There's a minimum thought requirement,
particularly for newspaper letters pages.
Actually writing your own letter indicates
that you have, however minimally, thought
about the issues.
Form letters simply encourage knee-jeck
reactions.
The form letter producers want knee-jerks,
of course.
If you actually wrote your own letter, you
would start thinking about evidence, details,
problems and shades of grey.
And before you know it, you're arguing over
what policy is appropriate, thinking as an
individual and removing the illusion of a
united front.
Activists of any stripe just hate
that.
Smalltalk is interpreted, whereas C normally isn't.
Smalltalk uses a JIT.
The development environment automagically
converts between raw instructions/stack frames/contexts
and a object-based representation that
debuggers and other introspective code can
use.
Once inspected/modified it can convert back
to raw form.
To a certain extent, C debuggers have the same
capabilities, since something like gdb can
print you a pretty symbolic stack trace when
you type in "where".
It must be able to decode the stack.
So -- presumably with a few modifications to
the way stack frames are handled and with
additional information, such as local
variable names and types, supplied to the
debugger -- there's no particular
reason why C development environments
can't have the same functionality.
It's just that, traditionally, that's not the
way things have been done.
There is a common class of pointer errors
in C.
There are a number of other tools that
can help with memory leaks and rogue pointers.
I don't see why a sufficiently sophisitcated
debugger can't include that functionality,
particularly if you could get more meaningful
access to instrumentation from the source code.
But, if you're writing application programs
at least, you should probably have a good
think about why you're choosing to write
them in a glorified assembly language intended
for systems programming in the first place.
they don't want to publish the areas of the shipwrecks, but anyone with the money or power to go dig up ships has some ethics in them
This has not been true in the past.
Why should it be true now?
Anyone with a diving suit can start stripping
a wreck of valuables.
Here's an example: the wreck of the Loch Ard,
one of the worst and most historic wrecks
off Australia's coastline.
When it was found, it was blown apart by
drivers who wanted the lead ballast to sell.
Archaeology is great fun for those of a geek-like bent. Digs are always in need of volunteers. It's healthy, outdoor work. It's got lots of intellectual interest. There's lots of opportunity for socialising. There's lots of opportunity for beer -- and you'll need it after working 8 hours in an open field.
Plus, if you have programming or other scientific skills, archaeologists will love you. And you'll love them. Archaeologists reduce patterns of dirt to information and they're very organised. If you're a programmer they're the best "clients" you'll ever have. (I speak from experience here, having worked on various archaeological doings in the UK, France and Australia.)
The graphs show the uptake of Linux in the 90s.
(Fair enough, since it would be a little difficult
to plot usage of Linux in the 80s.)
And, surprise, it shows an upward trend.
However, that was the 90s, when everyone and
their dog was buying PCs by the truckload.
So it would only be expected that Linux use
also grows, simply as a result of more hosts
being available.
The interesting measure is what proportion of
active PCs run Linux.
Oh, and with 3-5 data points, you can't really
tell if the trend is linear or exponential.
From what I remember, when the alt.gourmand
Usenet Cookbook collection was published, the copyright existed on
the collection of recipes, rather than the
recipes themselves.
I gather that this is standard practice in the
cookbook world, where recipes turn up in other
people's cookbooks.
The reasoning behind all this is that
recipes are techniques, which are not
copyrightable.
The rec.food.recipes archive
rec.food.recipes archive contains more
information on this.
It also contains examples of what sort of
restrictions can be placed on the collection.
Since the GFDL is, essentially, a copyright
license, I don't think that the terms of
the GFDL can be applied to the recipes in the
collection.
Anyone can take the recipes and use them in
other works and not be bound by the GFDL.
Which, by the look of it, violates clause 4 of
the GFDL.
Mind you, as far as I'm concerned, that's fine.
Recipes should be freely exchanged and
published.
That's what is allowed now under copyright.
The GFDL seems to be an additonal encumberance,
since the collection could be placed in the
public domain in any case.
1984 has practically nothing to do with
technology.
The only technological innovation in 1984
is the two-way TV screen.
But it's hardly used; it's just there as a symbol
of omnipresent surveillance.
Other than that, practically everything else
is technology common to the 1940s.
Hardly surprising in a book that was originally
titled 1948.
The "technology" in use is a social technology.
Winston Smith is caught by O'Brien recognising
someone who wants to rebel -- and the Party
having a channel for rebellion set up to catch
thoughtcrime.
Smith's neighbour is shopped by his own child
for muttering in his sleep.
Room 101 is there to ensure total surrender to
the Party; you're willing to utterly betray
your own friends and you know it.
The perpetual war, the 2 miniutes hate, the
rewriting of records are all designed to keep
people aligned to a single goal.
Orwell obviously didn't think that Stalinism
was inevitable.
Either that, or he spent a lot of time whistling
in the dark in his other books and essays.
But he did think, in essays such as
The Prevention of Literature or
Politics and
the English Language, that intellectual
liberty and a commitment to the truth were
essential in fighting totalitarianism.
One of the things I find interesting in
1984 is that the Party is more or
less self-enforcing.
The Party members themselves ensure adherence
to the Party.
This particular piece of psychology is hardly
dead and gone and I think that there's very
little evidence to suggest that technology
ameliorates it.
On the contrary, technology, such as discussion
forums, often allows the enforcement of a closed
world-view, since opposing views can be easily
flamed out of conciousness.
(Hell, you can easily come up with analogies
for the Slashdot versions of "thoughtcrime", "the
2 minutes hate" and "Oceania has always been at
war with Eastasia".)
If I've the copyright myself on a code, I don't need the GPL and am not bound to it.
That would be my reasoning, as well.
However, the FSF's
comments
on X-Windows suggest that they have something
stronger in mind. Unless it's just FUD on the
FSF's part
Everything clear? The GPL does not require everyhing to be always open, that's FUD, often spread by people who don't really understand the license in detail.
It doesn't require everything.
But there's a lot of stuff shy of everything
that's still enough to cause disquiet.
An obvious example is clause 2a of the
GPL, which implicitly covers things like
plugins or kernel modules (or so the
FAQ
says, anyway.)
Another example is the making of the
readline library GPL, rather than LGPL,
and the reasons
for it. Use of the library creates an interesting
tail-wagging-the-dog effect, as far a licensing
is concerned.
Wrong, the code _you_ did soly for yourself you can do with whatever you want. The GPL only allows some certain rights for others to you use your code. You of course have any rights on your own code you want. The GPL has to respected if you want to take over other peoples code (merged with yours) as you need the extra rights the GPL allows to do this at all.
This doesn't seem to be the attitude of the FSF. See this article
where they criticise X/Open for changing the terms of their license.
The logic behind this appears to be that, once
you release something under the GPL, then
you're bound by it, too.
If software design were as visible as a bridge
or house, we would be hiding our heads in shame.
Good houses often look like ropey software:
they have external wires going everywhere;
they have an extension built to house the cat;
you can get to the dining room by two
completely different routes;
there's a priest hole that's been forgotton
about;
there's a secret passage from the library
to the kitchen;
there's a shortcut through the long
gallery to the blue room.
(Yes, of course I'm talking about an
English country house.
Software is complex and doesn't correspond
to a two-up, two-down.)
A good house got that way to serve
the needs of the people who live in it.
Not to satisfy the desires of the latest
school of architectural thought.
I always remember a TV program on a
house designed by a minimalist
architect.
It looked like utter hell to live in;
very aesthetically pleasing, though.
Tom Wolfe's From Bauhaus to Our House
is a pretty amusing read on what happens when
aesthetic theory gets hold of the steering
wheel.
I like good software aesthetics.
And there's a lot of advantages to having a
good, well structured design that can accomodate
the vagaries of user needs gracefully.
But, ultimately, the aim is to provide something
that is comfortable and natural to a user.
The result is often going to look organic and
messy.
Not that this doesn't fill me with a certain sense of irony. The US courts have been quite happy to extend their jurisdiction in civil cases to whereever it pleases them in the past. So one could say that the precedent has already been set.
Wouldn't this require the cooperation of the
national government of the defendant's country
of origin?
More or less, yes.
There's an
opinion piece
in
The Australian
which mentions that they'll probably have
to apply to a US court to have the damages
awarded.
The US court is unlikely to award damages
unless it meets the US criteria for defamation.
Not that this doesn't fill me with a certain
sense of irony.
The US courts have been quite happy to extend
their jurisdiction in civil cases
to whereever it pleases them in the past.
So one could say that the precedent has already
been set.
Fine Art basically means
"Art produced or intended primarily
for beauty rather than utility."
By extension, the term is often used to
mean forms where only that definition can apply:
painting, sculpture, etc.
(Yeah, yeah, I know.
A painting of Charles II
descending from the heavens appointed by God
has quite a lot of utility.
But you know what I mean; short of burning
it as fuel, a painting doesn't have much
direct use beyond covering up a stain on the
wall.)
This ends up meaning that craft-based arts
such as ceramics, jewelry, printing and, by
extension, computer generated art, are not fine art, even if the objects produced do meet the
beauty over utility requirement.
Because some other areas of these arts are
used to produce things which are useful as
well as beautiful, the mud rubs off and sticks.
Computer generated art is in the same boat.
This is pure snobbery, of course.
But, if you follow the money, then it associates
status with the conspicuous consumption of
useless items.
Which is what humans, particularly rich,
socially connected humans, do.
So the people who say that computer generated art
is not fine art are are sort of correct; it fails
the "utterly useless" test and
is, therefore, socially unacceptable.
This may not mean much in purely objective terms.
But, in patronage terms, it means a hell of a lot.
I took an OO class in smalltalk. It was a waste of my time. I already knew C++ and Java at the time and instead of learning OO concepts, I learned how to program smalltalk, a skill I will never need.
Did they teach:
The use of class extensions in packages/parcels?
The use of blocks (closures) as first order
objects and pieces of functionality?
The consistent use of objects, rather than
using special cases for primitive types and
functionality?
All these things are possible in Java and C++ --
well, not class extensions --
but only in a roundabout and difficult way.
As a result, programmers in these languages
tend to avoid them.
Smalltalk allows a clean expression of these
concepts.
Something, I think, that makes it an excellent
teaching language; you can learn OO programming
rather than hacking around obstacles.
Although it does lead to a certain disgruntlement
when the ex-student realises how far behind
Java and C++ are in OO terms.
Still, one can't have everything.
The features listed as Good Things in Python:
object-orientedness, consistency of approach,
ease of development, dynamic typing and so on.
They all exist in Smalltalk.
Smalltalk has the advantage of considerably
greater maturity in terms of virtual machines
and development tools.
Yet Python seems to be gaining popularity as
Smalltalk wanes.
This isn't a dig at Python; something good is
something good, no matter where it turns up.
But what is it about
Python which makes it popular, while Smalltalk
never seems to quite make it?
The first and most obvious thing is Smalltalk's
apalling marketing decisions.
Run-time licenses, expensive development
licenses and costly support obviously went
some way to raising the bar, although
there are now good, free versions of Smalltalk
available for non-commercial work.
But ParcPlace not taking up Sun's offer to
bundle VisualWorks with SunOS.
Ow!
I don't know what Python's memory footprint
is, but Smalltalk's can grow quite large.
With Java, perl and Python around, resource
hunger no longer seems to be a crippling issue,
though.
Well, HAL, with his single eye,
knocking off the crew one by one, makes a
pretty good
Polyphemus.
The problem is that anything where you've
got a long voyage with people being killed looks
like the the Odessey.
Greg Egan's
The Plank Dive
has an offhand comment about these sort of
myths being strange attractors for
pre-literate stories.
Everything created by the Feds exists in the public domain, yes; but public domain works can be used in GPLed programs.
But the reverse is not true; a GPL'd program
is not in the public domain.
Placing something in the public domain means that
it's there for people to use as they see fit,
including modifying it and charging for the
copyrighted, modified version -- the original
is still in the public domain, of course.
The GPL explicitly prevents this.
The viral nature of the GPL means that
modifications to GPL'd programs as also GPL'd.
(The intentions of the GPLs creators seem to
be that this even applies to linkable modules.
Although Linux has a permission,
courtesy of LT, for things like device drivers.
See
here
for comments on this.)
So, if the government wants to create a modified
version of a GPLd program, for the public good,
there's a potential conflict.
This is a good example of where the use of
free in the FSF's definition of
free software becomes rather
questionable; it's not so free as ye olde
traditional public domain.
The GPL provides a benefit in terms of preventing
incompatibilities, drift and hidden extensions.
Something which is a good thing.
It also prevents free use of publicly available
information, which is not so good and something
public institutions need to keep a wary eye on.
Or would they use reverse logic and give a list
of allowed sites, with the person having to
petition for each specific site.
This would destroy any real use of internet.
From a Chinese officials POV, this approach
would likely be preferable.
Once you have this, you can have official
censors going through sites, like a sort of
human Google, deciding what is viewable by the
fair eyes of the great unwashed.
After a while, entire sites would be rated
as "trusted", reducing work somewhat.
Think of it as firewalling; that which is
not explicitly allowed is prohibited.
"any real use of the Internet"
contains a raft of cultural assumptions[*]
From the Chinese government's POV, what
they're doing is no different to a company
installing filtering software to ensure
that employees only use the Internet for
approved purposes.
Just with a rather wider scope.
And you can never go home at the end of the day.
[*] Assumptions that I'm pretty happy
with, incidentally.
Re:Most ignorant comment in the history of mankind
on
Fission in a Box
·
· Score: 1
For all we know, an alien species had a base here.
Well, I hope they got a visit from their
regulatory authorities.
No proper waste handling, just leaving it
in the ground with water running over it.
Given the shape of the reactors, no containment
buildings.
And all that unused uranium lying about the
place when they were finished.
And water moderated, too; you'ld have thought
that they sailed through space in caravels.
Shocking. Simply shocking.
Re:Most ignorant comment in the history of mankind
on
Fission in a Box
·
· Score: 1
The waste from nuclear reactors and reprocessing is in no way comparable to what is present in nature.
You might want to have a look at the
Oklo Fossil Nuclear Reactor,
the product of a natural nuclear reactor
in Gabon.
It's thought that, about 2000 million years
ago, a water moderated chain reaction started
in uranium rich soil and ran for about 1 million
years.
(How do they know, well there's all these fission
products lying about the place.)
...but it will still be worrisome to deal with for something like 100,000 years.
This 100,000 years number seems to pop up with
wild abandon.
IIRC, this figure appears to be the half-life of
one of the daughter products of the decay chain.
But with a longer half-life comes less
activity.
How close is this to background radiation?
And does it make more difference to radiation
exposure than,
for example, living at 700m above sea-level
does?
None of the above is an argument for not
treating nuclear waste carefully.
But requiring massively different standards
of risk control when the nuclear word is used
doesn't help anyone.
And it may be actively harmful: Opposition
to nuclear power has led to more coal being
used for base-line power production, leading to
massive amounts of chemical and radioactive
pollutants being spewed into the atmosphere.
Not something I feel particularly comfortable
about.
My personal hunch is that if you conducted a random poll of developers with *significant* experience with both languages (say, a minimum of 1 year full-time experience with each), probably 90-98% would agree with this.
I'll put my hand up: 5 years Smalltalk,
2 years Java.
I think Smalltalk is a much more pleasant
language to program in: closures (blocks),
the IDE (especially senders and implementers),
no casts and numerous other features.
(Although
Brewmaster
provides some of the nicer IDE aspects.
Smalltalk also has software engineering features
that I would love to see in Java, particularly
open classes.
Open classes prevent the hell of library classes
not anticipating your every need; it's one of the
nicest things about Smalltalk and every other
OO language seems to have utterly missed the
point.
(MultiJava has these).
What Java does have is standards, such as
EJB, JFC, JMS and so on.
They may not be particularly good or coherent
standards.
But they do encourage component development
and interoperability.
Smalltalk is good at the core language level,
but differences in GUI building and other
libraries between vendors can make Java a
more attractive option if you want to avoid
legacy lock.
Slashdot Mirror
Actually writing your own letter indicates that you have, however minimally, thought about the issues. Form letters simply encourage knee-jeck reactions.
The form letter producers want knee-jerks, of course. If you actually wrote your own letter, you would start thinking about evidence, details, problems and shades of grey. And before you know it, you're arguing over what policy is appropriate, thinking as an individual and removing the illusion of a united front. Activists of any stripe just hate that.
Smalltalk uses a JIT. The development environment automagically converts between raw instructions/stack frames/contexts and a object-based representation that debuggers and other introspective code can use. Once inspected/modified it can convert back to raw form.
To a certain extent, C debuggers have the same capabilities, since something like gdb can print you a pretty symbolic stack trace when you type in "where". It must be able to decode the stack. So -- presumably with a few modifications to the way stack frames are handled and with additional information, such as local variable names and types, supplied to the debugger -- there's no particular reason why C development environments can't have the same functionality. It's just that, traditionally, that's not the way things have been done.
There is a common class of pointer errors in C. There are a number of other tools that can help with memory leaks and rogue pointers. I don't see why a sufficiently sophisitcated debugger can't include that functionality, particularly if you could get more meaningful access to instrumentation from the source code. But, if you're writing application programs at least, you should probably have a good think about why you're choosing to write them in a glorified assembly language intended for systems programming in the first place.
This has not been true in the past. Why should it be true now? Anyone with a diving suit can start stripping a wreck of valuables.
Here's an example: the wreck of the Loch Ard, one of the worst and most historic wrecks off Australia's coastline. When it was found, it was blown apart by drivers who wanted the lead ballast to sell.
Plus, if you have programming or other scientific skills, archaeologists will love you. And you'll love them. Archaeologists reduce patterns of dirt to information and they're very organised. If you're a programmer they're the best "clients" you'll ever have. (I speak from experience here, having worked on various archaeological doings in the UK, France and Australia.)
The archaeological fieldwork server at http://www.cincpac.com/afos/testpit.html lists digs looking for volunteers.
Oh, and with 3-5 data points, you can't really tell if the trend is linear or exponential.
A pity he's gone.
The rec.food.recipes archive rec.food.recipes archive contains more information on this. It also contains examples of what sort of restrictions can be placed on the collection.
Since the GFDL is, essentially, a copyright license, I don't think that the terms of the GFDL can be applied to the recipes in the collection. Anyone can take the recipes and use them in other works and not be bound by the GFDL. Which, by the look of it, violates clause 4 of the GFDL.
Mind you, as far as I'm concerned, that's fine. Recipes should be freely exchanged and published. That's what is allowed now under copyright. The GFDL seems to be an additonal encumberance, since the collection could be placed in the public domain in any case.
1984 has practically nothing to do with technology. The only technological innovation in 1984 is the two-way TV screen. But it's hardly used; it's just there as a symbol of omnipresent surveillance. Other than that, practically everything else is technology common to the 1940s. Hardly surprising in a book that was originally titled 1948.
The "technology" in use is a social technology. Winston Smith is caught by O'Brien recognising someone who wants to rebel -- and the Party having a channel for rebellion set up to catch thoughtcrime. Smith's neighbour is shopped by his own child for muttering in his sleep. Room 101 is there to ensure total surrender to the Party; you're willing to utterly betray your own friends and you know it. The perpetual war, the 2 miniutes hate, the rewriting of records are all designed to keep people aligned to a single goal.
Orwell obviously didn't think that Stalinism was inevitable. Either that, or he spent a lot of time whistling in the dark in his other books and essays. But he did think, in essays such as The Prevention of Literature or Politics and the English Language , that intellectual liberty and a commitment to the truth were essential in fighting totalitarianism.
One of the things I find interesting in 1984 is that the Party is more or less self-enforcing. The Party members themselves ensure adherence to the Party. This particular piece of psychology is hardly dead and gone and I think that there's very little evidence to suggest that technology ameliorates it. On the contrary, technology, such as discussion forums, often allows the enforcement of a closed world-view, since opposing views can be easily flamed out of conciousness. (Hell, you can easily come up with analogies for the Slashdot versions of "thoughtcrime", "the 2 minutes hate" and "Oceania has always been at war with Eastasia".)
An obvious example is clause 2a of the GPL, which implicitly covers things like plugins or kernel modules (or so the FAQ says, anyway.)
Another example is the making of the readline library GPL, rather than LGPL, and the reasons for it. Use of the library creates an interesting tail-wagging-the-dog effect, as far a licensing is concerned.
It is worth examining the license in detail ...
They're worth watching to see what the pitfalls are likely to be.
Good houses often look like ropey software:
(Yes, of course I'm talking about an English country house. Software is complex and doesn't correspond to a two-up, two-down.)
A good house got that way to serve the needs of the people who live in it. Not to satisfy the desires of the latest school of architectural thought. I always remember a TV program on a house designed by a minimalist architect. It looked like utter hell to live in; very aesthetically pleasing, though. Tom Wolfe's From Bauhaus to Our House is a pretty amusing read on what happens when aesthetic theory gets hold of the steering wheel.
I like good software aesthetics. And there's a lot of advantages to having a good, well structured design that can accomodate the vagaries of user needs gracefully. But, ultimately, the aim is to provide something that is comfortable and natural to a user. The result is often going to look organic and messy.
For example, the Absolut Beachwear case.
More or less, yes. There's an opinion piece in The Australian which mentions that they'll probably have to apply to a US court to have the damages awarded. The US court is unlikely to award damages unless it meets the US criteria for defamation.
Not that this doesn't fill me with a certain sense of irony. The US courts have been quite happy to extend their jurisdiction in civil cases to whereever it pleases them in the past. So one could say that the precedent has already been set.
By extension, the term is often used to mean forms where only that definition can apply: painting, sculpture, etc. (Yeah, yeah, I know. A painting of Charles II descending from the heavens appointed by God has quite a lot of utility. But you know what I mean; short of burning it as fuel, a painting doesn't have much direct use beyond covering up a stain on the wall.)
This ends up meaning that craft-based arts such as ceramics, jewelry, printing and, by extension, computer generated art, are not fine art, even if the objects produced do meet the beauty over utility requirement. Because some other areas of these arts are used to produce things which are useful as well as beautiful, the mud rubs off and sticks. Computer generated art is in the same boat.
This is pure snobbery, of course. But, if you follow the money, then it associates status with the conspicuous consumption of useless items. Which is what humans, particularly rich, socially connected humans, do. So the people who say that computer generated art is not fine art are are sort of correct; it fails the "utterly useless" test and is, therefore, socially unacceptable.
This may not mean much in purely objective terms. But, in patronage terms, it means a hell of a lot.
There's a similar article called Why Microsoft is Wary of Open Source by Joe Wilcox and Stephen Shankland on CNET.
Did they teach:
All these things are possible in Java and C++ -- well, not class extensions -- but only in a roundabout and difficult way. As a result, programmers in these languages tend to avoid them. Smalltalk allows a clean expression of these concepts. Something, I think, that makes it an excellent teaching language; you can learn OO programming rather than hacking around obstacles.
Although it does lead to a certain disgruntlement when the ex-student realises how far behind Java and C++ are in OO terms. Still, one can't have everything.
This isn't a dig at Python; something good is something good, no matter where it turns up. But what is it about Python which makes it popular, while Smalltalk never seems to quite make it?
But ParcPlace not taking up Sun's offer to bundle VisualWorks with SunOS. Ow!
There's got to be more to it than that, though.
The problem is that anything where you've got a long voyage with people being killed looks like the the Odessey. Greg Egan's The Plank Dive has an offhand comment about these sort of myths being strange attractors for pre-literate stories.
But the reverse is not true; a GPL'd program is not in the public domain. Placing something in the public domain means that it's there for people to use as they see fit, including modifying it and charging for the copyrighted, modified version -- the original is still in the public domain, of course.
The GPL explicitly prevents this. The viral nature of the GPL means that modifications to GPL'd programs as also GPL'd. (The intentions of the GPLs creators seem to be that this even applies to linkable modules. Although Linux has a permission, courtesy of LT, for things like device drivers. See here for comments on this.)
So, if the government wants to create a modified version of a GPLd program, for the public good, there's a potential conflict.
This is a good example of where the use of free in the FSF's definition of free software becomes rather questionable; it's not so free as ye olde traditional public domain. The GPL provides a benefit in terms of preventing incompatibilities, drift and hidden extensions. Something which is a good thing. It also prevents free use of publicly available information, which is not so good and something public institutions need to keep a wary eye on.
From a Chinese officials POV, this approach would likely be preferable. Once you have this, you can have official censors going through sites, like a sort of human Google, deciding what is viewable by the fair eyes of the great unwashed. After a while, entire sites would be rated as "trusted", reducing work somewhat. Think of it as firewalling; that which is not explicitly allowed is prohibited.
"any real use of the Internet" contains a raft of cultural assumptions[*] From the Chinese government's POV, what they're doing is no different to a company installing filtering software to ensure that employees only use the Internet for approved purposes. Just with a rather wider scope. And you can never go home at the end of the day.
[*] Assumptions that I'm pretty happy with, incidentally.
Well, I hope they got a visit from their regulatory authorities. No proper waste handling, just leaving it in the ground with water running over it. Given the shape of the reactors, no containment buildings. And all that unused uranium lying about the place when they were finished. And water moderated, too; you'ld have thought that they sailed through space in caravels.
Shocking. Simply shocking.
You might want to have a look at the Oklo Fossil Nuclear Reactor, the product of a natural nuclear reactor in Gabon. It's thought that, about 2000 million years ago, a water moderated chain reaction started in uranium rich soil and ran for about 1 million years. (How do they know, well there's all these fission products lying about the place.)
Agreed, but it is less chemically toxic than many other natural toxins. Botulism toxin is, I believe, top of the league table here. (And people inject this stuff into their foreheads!)
This 100,000 years number seems to pop up with wild abandon. IIRC, this figure appears to be the half-life of one of the daughter products of the decay chain. But with a longer half-life comes less activity. How close is this to background radiation? And does it make more difference to radiation exposure than, for example, living at 700m above sea-level does?
None of the above is an argument for not treating nuclear waste carefully. But requiring massively different standards of risk control when the nuclear word is used doesn't help anyone. And it may be actively harmful: Opposition to nuclear power has led to more coal being used for base-line power production, leading to massive amounts of chemical and radioactive pollutants being spewed into the atmosphere. Not something I feel particularly comfortable about.
I'll put my hand up: 5 years Smalltalk, 2 years Java.
I think Smalltalk is a much more pleasant language to program in: closures (blocks), the IDE (especially senders and implementers), no casts and numerous other features. (Although Brewmaster provides some of the nicer IDE aspects.
Smalltalk also has software engineering features that I would love to see in Java, particularly open classes. Open classes prevent the hell of library classes not anticipating your every need; it's one of the nicest things about Smalltalk and every other OO language seems to have utterly missed the point. (MultiJava has these).
What Java does have is standards, such as EJB, JFC, JMS and so on. They may not be particularly good or coherent standards. But they do encourage component development and interoperability. Smalltalk is good at the core language level, but differences in GUI building and other libraries between vendors can make Java a more attractive option if you want to avoid legacy lock.