As I detailed in my treatise on branching/forking, on a large project heading towards ship date, it’s ok to branch supplier, but only when the consequences are understood. Part of the problem with branching is that it can be very hard to know the rate at which your 100s of engineers are changing bits of code. There is often no central policy maker/code submission approver.
On my last project I wrote some simple python to interface with perforce to generate a graph of the number of supplier files which have been modified. Since perforce can give you the contents of a file on a date or change list, it’s posible to construct a branching trend going back as far as you need.
The graph below shows a period during the first two months or so of a S60 based project. The initial peak in symbian branches came from modifications to support the baseport made in situ to the Symbian kernel code. (We managed to un-branch these quite quickly).
Figure 1: Branching report (number of files branched per supplier)
This sort of visualisation is really important for a project leader to keep an eye the code base. But it also represents one metric of the fitness for purpose of a mobile platform.
Symbian have published a paper which I wrote for them. It’s based on the iTunes remote control prototype which I provided on my blog some time ago.
iChoose – an Introduction to AJAX on Mobiles
This paper provides a short introduction to AJAX (Asynchronous JavaScript and XML) and discusses how its concepts transfer to a mobile environment. In order to highlight the technologies and APIs which are associated with AJAX, an implementation of a Web client/server application is provided (iChoose), which can be used to understand how the flow and distribution might work in a typical AJAX style application
I’ve written before about how I don’t think on-line click through adverts work. I’m pretty sure that most people are exhibiting banner blindness – the brain’s way of shutting out information scattered around the top and sides of web pages which are matched as adverts and moved into a blind spot.
Lame attempts to disrupt this self conditioning include rolling those annoying flash adverts across the page and have it obscure the main article. This seems an especially popular approach for questionnaires, with everyone from the register to NHS direct utilising the idiom.
Here is an interesting paper from HP concluding the results of a behavioural experiment which compares brand image and advertising power of print outs verses on screen banner adverts.
The results show a significant improvement in brand recal and perception with printed out brands compared to on screen only.
It would seem to suggest that it’s better to plonk your car rental ad on the google maps printout. Even if this means having the user type in a URL.
When I print something out, it’s either because it’s a longish PDF that I wouldn’t mind reading on the train or an itinery, which I will be checking frantically all through the trip to octuple check that the airplane times.
Initially printed web pages do not seem attractive and are usually represent a tiny fraction of the content a user browses for. However this very selectivity in printing may actually result in ‘multiple hits’ from many different viewing angles (breakfast table, airport, bed) – which may make it a premium spot, like the ads between corrie.
Oh c’mon, if anything is a sign of a bubble it’s the hippy start-up voice, you know the one. The kind of voice which girls do when talking to two day old puppies. It says “sowwy, website down. Our fungineers are working ever so hard at fixing the problem. LOL ”.
If I really believed that these corporation’s employees worked in a Charlie and the chocolate factory style research facilities and drank nothing but innocent smoothies, and shat out corn syrup everytime they went for a dump, then fine.
But really, the reality is closer to a sweat shop than a sweet shop I’m sure.
Here is the worst one so far….
Flickr partnered with Picnic to integrate photo editing into their service. On loading you get a progress bar and instead of the usual “loading swatches”, “listing fonts” state messages familiar to photoshop users, you get messages stating “painting sky, sprinking dew, fluffing clouds”.
So the photo shop messages are not so great, but they are useful for helpdesks for dealing with problems with application startup. If it freezes on “listing fonts”, then that might be quick and simple bit of information which the user can communicate.
Aside from that, there is no state entitled “unwrapping scotch eggs”, which in my books does not a picnic make.
I broke my leg around 2 years ago in car crash, and had been having a bit of pain recently. I didn’t get a NHS appointment very rapidly and ended up seeing my consultant privately.
I was a bit shocked when I saw the x-rays. The titanium rod which had been drilled and hammered into my femur had snapped in two.
Apparently it’s metal fatigue which would have caused the problem, rather than a sharp blow. (Imagine a shatter proof ruler being flexed again and again till it eventually snaps).
I would have known if enough force had hit the metal to break it because my leg would be in several pieces.
Anyway, I asked if I could get a print out of the xrays, and the xray department offered to burn me a CD instead. Fantastic!
The CD contains the raw XRAY data, and a program which can view the data and export JPEGs. So I was told to take the disc along to the NHS clinic to avoid a long wait in radiology, which is possible since the royal london use the same computer system.
We are so used to exchanging ms office document with each other in business, that it’s difficult to appreciate the limits in place when exchanging data between two different health care systems.
The CD is useful to me the patient, since the I can consider the images after the appointment and go over what the consultant said with my family etc. But also the source data is there is a well defined format which can be used in other health care settings.
I nearly wept…
Mobile is not a very friendly environment, and it has been interesting to see how google deals with roll out of services in such a fragmented landscape. I thought that by now we would see some sort of mobile google app layer – some sort of runtime which can be installed on every handset so that google services could be deployed seamlessly. I reckoned that google gears was the key, and that getting that into the browsers to support off line apps would set the initial conditions for google dominance.
Mobile gears was announced for windows mobile, but that only addresses a tiny fragment of the market. We can see that Gears is evolving with the geolocation API, but what about Gears as a credible mobile development API? Google doesn’t need Android to succeed to make money, it needs a widely adopted platform to deploy it’s services and a platform which provides a good user experience and some place to stick adverts.
Not only does it cost a lot of money to run a service over many incompatible handsets, but technically it’s a gnarly job. Android feels like a backup plan – a seed for the market sponsored by google, rather than core business.
Why not Gears on S60? The S60 browser in S603.2 supports plug ins for rendering flash and video, in fact just about every browser support Netscape style plug ins for handling media content. So, what would be ace for google is to have gears running on all major browsers for their AJAX apps.
I’m not sure what the industry tone here is though. Nokia have much to lose due to their ’service company’ aspirations if Gears installs across multiple handsets. Because once you have a single sign-on to services from google, it means that bariers are lowered for existing subscribers to try out new services (you really can’t make users on mobile type in their email address, password and memorable phrase each time).
Problem is that Google is a much stronger service brand, associated with search, mail, maps, even calendar. I’ve unintentionally transitioned to using google calendar after trying it out for a few months, and i’m desperate for a mobile client which I can also use off line.
Can Nokia block or are they blocking the deployment of Gears by making it difficult to integrate unsanctioned plug ins into the browser? If not, then were is Gears for S60?
I can certainly see why Nokia would prefer to go down the silverlight/widget route rather then opening up the browser.
What I am finding particularly alarming at the moment is the term UIWOW, which is a term typified by a recent Nokia document called mobile design showcases. UIWOW seems to mean, “cool moving and cross fading goodness…yay”. From what I gather, UIWOW is measured on a scale between 0 and 10 nerdgasms per second.
Now I’m all for exploiting the aesthetic usability effect, which supposes that in two UIs which behave the same (in terms of clicks and key presses); the more attractive UI will be perceived as easier to use. But having engineers bolt on UIWOW for the sake of eye candy, is not really a good idea.
I can sort of relate to the term UIWOW, because “WOW” were my words when I first saw a Silicon graphics workstation with its rotating 3D menu structure. It was so incredibly fluid and smooth. My main thought was that the hardware was so powerful that it could waste time on animating UIs. It’s cool for a while, but actually the delay between screen transitions is too long and it gets a bit frustrating.
I found a video on youtube, skip to about 4:20 to see the menu, it’s very impressive for a system conceived in the late 80s.
My objection though is the use of UIWOW to mean “Better UI”.
UIWOW was also pioneered in the game industry. Every console game had their own fonts, and fancy rotating menu display and almost all of the designs inconsistent, showy and frankly quite rubbish.
Animation and motion can be powerful tools for highlighting a change of state and for assisting in establishing a narrative for a UI.
The original iPod is a fantastic example of an affective use of motion. Each level of the menu slides from the right sliding screen combined with a button which always takes the user back ‘up’ a level. Without the screen slide effect, there flow from menu to sub menu is disrupted and the result is a poor user experience.
As a corollary, imagine on your PC that every time you launched an application, the screen goes dark and fades back in once the application has opened. You then have to guess what the changes to the display were. In fact this is almost what windows vista does when it queries for user granted privileges.
Designing a UI effect which is consistent with the application model can really improve the usability and flow of your application. But, as my mate Nigel says, if you can’t do a smooth transition (synchronized to the LCD refresh rate), then don’t bother. Get advice from someone who is experienced; don’t rely on UI WOW as garnish, it’s distracting and often embarrassing.
One of my favourite books on design is called the Universal principles of design; it’s a reference book with one key idea per page/two page spread. It has sections on the golden ratio, Fibonacci, symmetry, grid, aesthetic usability, Fitts’s law etc.
3D UIs, “UIWOW” have been around for decades, and so have good design principles.
”.
