Calling them all "planets" would be like starting to call all schools of whatever level "university" - a status grab that would ultimately achieve nothing but a devaluation of the more prestigious term, and a muddling of the underlying factual distinctions.
Good thing the British Govt isn't responsible for designations of astronomical objects, then.
It does make me wonder though - how many legitimate e-mails contain an attached zip and an attached image? I'm surprised that combination doesn't trigger an auto-delete.
An easy language is not necessarily expressive enough.
An excellent point - wish I had some mod points.
It does, however, raise the issue of qualifying and quantifying expressivity. I think a strong type system is a definite plus - if you look at a language like SML, the type of a function usually tells you what it does without any further documentation. Language support for modularity, to allow top-down design which maps cleanly to a top-down implementation is also good - OO and a true hierarchical package system (rather than e.g. Java's where the apparent hierarchy is only for logical structure) do this well. Finally, I agree with you that good libraries are important - the more in the libraries and the less in the language, the better, unless taken to the extremes of Turing Tarpit languages.
I think you're rather harsh on Java - I've been working with it in industry for nearly two years, and I find it a fairly elegant language (with excellent libraries). I'm looking forward to the improved type system of Java 1.5, which will be almost as expressive as the basic SML type system (i.e. without modules).
Functional PLs, which you mention, are indeed elegant for algorithm development (provided the language is well typed) or fairly mathematical work, but the functional paradigm isn't really suited to I/O, which necessarily implies functions with side-effects. (And, as a side-note, Logo does include list processing commands - it's not at the level of SML or LISP, but it's probably more expressive than you think).
In summary, I think there are languages available at the moment which good programmers will find expressive enough but which won't burn bad programmers too badly - I'd claim Java as an example, and possibly OCAML, although I haven't really delved into that. Maintaining code by someone who writes either without understanding the paradigm will be a major headache (I'm currently trying to tidy some imperative code written in Java with as little use of objects as possible - ugh), but I doubt anyone will ever write a language which forces code to be written in a maintainable way.
More and more computer science departments are teaching using Java. Why? because it helps people to understand how the computer works? no. Simply, because it's what the industry is using.
Why should CS departments use low-level languages to teach programming? The important concepts to get across in a programming course are algorithmic complexity and basic software engineering - writing for maintainability. The only programmers who really need to know how hardware works are those writing kernels and drivers.
There seems to be a tension in your argument. You complain, it seems, about
Programming languages which are meant for the lowest common denominator. Those languages are used which are not the best ones, but for which you can find enough adequately skilled work force. Mediocrity begets mediocrity.
but your main proposal seems to be
If you want programming to not-stink, MAKE IT EASY FOR THE PROGRAMMERS. It really is that simple.
Is the conclusion we're meant to draw that the easier a language is to use, the harder it is for good programmers to use?
Re:Expensive Electronics Cheap Scams, not taken do
on
eBay Fraud Vigilantes
·
· Score: 4, Funny
I can't remember when I ate anything that was as good as the ads.
Nothing beats an IBM full-page magazine ad with mustard.
My technique for that is simple: use citeseer to get the names of the authors of a few relevant papers, and then use Google to find their home pages. A rather high proportion of academics have their papers for download, even if the journal which published them wants you to subscribe before you read more than the abstract.
The biology teachers at my dad's school can't read the syllabus online because the school's filter blocks it. Whatever measures you take, filters will be imperfect.
At what point do we say the security we have is good enough?
Indeed, that's the big question. The problem is that the first stage in designing security is writing the threat model - and "People attempting to kill at random" is about as specific as we can get. Even looking at air travel, the threat model is unclear - it used to be straight hijacking for hostages, but now the aim could be to take control of the plane and use it as a bomb, or simply to destroy it in order to further disrupt air travel.
Do you really think two or three guys armed with boxcutters are going to take down 30-40 passengers (not to mention the sky marshal)?
Depends how good their intel is. If they board the flight knowing who the marshal is and where he's sitting, the first thing they do is stab him and take his gun. If there are no firearms on board, it comes down to hand-to-hand combat with improvised clubs and blades. Bearing in mind that planes are quite narrow, a team of four who've trained and planned the weapons they'll improvise certainly have a fighting chance against 30-40 passengers.
Locking doors is no good. What you need is time-locked doors, or even better doors which can only be opened after a certain length of time, in a certain geographical area, from outside. That way terrorists can't force the pilots to open them by killing the passengers one by one.
Since terrorism (in the 9-11 sense) is a response to U.S. tyranny elsewhere, ending the tyranny would end the terrorism.
What do you mean by "(in the 9-11 sense)"? On a scale that kills 50+ people? Perpetuated by Muslims against the US and her allies? Perpetuated by al Qaeda?
Even were the US to pull out of every other country, and return to the isolationist state of the 30s, it would still be possible for blame to be cast at it for a decade or two. Moreover, whereever a country was governed by current allies of the US (e.g. while Pakistan remains under Masharraf) the recruiters can spin conspiracy theories. In think you're optimistic in thinking that the hatred for the US will go away just because the US is acting reasonably.
No, the IRA was "Kill all British/Protestants wherever you find them."
"Except in the US because we don't want our funding to dry up." I'm guessing that grandparent poster is from the US (partly because they refer to I-95, and partly because they don't seem to have heard of the bombings by ETA and FARC in the past year).
Slashdot Mirror
Copyright protects the expression of an idea.
You must be the only person who thinks that the appeals process will take less than a couple of years.
"As the Gates Cambridge Scholarship program enters its third year"... - written this year, because it refers to Cambridge's "new Vice Chancellor, Professor Alison Richard".
Three years is a long time?
You missed it
Without wishing to be flamebait, "historical inevitability" immediately puts me in mind of Marxism.
I simply route everything with HTML content straight to the bit bucket.
It does make me wonder though - how many legitimate e-mails contain an attached zip and an attached image? I'm surprised that combination doesn't trigger an auto-delete.
It does, however, raise the issue of qualifying and quantifying expressivity. I think a strong type system is a definite plus - if you look at a language like SML, the type of a function usually tells you what it does without any further documentation. Language support for modularity, to allow top-down design which maps cleanly to a top-down implementation is also good - OO and a true hierarchical package system (rather than e.g. Java's where the apparent hierarchy is only for logical structure) do this well. Finally, I agree with you that good libraries are important - the more in the libraries and the less in the language, the better, unless taken to the extremes of Turing Tarpit languages.
I think you're rather harsh on Java - I've been working with it in industry for nearly two years, and I find it a fairly elegant language (with excellent libraries). I'm looking forward to the improved type system of Java 1.5, which will be almost as expressive as the basic SML type system (i.e. without modules).
Functional PLs, which you mention, are indeed elegant for algorithm development (provided the language is well typed) or fairly mathematical work, but the functional paradigm isn't really suited to I/O, which necessarily implies functions with side-effects. (And, as a side-note, Logo does include list processing commands - it's not at the level of SML or LISP, but it's probably more expressive than you think).
In summary, I think there are languages available at the moment which good programmers will find expressive enough but which won't burn bad programmers too badly - I'd claim Java as an example, and possibly OCAML, although I haven't really delved into that. Maintaining code by someone who writes either without understanding the paradigm will be a major headache (I'm currently trying to tidy some imperative code written in Java with as little use of objects as possible - ugh), but I doubt anyone will ever write a language which forces code to be written in a maintainable way.
Courtesy of Google News, no subscription required.
My technique for that is simple: use citeseer to get the names of the authors of a few relevant papers, and then use Google to find their home pages. A rather high proportion of academics have their papers for download, even if the journal which published them wants you to subscribe before you read more than the abstract.
The biology teachers at my dad's school can't read the syllabus online because the school's filter blocks it. Whatever measures you take, filters will be imperfect.
Is it just that British eccentrics take a different form?
Surely those copy-protection pseudo-CDs work in more CD players than any of the vinyl records you've bought?
Goes to prove you should never wash your hands. What, why's everyone looking at me?
Locking doors is no good. What you need is time-locked doors, or even better doors which can only be opened after a certain length of time, in a certain geographical area, from outside. That way terrorists can't force the pilots to open them by killing the passengers one by one.
Even were the US to pull out of every other country, and return to the isolationist state of the 30s, it would still be possible for blame to be cast at it for a decade or two. Moreover, whereever a country was governed by current allies of the US (e.g. while Pakistan remains under Masharraf) the recruiters can spin conspiracy theories. In think you're optimistic in thinking that the hatred for the US will go away just because the US is acting reasonably.
In fact, a couple of years ago the UK Robot Wars banned flippers. Dunno whether the ban's still in place.