and GPRS.

They are different communicating protocols that are faster than before. We had only 9.6 kilobits for communication and also charging was very problematic because once you made the connection, you paid for time, not for data transfer.

If you imagine that you go with GPS and you want to do some field work and send it to your office, if you will pay for the whole day, it's not going to be very cheap. So GPRS, for example, brought us payment for data transfer. So now it is quite cheap to go with GPS and do some field work.

Well, and what will the third generation of GSM bring us? You will see in a few years maybe some Asian colleagues can tell us more because in Japan they are already third generations running. And the other technology, I think what will be very concurring to third generation

of GSM is YC technology. It is much faster and the difference is between GSM networks and YC is that YC is not licensed. So this is why I think it's going to be a big concurrent because third generation

of GSM needs licensing if you want to, if you are an operator of GSM and you want to build new technology based on UMTS, you have to pay a lot of money for license. And if you want to build a network with YC, it's free, you just buy hardware and you can go on.

So this is why I think there will be some fighting between these two. Well, the other thing we need to build mobile GIT, we need mobile devices. Recent devices are very powerful, small. For example, we have here IPEX3870. Just remember your desktop computers you had three years ago.

I think it was quite similar. We can see there is 32MB of read-only memory where operating system and several software can be stored. Then you have 64MB of RAM, 200MHz processor and several other functions.

If you want to use it for communication, some expandability is required. We have a lot of expanding modules. For example, here is GPS, here is wireless card, we have expansion of memory, this is 1GB of micro drive, it's very small, I can show you later how it looks like.

So we have other building stone and now what we need is, we need communicating protocols and we need some operating system background because most of these PDAs come with Windows CE or as we have seen just a while before, when they are from Palm, they are with some

kind of UNIX-based system but very small. So I think it's very good to run Linux on handheld because those recent but I must say more expensive devices as it was said before, like IPEX are very powerful to handle

Linux, they can work with it. And you can change IPEX from digital diary almost to network server, there is wireless network support for wired network also because Linux is designed for networking. The other fantastic thing is scalability of the system and security because if you

want to use such system, let's say for some rescue services, police, army and so on, you need encryption and it's implemented in SSH. And what is the other fantastic thing, you have free operating system and free development environment, you can port a lot of software that is already existing for

x86 architecture and have it on your IPEX which runs on different architecture. Now let me say something small about Linux on IPEX, we can have fully capable x-window system as you can see, here is an example with GIMP.

GIMP is started from a different computer, it's not running locally because GIMP is quite big to fit in memory, of course we can fit it to some expansion card. But this GIMP is started on Linux server and output is redirected to IPEX. With wireless connection you can do such things anywhere in the field, so if you want

to do some very difficult computation, you can have it done on another computer and see the display, of course you can do it even locally because it's quite powerful. Well just by the way it's not connected to GIS but partly it is, there is even MP3 support, MPEG and so on, what is the connection, if you want to use such system

let's say for public, like I don't know public radios or let's say in library, you can navigate a person in library to some place with GIS and then you can tell him something with MP3, some information. Well Linux on IPEX is, we can say supported by Compaq and DEC developers

and there is also native compiling environment provided by Cambridge Research Labs. It is not very easy to change operating system on something like IPEX, as you can see I will make it smaller, allow just how it is.

There are no disk drives and no keyboard so it's quite difficult, the only thing you have is a serial line so you have to be enough with that. Several fantastic things were developed, one of them is Boot Blaster which is

such application that is able to flash a new boot loader to read only memory of IPEX. And then it is easy to communicate with IPEX on serial line from your desktop computer and you can have such, it's called root image of Linux which is copied to read only memory.

That root image contains support for wireless so then you can connect to your wireless network or of course there is support for USB networking and all other possible networking, all other networking possibilities and you can follow with installation of Linux.

There are some screenshots just to show. So this is, as I said we met all those building stones that we need for mobile geo-informational technologies and the last thing is that GIS. So here in Trento we decided that it's a good idea to have mobile graphs

and here you can see the first picture, it is taken just the first day. We were lucky with compiling. You can see monitor here, here is a graph shell and then we followed. So how it is done? We took existing libraries and functions of graphs,

we just selected a small portion of them because graph is very large and we compiled it for IPEG architecture. So mobile graph is now capable to show raster, vector, sites, it's able to do query,

you can input sites, you can use measure functions, you can use zooming and we are developing enhanced GPS support because support for GPS in graph is now done by external module and it's not real time and what we need is real time application that will directly read data from GPS

and send it, show it on display and send it to server if you want to be connected by wireless. These are the modules that were reported. So as I said about vector, raster and so on, we have some setting of GIS environment,

we can remove, copy, we can change map sets and what I forgot, there is also TCLTK support which is an alternative for command line but I will talk about a little problems about that. Now how did we compile? There are more possibilities.

One possibility is to build a tool chain which is a combination of tools that you can build on x86 machine on your Linux machine and then you can use it for compiling but it's quite complicated and it takes quite a long time to build such cross compiler.

So we decided that we will use an advantage of having native compiler available on Cambridge Research Labs. There is a skiff cluster, it is a cluster made of IPEX, it looks very artificial and you can log in to this cluster and copy there your source code,

you can compile there and download binaries. There is one more possibility to go on, you can build on your own compiler and run it on standalone IPEX but it's quite huge, you have to have more libraries for Linux and you have to have all libraries for compiler

and all libraries that you need to compile your software. I think the best way if you want to try yourself is to use this skiff cluster that is available and running almost 24 hours a day. Well here are some screenshots of our demo system that we created.

First is just AutoPhoto, the other one is AutoPhoto with vector overlay and the third one there is a bit much or too much information, this is one raster which is overlaid, here you can see black part,

this is transparent black and white raster of technical map but it doesn't fit very much, it's only for demo and also you can see some sites with labels, this should be civic numbers of houses but I'm sorry for the data, they don't fit very much. Here we have technical map again zoomed to be more visible

and again we have sites and also here is grass shell, I performed query for I think it was number 14 here and here we have answer of course, it needs some graphical interface to answer in more readable way. Here is example of use of TCLTK menus,

it looks nice here but in real use it's not so easy because once you start opening TCLTK menus and they are covering your map and you result with 20 open TCLTK menus and absolutely un-visible map. Of course with this X window manager you can change displays,

so I'm using it like on first display here I have grass shell, on second display I usually have only the map and on third display I'm using TCLTK menus, it's possible and it works quite nice but still we are calling, calling for new graphic user interface, must be.

So what are the problems and difficulties? As I said graphic user interface, input for IPLEX is limited only to touch pen, so it is a problem for if you want to use command line. The other limitation is small display, so this is a problem if you have graphic interface, so we really need something intelligent and graphic to help us.

Then we have memory space, if you want to have everything there, it's very big and you are not able to fit it just inside the unit, so you need extensions. If you use external modules you have such jackets, but once you have, I don't know whether you can see it,

but you have only two slots, if you have double PCMCI slots, so once you place wireless here to be connected, then you place some memory cards, but you have no space for GPS, so you can solve it by wire of course, but then you are working with wires and I think that's not the best way to do.

Well, we can solve this problem a bit with use of nice features of Linux, we can have NFS mount and have our disk at our office and use it through wireless, it's also a possibility, we can solve that. And then there is another problem, if you want to use compact-place GPS, which is this one, with VLAN cards,

this is a problem of constructors, it's impossible to fit it. There is no way that I am talking, I am saying this because of Helena, because she was asking me to tell her something about the problems of use of this. And the other problem is the vulnerability of these devices, we need someone to produce this in some strong rubber

or something like that, because once it starts to rain, once it is cold, once it falls on the ground, I better don't want to know what will happen. Now let's talk something, let's say something about possible practical applications.

As I said, you can use it for libraries, you can use it as instance, tourist guide, just imagine that you are walking somewhere, I don't think that you will like it in mountains, but let's say walking somewhere around the town, in parks or through the city,

you can have GPS with you, and the system knows exactly where you are, then you are standing in front of some nice, I don't know, fountain or something like that, you want to know what it is, you just click on your device, and it will, through wireless network, it will download stream, multimedia stream, and it will show you the information, tell you what it is and so on.

You can use it as intelligent card navigation system, you just find a badly parking car, you make a point from GPS, you send it directly to your database, and you are not home with your car, and you already have nice envelope from police in your mailbox, so it's very fast. You can also capture data in the field,

as we had a while before for geology or biology and so on. Also there is nice use for technical infrastructure companies, like, I don't know, gas company, electric power and so on. And here in Trento, there is Wilma project going on, it's short name for wireless internet

and location management architecture, which is project that is, that should make such open air lab from Trento, with wireless connectivity and with several applications for several organizations or for public.

What to do next with this? First is the graphic user interface, it's really a problem. Then we need improved GPS support to support multiple clients, for us to be able to see other workers with IPAC and so on. We need to add some other GIS functions,

perhaps other functions, and IPAC is not the only PDA which is able to do it, as we have seen in this keynote speech, so we should port and develop for other handheld devices. Here are some important web links, I just want to point this one,

if you want to start using Linux on your PDA, here is really everything written about that. And if you want to start using mobile graphs or this one, this experimental package of course, you can see this page on the main graph pages, there is a link. And you can download sources, binaries,

or if you will have Linux installed on your PDA, you can use package, it's already packaged, so you can use package manager and just install graphs by very easy way. Thank you for your attention.

Thank you, any question? Okay, Fernando Jose, a simple question is, how much is this baby grass? No, how much is the price?

Sorry, it takes about, I think now it's about five or six megabytes. The price, price of mobile device? Oh, I couldn't hear you. The price of this?

It's about just this IPAC, I will just play Lego again with that. This is about 1000 euro. This is just this. Then you need GPS, this is about 300 euro. It's cheap one, but very nice. Well, and then you need some jacket. This is about 50 euro, this one.

This one is much more, this one is about 400 euro. And if you want wireless card, it's again 120, 130 euro. So all together, you result in something that is similar to notebook. So it's not a solution for private use,

I think, or for student use, as we were shown with cheap PDA and very cheap GPS. It's a solution for company that really does field work and needs to see the data, I mean map or auto photo and so on. Or it is a solution for rescue services

that need data support in this field with online maps and so on. I just want to point out that there is a similar kind of application developed by Media Lab Asia available on GPL, and it has a GPS support.

Maybe you should take a look into the code so that you don't need to write the code again. And another thing is if you can have a barcode reader, I think it is useful, especially for gas and electricity building and things like that, you know. Well, I think maybe it is possible because there is a compact flash support,

so I think if you just have a look at internet, I think you will be lucky to find barcode reader. There are really many, many extensions. And as you said, for that GPS support, there are several nice applications, so we will have to learn from them and implement.

But now there was no time to finish it. It was just finished, and then like that. It's experimental, as I wrote. Instead of using PDAs, didn't you think about such computers like Husky, which are designed for field work?

They are very robust, and so you don't need to be aware of having damaged those devices. Yes, and what is the price? Much cheaper, I think. Don't you think so? I don't know. This is the first time I hear about these devices,

so maybe... These are computers built in the UK, and they are designed for field work. We can try. For work in hard conditions. We can try, but at this time we have this because it was available and quite acceptable

to be bought from grants. I think that the limitation of these is really that they are not designed for field work. Well, in fact, they are designed to be used in a mobile way, but they are not designed to be used in rain, snow, and in heavy conditions.

Okay, thank you to all. We'll meet again at half past 14 here, so I...