Handful of Nothing

Apart from the title, that is just to grab your attention I really wanted to make sure some people understand one thing: Knowing how to criticize is a crucial thing to have your message heard. Also I wanted Maemo Developers and some heavy users to know another thing: The most important thing your software or your device must have is a target group, the group to which you design and develop your user interfaces.

So, the guys over tabletschool, were very kind with our project, but What I want to talk is on the end :

There’s been a lot of rather strong feedback (EXTREMELY strong, some of it!) for this article already, which is the main reason the ITS has avoided editorials. :-)

This would make me sad, if I was not surrounded by great Developers (physically or in the net - IRC, Gtalk) that are concerned with more than their own belly button. Let me clarify some thing that may help you understand better what is “User Centered Design”:

We picked a user profile. We analyze their needs, We see the competitors, We try to solve the USERs problem and most important: we try to solve that “User” profile problems, not EVERYONE’s problem. Why? Because it’s simply impossible to do so.

With that said, I then ask all the thumbs down guy here on maemo planet to read the following sentence:

“If you don’t agree with the basic concept behind canola, the simplicity and focus on consuming not editing, then Canola is not for you.”

But hei! What is the beauty of open platforms? : You have choice. You have the Hildon UI, you have KDE so you’re actually more well served than those users that do not care for the 95% of features you miss in Canola or any other simplified UI. They care about User oriented interfaces, not a powerhouse interface, and I can be wrong with that but the massive amount of emails I receive gives me the hint that there are a lot of user who wants the simplicity way.

Don’t get me wrong, this is like Itschools not about technical matters. I’m the one here in the company who says: I don’t care what Acronym you’re going to use, I just want the best user experience with everything that we can do on it. I want the cherry on top of the user’s pie. If you don’t think it’s important, pass it forward there are a lot of people that do care about.

And just to make clear: ~80% of the email asked for features, complained about bugs, requested improvements but none asked to include regular UI elements on the software. Why? They understood the goal.

So please, stop felling “hurt” when people talks good things about things you don’t agree. I think Hildon is a desktop like UI, and thus provide a desktop like experience. Canola tried to fill the other gap. If users think they should have more Canola, it’s their rights, as much as you think Hildon should “be ok for them”. It’s not. That’s why the use Canola (or the other apps) and that’s why We focus on python development. Because we believe we need to do more, in a simple and python give us that power.

And about the whole Canola thing: If you also really want to criticize, I know it takes time, but read the forums in ITT. Look at our way of dealing with users. We do treat them as co-developers. We do listen, we do change. But we stick to our goal.

Also to make clear another thing, this time for the tabletschool: http://tableteer.nokia.com/tableteer/os2008/team_marcelo.xhtml, INdT stands for : Nokia Technology Institute and Canola only exists because Nokia indeed wanted it to exist. As I said Canola is in opposite direction to Hildon, but that doesn’t mean Hildon is not good. It just means that the user focus is different.

Concluding: Please understand the concept of user centered design, before trying to compare things. And please be ok with the fact that it was not made for you. As it probably was not made for the one who made it. Or do you really thing all the Devs in the team were happy to hear me shout that the thing was not “smooth” and fluffly enough?

There’s a LOT MORE after the click :) But please, bare in mind that the topic here is not about Canola, or hildon or this or even your taste. It’s about being humble enough to figure out that sometimes you may not be the target.

Continue reading…

Jeff Mancuso, dev of expanDrive :

With high-level languages and good libraries, small teams can create great products at a rapid pace. We realized that we could write applications for the desktop in the exact same way. We rewrote SftpDrive from top to bottom in Python, with a GUI in Objective-C. It’s called ExpanDrive, and it took 1/3rd the time that SftpDrive took to develop.

read the nice post here

That’s what happened to us on Canola and other projects development, and no Canola2 was not a rewrite, it was a brand new application from scratch, using totally different technologies and we also achieved the same performance, including : more features more polished and also : more maintainable and the best one : easier to get a new dev on board (start time quite small for simple / medium tasks)

Of course this is not a silver bullet, a magical way and you should only consider giving it a chance if you haven’t : ) of course there will always be “no go”s and other reasons not to do it, but for a lot of user space applications it’s just an amazing tool to work on.

Hei guys :)

I would like to say that this is coming, and if everything goes right you will have it at least for linux :)

Video :

It’s fully functional, has some small problems in fullscreen (some screens are not yet ready for fully scalable uis ) but you can test everything in windowed mode.

Also, there’s two more tutorials coming, and gustavo published one simpler one for those that were not too keen to try the more complex youtube source code :

Gustavo’s simple model tutorial for canola2 plugin development

We think that if we can release this tar file (the generation is also automatic, so it will be updated with each new release) it would be quite nice for developers that want to try out.

For windows and mac users there’s a solution (not a good one but..) you can run a live cd of ubuntu (in vmware client for example) and do all development inside, the same way you would do it on the device.

Kudos for fleury and the rest of the team, and I will now try to do what it takes to release this asap :)

Marcelo

I made this video while ago, but it’s nice to see that are step by step moving to it :)
Youtube is coming to beta6, and we are working as hard as we can to make Upnp available for Beta7. We are just not satisfied with the user experience as it is. Of course I need to add this here or the developers will kill me: don’t take the video for granted, bad things can happen on the dev path blabla :) you got the point!

From our experiences we know that those guys will generate a lot of requests / bugs and specially Upnp will be a lot of work to fix compatibility to custom servers. That’s why we are trying to improve Canola2 itself to the maximum so then we can focus at least for a sprint on fixing youtube, upnp stuff.

Again, time is the only villain : /

Info : you can see that the menu is actually “media servers” then a ball for the upnp devices (names will probably change) this means that we are open one room, just for those guys, so if people want to implement support for other servers/standards they can be placed here :)

Yes, one more beta in this long road to final version :) if it is going to ever be a final version (just to be trendy we will be always on beta).
This is a huge update on the core of Canola2, and we tried to make it as stable as possible.

There was a small character bug on the scanner, fixed and already available (maemo2 packages) and apart from that no more serious problems. Of course we will try to see how to improve things, and we are investigating some video related issues, but as far as we could have information, it’s more the case of too high resolution videos (that plays on mplayer out of canola) but not on embedded mplayer.

At least seems that the problem with the scanner is gone, some people still have crazy files that don’t get indexed, but the least we can do is receive the files, test, and if it’s a bug we fix and release it again :)

Also: youtube and last.fm are not there, because we need a little bit more time for 2 things:
1. to separate bugs from beta6 to bugs from youtube and last.fm
2. to fine test it more. We had some other projects priorities and those 2 plugins were delayed :/

But, I we’re more than happy to receive your feedback, and if you send via the site, make sure you place your email, I answer every single email myself, no bots allowed :)

Marcelo

After a crazy week, we finally did it. Of course there will be now new bugs, but that’s life and we are ready to solve them. If any serious bug appears we will release a fix asap.

Hello everyone!

So we made it! If you have one of the following supported distros on your devices you can already install it quite easy going to the website and hitting install now. Just make sure you unninstall canola1, if possible reboot (to get rid of the canola scanner) and you ready to go.

While we do want to receive as many bugs (send them there) and suggestions / critics etc, I can already say some of the things that we know already, or wasn’t implemented because of time, and it’s already on the todo list:

1. Scrolling sometimes “stick” when I click too close to the right border of the screen.
   1. yes, indeed we placed a ” scrollbar ” like area there to allow you to scroll faster on bigger lists but unfortunately the device cannot handle pretty fast animation so theres no visual clue what’s happening. I found myself thinkting that this was a bug, when I realized that gustavo made it and I asked him to increase the area.
   2. Solution: for now trying scrolling a little bit more to the center of the screen.
2. First touch on main icon is slow to start the animation. Yes, We are hard working on this, as this is quite the first impressions. Just because too much things are being loaded in first use. After that is quite the speed we want.
   1. We will fix it =) don’t worry.
3. Scrolling sometimes looks ” locked ” : actually you’ve reached some of the limits, and there’s no visual indication of that happening (actually there is… the scrollbar but it’s easily missed)
   1. So we are going to implement some kind of bump (unfortunately) to show off that you’ve reached
4. There’s no 770 support YET. Sorry, 770 demanded a lot more work, but we are working on it and it will be supported. HE is ” theorically” a BORA distro, but we don’t want to mess if your device so.. please allows more time to test and make sure we are not going to screw your precious installation
5. The way we wanted to support the hardware keys are not yet implemented so also hold on on this one ( I’ve posted about it before)

So, there’s a lot of rough edges to polish and we promise that we will be working hard on this (of course after the parties) and the final release will be as complete as we promised.

We really need your feedback this time, and we count on you.

So hit the Canola2 site link button and make a hit

Hi, as promised I’m going to talk a little bit about the photo viewer on the new Canola. This was a little bit controversial design and got of course the normal suspicious overview by the dev team. I fully understand them, as most of the time I have an idea but it just don’t fit in all cases, and they are a great in setting “hei dude calm down” points and to of course keep my feet on the ground sometimes.

But this time, I was really after this. I had a problem with thumbnail pages that are a really solid and fixed grid. I always thought that you photo view should be the closer to the real thing, and real thing has a real human made arrangement, where you try to combine and use the empty space as much as you can. So the goal was: How to be more “real”? More “personal”, more “human” or more “organic”?

Of course a real simulation is not possible (yet) due to the same problem like always (hardware performance) but come on! We are on a mobile platform! And this is running today only on high end computers or videogames like the Playstation 3.

But one good thing can always comes out of a restriction. After analyzing the photos, even a very small difference between the photos can give this more realistic look and feel. So it is possible to do that only by having landscape and portrait oriented photos. Then you put some small png images to the corners to create the 3D illusion and we finally found a nice format. We defined that at least height was going to be equal to ease some calculations in the thumbnailer and also in the drawing engine. But this simple detailed gave a much more “organic” feel to the view.

The internal screen

So, we are on a touchscreen based device, and as an Interaction designer I’m aware that “second click” or click and open menu are not a good solution for this kind of device. And personally I think contextual menus on touchscreen devices that appears when you hold the stylus down just one of the worst decisions ever. It’s again an attempt of bring the same behavior from the desktop to the mobile without adapting it.

To solve this we decided to have a “almost” fullscreen preview of the chosen photo and taking advantage of the widescreen and also creating consistency with the album art browsing screen, we decided to go this way :

This screen allows us to create a area for image information and maybe buttons to connect it seamlessly with other mashup content in the future, but also gives us contextual menu on the corner. We are actually not going to use in the december release this menu, as most of the options will be available in the fullscreen view but this already enable plugin developers to expand the functionality creating for example : Twango uploader plugin, Flickr, Picasa Web album or any other kind of send to application.

And to end the sequence, there’s the fullscreen view. But some of you may already say : hey what if I don’t want to see this intermediary screen? Well, you can double touch the picture and go straight to the fullscreen (actually this will be hard on low sensitivity screen) but solves the problem adding the shortcut.

The fullscreen view brings you the commands like the player : play, previous, next and in the other side: zoom and rotation.

The screen appears to not have enough hit area if you think that the Internet Tablets has a high bevel as a border, but the hit area on these screen will be actually bigger than the buttons, to allow you to click always correctly.

What’s more in and really counts: We gave an special attention to the details to make the whole usage more “smooth” so even when you have slightly bigger load time (bigger pictures being rescaled) you will have nice fade in and out transitions to reduce the loading time sensation, as well some sequential loading transitions to give the illusion of flow.

What’s out: We are not able now to make a fully powerful zoom feature, or other transitions (mainly tearing and redrawing the wholescreen smoothly limitations) and features like renaming or tagging. But as I previous stated, we already have the place to do so (menu) and there’s also small things in my head that I want to put to work next year that will allow better management, even with my belief that Canola should be more a player than a manager of media.

The Drawbacks

We all know that not everything goes the way it’s planned. There’s of course bad things about this design but I really like the way it came out.

1. What I think is really going to commented by some people is the different aspect of organization (by rows) and that some of the photos in a sequence will be out of the screen. This will be the price to pay, but I think that in this particular case it will worth.

2. The other drawback is that if you don’t have at least pictures in both orientation you’re will end up not seeing the effect and worst: if you have only regular pictures the “tearing” appears a lot more.

3. The need for a really fast thumbnailer: EFL came to help again, but this is quite complicated when you think about leaving room to allow users to manage it, and control how it works.

4. Memory consumption when kinetic scrolling : we need to provide a great experience to do not be boring. so if the user “throws” the screen to the side, it should “answer” to that. This is for example what made the iPhone UI so “attractive”. But how to do this if you have 100 or 500 images in a device where you should be really careful with memory and mainly with reading data speed? This gave us a lot of headaches. (Leo knows it)

We also did a regular, static grid version (actually used in the album browser view) so if people doesn’t like the more organic one we can maybe provide a way to fall back to the regular way.

I really believe the gorgeous touchscreen with have to work with, and it’s unique widescreen nature provides the perfect background to do something like this. So the photo view is truly an exploration of how to create exclusively for the horizontal use in a high definition widescreen. If we have this, let’s use it to it’s maximum.

Real demonstration of this screens : you can take a look in the video posted previously here (sorry for the low resolution)

Conclusion: The drawbacks are there off course, but I’m really please to push the boundaries of the platform again, and now on both fronts: design and development.

Next topic will be player screens and how Car users will really have a reason to do not worry about the smaller buttons on this version.

This was the final piece of the fast lecture I had the pleasure of presenting at WUD’07 Recife. After talking about UX practices, our team, and how things can change when you do user centered design I made a small demo of the new Canola explaining some design choices and how it evolved from the past version.

The video is a nice opportunity to see the UI for music and Photos, that even not being the most important on Canola is now for sure one of my favorites.

Bellow you can also watch the video of the same UI but now running on a real device. Sorry for the quality but here the goal is just to show how smooth we were able to achieve with EVAS. The guys are just out of this world.

Click here to watch the second movie

Click the image for more info.

INdT giving sequence to it’s software tech days is going to make another on november 20th and 21th in Joinville. This is a great opportunity to get in contact with Maemo development. Not only you will have introdutory course on developing for Maemo but also be able to attend small classes on Python for Maemo, Qemu and even the microBrowser extension courses.

the best part? it’s FREE! =) as in free beer!

Maemo may not be giving away 10 millions dollars, but it’s going near everyone to teach on person how to do stuff :) :

Well, it’s not that, but this shows how easy is to create your own plugins. As a testing ground for our plugin architecture Kenneth is implemeting some Carman screens / sensors and data acquisitions.

Canola Menu

For those not interested in plugins, carman etc It’s really a peek into the most recent version of the main menu. Apart from the video quality you can see how smooth it is. The point for me is why to design a menu like that

Actually it was only a matter of observing the users and gathering feedback on the first version. Things that were clear and already posted here:

1. Scrolling on main menu : BAD. Why? A lot of users simply didn’t know that there was more options when scrolling down.
2. Arrows everywhere : BAD. Solving the problem above with arrows was the worst thing ever. It didn’t quite solved it and made the ui a stage for almost useless visual components.
3. Arrows not grouped : the arrows also were not “grouped” together. While the up and down were near by, right and left were quite apart, but there’s was not better way (IMHO) to do it. But it doesn’t justify the mistake.
4. Useless scrolling and not enough room: While it doesn’t appear too much with a 3 item only menu, it would be a hell to have 5,6 or more items on canola menu. Therefore the need to use something much more simple: the grid. You can run, you can hide, but in touchscreens the grid most of the times is the best, time saving view to use.
5. Textual internal menus are bad. While text can make things a lot clearer than icons, it also demands a lot more room (sometimes) and can be trick for a grid view. So to solve the problem of space we have a grid into a grid menu, or like a called here: menu on menu. The only difference is that we can fit up to 12 items on the first level then 10 items on internal menu. This takes canola now to a incredible capacity of 120 features in a range of 2 clicks. PS: WE NEVER going to place 100 items there =) but flexibility is never enough.

So this was the starting point to design the main menu, and after a lot of iterations we ended with a cleaner, simpler yet visually appealing menu that provides 2 level access to a number of applications much bigger than we need now and in the near future.

From the first 2.0 video, here you can notice that we really killed the scrolling. To keep main icon hidden is a incredible mistake that I really apologize for doing it ;)

Also in the image you can notice that we brought down some elements size. Actually we want to keep a small “compatibility” with the older Canola, and enable touchscreen use at it’s maximum, but I really (personally) wanted to get rid of :

The back button (letting only on hardware keys)
The menu button (also on hardware, but now harder because in the n810 is on the keyboard : /)
The home button (also on hardware, but this can break the device’s home button functionality while on canola)

In a sense, what I’m looking for is a almost button free (or at least when possible) User interface. Let people touch things and do things by context not only buttons. I know it’s a little bit too much, but come on, you need to have a bold target.

Why? Simple: the carriers would really bend to openness and would be a important and strong influence on other’s platforms. To be able to create mashups will be just incredible.

Apart from that, I’m not that keen on Java, but it’s a nice move from google, and as Gustavo said to me tonight, a lot better than use Gobject / Gnome components. It just don’t fit. (sorry maemo on that point, but that’s why we also don’t use). Gnome is so 35 minutes ago - ;) and while it really fits well on the desktop is a drag on the mobile space.

It seems also that google has a very nice VM for it’s Java bytecode. I think sun is really happy today, or could it be Eric’s heritage? Only time will shows us. For me it smells Google + Sun + Apple against the others. (kidding… or not)

So Google, after all the speculation about the gPhone is going even higher: Android is out and it’s the Google mobile, open source, and free? version of Windows. To be more specific, Windows mobile.

I know this is a pretty simple conclusion, (actually a comparison) why Google, heading the OHA, is aiming in nicely providing a lot of people a open platform, stack, or whatever to everyone interested on creating devices. Nice nice nice move google and more, after buying Android, expectable.

To imagine that Google is actually going to make sure all device made by the involved companies can run it’s web services nicely (in or without a browser like the Gmaps on iPhone) gives you a hint of how geniuses this can turn out to be.

“With devices built on the Android Platform, users will be able to fully tailor the phone to their interests. They can swap out the phone’s homescreen, the style of the dialer, or any of the applications. They can even instruct their phones to use their favorite photo viewing application to handle the viewing of all photos.”

Being honest, if Google manages to pull this thing off as it is bragging the world it will, and also if it can convice the actual manufactures to keep it open this thing can be a major blast at least in the smartphone world (Maybe HTC is tired of giving MS money and motorola cannot sustain it’s montavista linux prices) and really put Symbian and WM to work. But, really how they are going to pull this out :

Android does not differentiate between the phone’s core applications and third-party applications.

It can be, maybe with this :

…a custom virtual machine that has been designed to optimize memory and hardware resources in a mobile environment.

I like it. To have another good competition against WM for me is enough reason, and I’m all in for big players in the mobile world.

The best thing is to see that EVERYONE is jumping on the mobile bandwagon. Nice opportunities ahead.

And to end the post …

Breaking down application boundaries
Android breaks down the barriers to building new and innovative applications. For example, a developer can combine information from the web with data on an individual’s mobile phone — such as the user’s contacts, calendar, or geographic location — to provide a more relevant user experience. With Android, a developer could build an application that enables users to view the location of their friends and be alerted when they are in the vicinity giving them a chance to connect.

Please Google tell me how are you going to convince the manufactures and carriers (ok they are in the alliance) to allow developers (with certificate and signed code only?) to access people’s Agenda, Money (calling) and so on?

Ps: Oh yeah! I was almost forgetting this : Why do not take the Linux fragmentation from desktop to mobile also? =)

If you always missed some good video management tool for all you internet video needs, here’s a good solution : The videocenter application.

The application is a hub for internet videos, local stored videos and in the future for other video sources, like mythTV (explained in arthur’s web site here). You are able to stream, download organize, and soon enough to do a lot more =)

You can check out a nice vide made by my buddy Arthur (Morpheuz) :

The application is already available as a beta in the garage website

As any application I think everyone involved would be pretty happy to receive feedback / bugs / suggestion so please help the team behind it.

So, today at #maemo a lot was talked about these “kinetic” or inertial scroll, but agreeing with someone’s point I said the following:

We should not just simple try to copy Iphone or other device interaction. It’s all about doing it, testing it for our devices, improving where it needs to be improved and the most important: trying to add more and more usability and meaning to that.

Agreeing also with Nikoo (from the comments) that we need more than just the scroll. The iphone actually did the second step for the alphabetical lists (has a letter indicator on the right) from where you can browse faster on your list. But what about the other steps? Or when the list is not alphabetically ordered? Or when you reach the end and you want to come back to the top?

Anyone who bought an Ipod and loaded it with some 10.000 songs must have been felling a little bit anxious about having to slide the finger across the virtual wheel for a lot of time. And to come back… slide again in the opposite direction.

For Canola, the good thing is : the inertial, or kinetic scroll frees a little bit the area where it used to be the scrollbar so a smart “scrolling” indicator can be placed there, that could also provide “jumps” in the whole list. I think that a combination of being able to jump across the list, use the finger scrolling and also to have the ability to hold down while scrolling (as a regular scrollbar) will be the best solution. With that we will not need to use dirt hacks (the key + arrow) used in the first versions of canola.

So I think this should be even more our way into making maemo a better “place” to learn, to improve, adapt, test and discard when not neeed Instead of just copying for the eye candy fans.