* Posts by Mark Wilcox

6 publicly visible posts • joined 18 Nov 2009

Why Nokia failed: 'Wasted 2,000 man years' on UIs that didn't work

Mark Wilcox

Oh and just to be clear

"The question as to why Nokia surrendered its independence lies in why it took so long to engineer a competitive UI, and then under new management, decided that it couldn't."

If you read the blog carefully (sorry, it's long) you'll find that Nokia hasn't decided it can't engineer a competitive UI. Indeed it still needs to do that to avoid completely nose-diving in the smartphone market this year. My speculation was that they don't feel they can compete with the likes of Apple & Google alone, their pockets are no longer deep enough.

Mark Wilcox

Timeline a little wonky

Thanks for noticing my epic.

You've got the timeline a little wonky. Hitchcock/Alf and the Maemo 5 GTK+ based UI toolkits were both started before the acquisition of Qt. Otherwise this is quite accurate reporting for you.

Oh, and I was actually working at the Symbian Foundation (freelance writing and then contracting before going permanent) before there were any employees at all.

I think it's too early to count Qt out yet within Nokia's strategy, even if it's not on Windows Phone.


Dell skunkworks brews ARM server future

Mark Wilcox

You need to look a little further ahead than next year

ARM are scaling up performance much faster than Intel are reducing power consumption. Electricity bills are THE major expense for a data center...

BTW, I believe AMD already has an ARM license via its purchase of a certain major graphics vendor.

The Cortex-A15 recently announced has support for sufficient memory for servers, hardware hypervisor support and significantly increased performance over the Cortex-A9.

As we've read here on El Reg, Intel has been competing via dodgy business practices involving large "rebates" (or bribes as I prefer to call them) for the last decade rather than innovating. 5 years from now the server market could look rather different.

Google open sources $124.6m video codec

Mark Wilcox

Theer's already NEON support too

For those asking about mobile hardware support, even at the low end this should run reasonably well on devices released in the next year or two. It's already got optimisations for the NEON instructions for newer ARM cores, which really blurs the line between "software" and "hardware" codecs. It won't be quite as high performance or power efficient but taking into account the lower screen resolutions toward the lower end of the range, it shouldn't be a big problem.

Nokia to cull Symbian in 2012

Mark Wilcox

Who was doing better before Apple came along?

Thanks for the response Andrew,

"Navikey(tm) was the best, while S40 was OK, and good enough."

Yes, Navikey was the undisputed leader in its day, but what was better than S40 in its time? I worked on it and Ericsson's alternative, even on Panasonic, wouldn't have dirtied myself with the horror of Moto's effort. Some of the Japanese stuff was several years ahead of course, but with the usual uniquely Japanese slant.

I maintain S60 wasn't badly designed for its time (although I didn't have anything to do with creating that, only some devices based on it a few years later). By the late 2nd Edition devices it was just horribly cluttered. Apple didn't solve the clutter problem with a cunning design, they just removed the clutter (and a lot of features that many operators considered "essential" until someone proved otherwise).

There's a lot of truth in Roger Nolan's piece, Nokia is engineering led and in some cases the power ends up in the wrong places - the engineering teams creating S60 had almost no accountability to most of their "customers" creating products when it came to the design and UX aspects.

Mark Wilcox

Misleading title, but largely accurate

It would seem reasonable to read from the public statements that Nokia are still talking about using Symbian in FAR MORE devices in 2012 than they are currently doing now, so "Nokia to cull Symbian in 2012" is extremely misleading.

They also seem to be planning to use Maemo in FAR MORE devices than they do now. All of which implies that they expect to use Series 40 far less, but continue to run a multi-platform strategy. Lets see if that turns out to be the case.

Have you any reference to Nokia denying that they'd replace Symbian with Maemo "at the top end". I don't believe they did, just denied they'd be replacing Symbian with Maemo completely and re-affirming their commitment to Symbian.

Since both Symbian & Maemo are both getting total UI re-writes from scratch with Qt, it seems a little early to call relative portfolio/market shares. Plus Maemo is currently a "one device at a time" platform, rather than a true multi-device platform, and there's a lot of work to be done to make the transition.

To say they're making the switch to Qt 5 years too late is probably true but a little on the harsh side. Qt will be shipping in devices next year and powering their native UIs in early 2011. In early 2006 Nokia was still king in smartphones and everyone was praising the N95 - iPhones and Android were nowhere to be seen. It's extremely rare for a market leader to radically change their offering to fend off as-yet unseen competition.

I don't believe Nokia has an inept record in UIs at all - the Series 40 UI was the best in the world for years, and the original S60 design was sound, it just got over-complicated with far too many features added badly over time and trying to make it seemlessly extend to a touch-based UI was always doomed to failure. Services is another matter - I only hope these are just the pains of a company fairly new to the market!

The Series 80 and 90 UIs needed to die because there was too much developer fragmentation - it isn't necessary to build an entirely new UI framework in order to differentiate device UIs. S60 just wasn't flexible enough to be the one-size-fits-all option.

All that said, I generally agree with the theme of the piece. The best article on Symbian/Nokia/Maemo I've seen on the Register for a long time!