World: r3wp

or just construct the pipes and pumps that would use seawater near 
to constantly water it? They transport oil across continets in pipes 
no matter how far the water is, they sould be able to make a constant 
suply in there.

or probably 100 other things that seem better than driving in with 
human controlled firetrucks and flying over with a helicopter pouring 
water on it.

at slight risk of looking like idiot: http://jankom-code.posterous.com/idiots-3-minute-solutions-on-cooling-down-the
Idiot's 3 minute solutions for cooling down the reactors in Japan

There is probably no money in developing equipment for handling nuclear 
accidents, if one only happens every 25 years, as it just adds to 
the cost of building reactors, reducing the financial incentive to 
build them in the first place. Also even though many reactor designs 
that are supposedly better than the current ones exist on paper or 
are in research, there is apparently not enough people working in 
the field of research and licensing to move such reactors into production.


There are reactor types that work entirely with passive cooling and 
can be evacuated for 72 hours before anything happens, but they are 
still at the research stage.


Putting the reactor in a big hole might be a good idea, but it depends 
on the location and how an earth quake would affect the hole.


It seems that many of these accidents are due to very clear design 
flaws or overriding specific safety procedures. That's a positive 
thing, because it means, it's not impossible to build very safe reactors.

The only thing they seem to robotisize is a Dutch company they asked 
to make aerial photos with those remote controlled geek helicopters

I thought those were always Made In China, but apparently the Dutch 
are good at making them seem high tech :-)

just so that we don't start feeling too smug about how things are 
done here in North America wrt nuclear plants:

http://www.counterpunch.org/stclair03162011.html

http://xkcd.com/radiation/

That's good info

http://www.i-programmer.info/news/150-training-a-education/2188-cmu-drops-oop.html

Why not just teach Javascript?

At least they can use it

There are these rare messages where I sometimes think there is some 
sanity left in the world after all

unfortunately, what I call flat OOP (limited inheritance & polymorphism) 
is very effective and functional programming isn't a substitute for 
OOP.  The fact that the page talks about OOP being anti-modular, 
IMHO, clearly shows a fundamental lack of understanding for that 
paradigm.


the problem here is not OOP, its how people have granted it the "golden 
hammer" status that it never should have gotten in the first place. 
 The problem is that people have diluted the core ideas behind OOP 
by bloating it out of its purity.   


When you look at the huge mess that are the current commercial frameworks 
like java or .net, then it does seems like OOP has somehow failed, 
but in reality, going back to basics and teaching how to leverage 
OOP properly would have been a better decision IMHO.

Hmm, a subset of C and ML. Maybe the anti-modular comment refers 
to modularity in the large, e.g. system modularity, which I agree 
with. 


I'm not sure about bloating out purity though Max. Yes, the three 
legs it stands on are easy enough to list as bullet points, but even 
early works (not going back to Simula's era) like Booch's OOAD talk 
about notations and other heavy additions, along with the view that 
we needed OOP to help manage complexity, because software is inherently 
complex.

when I mentionned purity I guess I should have used a more descriptive 
sentence.


I really meant to say, objects, being used as objects.   nowadays, 
OOP (the paradigm) is used for every part of software, even parts 
for which its ill-suited.


OOP is not about the language, its about the logical step after structured 
programming.   grouping things together.


why stop at OOP, they might as well re-introduce the GOTO as a viable 
pattern.  :-)


OOP when its used without all the "advanced" object patterns, is 
incredibly effective... just look at the Amiga OS which was almost 
totally OO in its layout and use while still being coded in C.

I like the Go model to OOP: No inheritance, polymorphism based on 
interfaces, object-based concurrency.

I don't know that I would say *after* structured programming. Grouping 
by binding code and data together is not the only way, though it 
can work well. I think we're on the same page. And I agree with Brian.

http://www.stuff.co.nz/science/4825382/Meteor-explodes-over-North-Island

Missed it!

Max, then there is no problem with teaching C, is there? :-) CMU 
now offers an optional follow-up OOP course where they can teach 
to do that in C. Just like you can teach to avoid goto in C (which 
can still be very useful, including long jumps)

its just plain stupid for a university of this caliber to shrug off 
50 years of CS cause someone high-up doesn't like it.


the grads won't get the training they need for actual real-world 
jobs.

I repeat, they offer an optional OO course

Training for mainstream jobs is the task of technical colleges, not 
of universities

@Graham and vice versa

http://blogs.unity3d.com/2011/02/27/unity-flash-3d-on-the-web/

Max, functional programming IS a substitute for OOP. whether it is 
better or worse is a different matter.

then why would they start adding OO concepts to functional languages?

(ignoring the issue wether that question even makes sense to ask, 
or wether the factual situation implied is indeed the case)


because the concept transcended it's OO roots and turned out to be 
more generally applicable, or even just _also_ applicable in a different 
context?

viz the success of subtyping (in programming languages) and subtype 
polymorphism which arguably rooted in "OO" (simula) have long since 
transcended this paradigm

http://www.google.com/intl/en/jobs/uslocations/mountain-view/autocompleter/index.html


New job offer at Google (yes, today, I'm posting in the wrong group) 
:-)

It's a bit late for April fool's isn't it ?

Not in Europe or the Americas -- still prime time for announcements 
of R3 and such things,

Yes, it is well past mid-day.

just after 9 AM here, so plenty of opportunities left. :-)

Henrik you really must stop living in the past ;-)

well, spin the Earth in the other direction, then. :-)

Henrik .. it's just a convention!

We could start GMT+13 in Scandinavia .. that would be easier than 
changing the earth's rotation

and NZ would be at GMT

except we have no Zulus here

since we are not rid of things like the daylight savings time yet, 
I'm starting to think it's easier to just change the Earth's rotation 
:-)

This is what the Danes think

http://physicsworld.com/cws/article/news/3100

We have constant daylight savings here. That's why the time in Jakarta 
which is two hours flight to the east of us is one hour behind the 
time here. The sun is at  its highest between 1:00 and 1:30 depending 
on the season.


There was a rumour that the Prime Minister who made the change did 
so because he was fed up with the senior civil servants playing golf 
before going to work.

If the poles reverse, we might as well change the timezones so we 
can still use our compasses to guess the time

too bad that swatch time never caught on. perhaps it was simply introduced 
at the wrong time.

My chat program based on Maarten's rubgy used swatch time

All these webinar invites I get .. no idea what the local time is!

then why would they start adding OO concepts to functional languages?

 - to attract more (ignorant) people to the languages. (unless you 
 consider typeclasses an OO concept...)

indeed, some languages use feature labels, rather than actual design.