Re: Dr. Ari Jaaksi on Maemo 5
 

Saying that "saving pennies per call" is foolish is, well, foolish.

Those pennies add up.

Re: Dr. Ari Jaaksi on Maemo 5
 

I did, well some pictures of it, I tried to get a real board but the build was delayed and we are still busy with the sw for it.

However it was just a measurement board, no new OMAP3 HW to show. Yet.

I'm attaching the images I had at the summit. We are going to use this to do detailed measurement of the next HW.

It has 5 high resolution double channels (current + voltage @ 50 ksamples per secons) and 32 low resolution channels (current @*10 ksamples per second), plus several digital inputs (to monitor OMAP3 clocking and activity states).

It will be completely open: once we publish it, you get also the schematics and the sw, both the one running on the onboard uC (ST ARM9) and the one for the host PC (scope-like SW).

I have to say that I was a bit disappointed that in general nobody was really interested in kernel development, but maybe the maemo community has developed itself into a userspace-focused group. And that's also what our marketing likes, so I suppose it's ok anyway.

Re: Dr. Ari Jaaksi on Maemo 5
 

Just a warning, I'm not a hardware guy. Not that I wouldn't like to be, but this stuff's a bit beyond me. So bear with me:

• You mean your test hardware will be open, or the final hardware? Or both?
• What is the ARM9 you mentioned? That's not the OMAP3, right?
• Your board has an ethernet port! Cool!

Like I said, not a hardware guy, but I wish I had known what you look like, I wanted to ask you some questions. I'm also disappointed that our top community kernel guys couldn't be there. We need hardware / low-level kernel guys in the community in order for the community to be healthy. Fanoush is my hero. :D

Re: Dr. Ari Jaaksi on Maemo 5
 

Apache Harmony, apparently :( so much for portability...

Re: Dr. Ari Jaaksi on Maemo 5
 

IIUC, it runs java compiled to a different bytecode.

Porting existing code should be closer to a recompile and further from a port to a different language, anyway. It does preclude use of closed binaries, though, unless someone writes a bytecode translator.

Re: Dr. Ari Jaaksi on Maemo 5
 

Yes, having a device that can run desktop framworks and rock from there is a good idea to have. A better one to use. However, we have to be careful not to mix the UI and UX of a 5in mobile to the UI and UX of a 5in computer.

Ideal world says IT can exist in the right context that the user needs, and accessories (docks, wireless keyboards, high speed wireless connects, etc) accent that and help it morph to what is needed without degrading the experience.

If this is what Nokia is doing, then no, they aren't late at all, just a step ahead of where half of the folks that like them are thinking near-term ;)

Re: Dr. Ari Jaaksi on Maemo 5
 

I don't. Check your settings. We may discuss your problem in a separate thread if you want.

Re: Dr. Ari Jaaksi on Maemo 5
 

OK. I have been thinking about our last discussions here, and there is something strange about it. They change focus.

At first, the N900 was going to take the world by storm because it would allow voip over umts /hsdpa (and massive savings, whatever). Then I pointed out that the entire present N-series and E-series line from Nokia allow voip over umts /hsdpa now. With their built-in software. It's not the future, it's the present. Yet I see no storm. (note that there is another reason why voip over umts is not a great idea, technically speaking, read end of post).

Then, the N900 was going to take the world by storm because it was a new "PC paradigm". Or whatever. I don't see how that could possibly be related to always on connectivity and I'd like to point out that the N800/N810 already are this "new paradigm", yet we do not see a storm of applications for them.

Then there is this:
What? They don't have anything? What have they been doing all that time? Let me recall you the timeline:
-end 2005: 770
-end 2006: N800 (actually early 2007)
-end 2007: N810
-we are now at end 2008, and talking N900 all dancing singing in a near future. And if they have not even started, we are talking end 2009 at best (manufacturing delays, etc...). Yet, we now have real competitors like: iPhone, Android, eeepc, etc...

Am I the only one to smell the stench of vaporware? At best, the maemo division is severely understaffed (which is never a good sign). At worst....


********and now for something completely different**********

The problem with voip over umts / hsdpa (or full time connection of anything over umts / hsdpa, actually).


Software engineers and designers of digital devices like to think that electronics is a nice collection of zeros and one. Unfortunately (for them), you still need the analog part. Amplifiers. Antennas. Heat sinks. Power.

This is the problem with umts (or hsdpa, which really works in a similar manner, check the umts article on wikipedia for more details). When the system was designed, spread spectrum was all the rage. umts is designed to run as spread spectrum on a bandwidth of 5 MHz. This is quite a wide bandwidth.

Unfortunately (for us), for spread spectrum to work, you need very good linearity. After all spread spectrum was designed by mathematicians and in the world of maths everything is nice and linear.

There comes the analog electronics engineer: "you can have wideband, linear, low consumption: pick two". Linear amplifiers waste power and get hot. At the time when umts was designed, people put that problem under the carpet expecting the technology to improve. The technology has improved, but not that much.

If you run your smartphone (even a modern one) to be always on over umts (which is necessary for voip), the battery lasts 4 to 5 hours. If you download full blast, the battery lasts one hour and the phone is piping hot.

The iPhone has the exact same problem, btw. Customers complain that if they really use it for light browsing during the day, the battery is exhausted in about 5 hours.

My e51 can run voip over hsdpa (and apparently vodafone did not yet block it). I tried once, the battery was exhausted in about 5 hours. It has a 950 mAh battery.

The N810 has a 1500 mAh battery and you all know how long it runs the show when you use it for something more than sleep mode. If you want the N900 to run voip over hsdpa for 8 hours a day, you will need a battery roughly twice as big. No improvement on battery technology is on the horizon, so it is going to be twice the volume and twice the weight.

This is what always online means. Think about it.

Re: Dr. Ari Jaaksi on Maemo 5
 

Thanks for the comments. This topic interests me since none of us are sure where exactly the UI in Maemo 5 is headed.


"The one and probably only thing that makes the tablets different from most competing and semi-similar platforms is the fact that they are, in fact, small desktop PCs and you can port desktop applications with little, sometimes no effort."

Really? That's the ONLY thing that makes the tablets mostly different?? Not the battery life? Not the nice touch screen display? Nothing else at all? UMPCs let you run Windows desktop apps - so the tablets about as distinguished from other small platforms as UMPCs? :rolleyes:


"Others might want UIs made only for mobile use and applications that are simple enough to work with these UIs. No problem - there's a number of such devices out there, go get them.)"

ummmmm, the tablets have a UI made only for mobile use... and mobile UIs are not just simpler than other UIs; they're different. how often do you saw a D-pad on a full size keyboard? zoom keys? switch view key? can we please have some respect for mobile UI design??


"Now, it seems that at least technology-wise Nokia stays with the desktop-experience here."

"stays"?? to date, the tablet have very little in common with desktop-experience.


My interpretation is that you gain little from using desktop frameworks if you dont do it because you want to run desktop applications on the device. So I hope we'll see future devices and generations of Maemo moving even closer to the desktop. In an ideal world, every popular desktop-application should be in the extras repository, properly hildonized"

be careful with generalizations. for example, very few desktop apps have support for a touch screen. so I hope you are not advocating that we run apps on mobile touch screen device even though they do not support the touch screen? I have seen some of these desktop apps shoehorned onto the tablet - they can't possibly work well and can I just say ugggghly!!

I think we probably mostly agree with you and I am just saying these things to try to get us to be more precise in our comments. :) I think maybe the proper point is that we want to LEVERAGE desktop apps, properly adapted and hildonized, and not just ported and running on a smaller device.

Re: Dr. Ari Jaaksi on Maemo 5
 

Where did you get that crazy idea? Available to the community, yes, internally ...