[Kde-games-devel] The iPhenomenon

[Josef Spillner]

Hey Ian,

some (probably too) long-ish Saturday morning thoughts from my place about 
your concerns. I partially agree, but I try to give some constructive ways 
towards improving the situation.

Am Samstag, 14. August 2010, 04:04:26 schrieb Ian Wadham:
> Lisa particularly likes that we can get localized apps, such as one for
> our Australian national broadcaster, the ABC.  It takes her direct into
> program lists, with hardly any "navigation" effort, where she can easily
> view and catch up on missed episodes, etc.

Desktop customisation aspects especially regarding the user context such as 
location are indeed waiting for more innovation. For example, when I roam with 
my laptop to another city on another continent, I want to specify my new 
location once (unless that information can be retrieved automatically anyway) 
and let all relevant desktop parts adapt to this context change. This includes 
the timezone setting, the weather plasmoid, recommended radio tunes and TV 
streams, smart bookmarks to local news websites, nearby people in the 
addressbook for chat and mail. Another context is language/culture - there's 
bug #245162 for one, but I think the issue deeply extends into supporting 
multiple personalities or pseudonyms such as separate private and business use 
of the desktop which is a necessity from a privacy point of view. Inspiration 
for that could be found for example in MozPET for privacy-enhanced Mozilla. As 
opposed to, say, Apple products which in this area remain stuck in the 90s :-)
So KDE could be a leader of context-aware desktop computing.

> You know me better than that, Aaron ... :-)  Actually, I am beginning to
> think of KDE as a walled garden.  As a programmer, I am locked into it. 
> As a user, Lisa feels locked out.

Admittedly there are lock-in aspects of KDE. For example, there are old Kivio 
files on my disk which I cannot easily edit anymore. Workaround such as 
installing old KDE versions in a virtual machine is not an option for users. 
The new kdelibs proposal for invoking a browser window with an information 
page for unknown filetypes could at least console the user, if somebody keeps 
track of old filetypes.
KDE also defaults to some specific online service offerings, often without the 
possibility to switch to alternative implementations of the same API (see bug 
#188952 and its many dupes) and without giving users guarantees about online 
data portability.

This means that the lock-in factor of KDE is definitely not zero, but overall I 
would say it's small on an absolute scale and very small relative to other 
platforms. What we lack and might increase the factor in the future is a 
coherent strategy to avoid lock-ins, e.g., through a policy. We have strange 
C++ ABI policies, which honour C++ implementation weaknesses of former times, 
but we don't have online integration policies which would touch on discussion 
points even present in mass media nowadays such as "Google vs. privacy", a 
topic covered at least weekly in newspapers and evening news. If we don't have 
a cool answer to that problem, journalists won't come to us to report about it 
and KDE remains unknown to the majority of the population.

> I often think of what R W Hamming said in his Turing Prize acceptance
> speech. Quoting Isaac Newton, who said that he could not have seen so far
> if he had not stood on the shoulders of giants, Hamming said, "In the
> computer world we stand on each others' feet".

Hamming is right on that one, I fear.

Zooming out of the relatively small KDE community, I see myself interacting 
with computer science and engineering communities (and KDE is one of them) and 
with amateur linguists communities. There are many connections between those 
two sets, many great computer scientists were and are linguists and many CS 
research topics are driven by natural language problems. Still, on the 
surface, the only connection seems to be (from my experience) the constant 
swearing of linguists about improper functionality of spellcheckers, printers 
and so on. My role as "computer guy" is supposedly to help out, but I can't, 
because KDE is also not optimised for linguists as users. Regarding your 
comment on Akademy 2008 and the target groups, it is clear that some features 
will always be more important to one group of users compared to others. KDE is 
certainly not the platform of choice for vast hordes of gamers either, but 
IMHO we have several features in this area which others don't have, so calling 
for despair is akin to devaluation of the potential.

> Sure, I would like to play around with touch-screens, but not at the cost
> of having to re-learn language, libraries, etc. and find that they are not
> all that different from C++, Qt and KDE, but just enough different to
> cause serious misunderstandings, bugs and heartaches.  Been there, done
> that!

Think about iPad-equivalent devices with large touchscreen but more choice in 
software, including a full Qt 4.7 stack. They do exist, like the WeTab. I'm 
currently trying to get such a device for some KDE integration work in a 
research project. Not having to relearn toolkits is certainly a boon.

> So guys, have a look at this Apple stuff and have a think about it.  Let's
> stop fixating on technical details so much and think about where we are
> going in the next few years.

I agree on complementing the technical view with a broader strategy of where 
KDE, and naturally KDE Games, is going to stand in two to five years. We don't 
have a clear roadmap of what goals to achieve across releases. "He/She who 
codes, decides" is still a suitable mantra within KDE, but it is not without 
problems and one of them is the apparent lack of long-term coordination.

What I don't agree on is having to look at Apple stuff for this to happen. We 
have our own brains to think about solutions, balls to counter resistance and 
butts to move so that change happens.

Josef