Become a fan of Slashdot on Facebook


Forgot your password?
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. ×

Comment Re:The kids love it (Score 1) 40

I was pretty sceptical about the Micro bit when it was announced. It seemed under-specced compared to all the other small board computers out there.

I was wrong. Talking to people who actually use them to teach, they (teachers and kids) love them. The combination of the on-board screen, accelerometer etc, and the toolchain all combine to make it really quick to get going and build something simple but fun. Like the old 8-bit micros, you can get going almost instantly.

Obviously they have limits, and something like an Arduino or Pi is going to be more use for bigger projects, but for teaching kids the basics and letting them have fun with them the Bit is great. These are designed as a first step, and that's what they are good at.

Comment Re:Well... (Score 4, Insightful) 532

Likewise. I'm hopefully sceptical :)

At worst there is an interesting effect going on that is worth further study and might provide some new insights into some aspects of physics, or simply improvements to experimental techniques. At best it has the possibility to revolutionise some aspects of space exploration.

I am sceptical that this will live up to the best case, but I really hope that my scepticism turns out to be wrong.

This is what science is all about. There's an odd effect, people are doing experiments, whatever happens we will have learnt something which may one day be useful. This is an extraordinary claim, it requires extraordinary proof, which we will hopefully get.

Comment A few obvious corrections (Score 1) 53

First, DES is 56 bit (near enough 60). Triple DES as per first mode (the authorised standard) is 168 bits. The article fails to distinguish, implying the authors are just a little bit naff. 3DES seems to be quite safe, as long as not used in DES emulation mode. And who the hell emulates a mode that was broken in the 80s?

Second, Blowfish was replaced by TwoFish, ThreeFish and Speck. Skein, an entrant to the DES3 challenge, makes use of ThreeFish.

Third, the Wikipedia page states it has been known for a long time that weak keys are bad. This particular attack, though, is a birthday attack. You can find all the ciphers vulnerable or free that you should be using. Anything not on the list is something you are solely responsible for.

In other words, this information is about as useful as telling up that Model T Fords weren't good at cornering at highway speeds. Below are some links, I can't be buggered to HTML-ify them.

I do not trust most encryption software these days, but that's because programmers these days are sloppy and arrogant.

Comment Re:An odd choice (Score 1) 14

In that case I would be asking what what Apple wants to do with distributed graph analytics because that was probably Turi's most interesting/unique product and expertise. They have a great library for handling extremely large graphs distributed over many nodes, and a lot of expertise in exactly how to do that really well.

Comment An odd choice (Score 4, Interesting) 14

I have to admit to being a little unclear as to Apple's plans here. I'm somewhat familiar with Turi's product offerings (at least, I was back when they were called Dato). It's more of pure data analytics tool than anything, and personally I found the underlying python libraries which are open source far more compelling than the point and click predictive analytics and charting GUIs which seemed to be their main product. And even on that front I would put more stock in scikit-learn, pandas, dask and the many open source deep learning libraries (mostly built on theano and tensor flow) if I really wanted to do machine learning and distributed machine learning.

Now don't get me wrong, Turi has some nice products, but they tend to be standalone suites designed to let front-line analysts have a nice GUI interface to basic machine learning tools, not "push the envelope AI". I really can't see what Apple would do with it beyond build up a business analytics suite to compete with Tableau and Azure ML. Anyone have any better ideas?

Comment This is for civil small claims (Score 2) 103

This is for civil small claims cases, in which lawyers are rarely involved and which are largely set up to support people litigating in person.

They tend to be more about arbitration of unpaid invoices or failure to provide a service that's been paid for etc. I have a couple of friends who have used the small claims courts either against non-paying customers or companies that have stiffed them. In all cases they attended in person and were supported by the court staff rather than lawyers, and they all had good things to say about the staff and the system in principal.

These are very much not cases where high paid lawyers square up against each other and slog it out in a dramatic battle of rhetoric. In fact I've heard from a number of people that the judges who preside tend to take a dim view of trained lawyers trying to steamroller or confuse non-lawyers on the other side. These are not cases involving complex points of law. If the case gets more complex then it may be referred to a higher court.

This proposal makes a lot of sense to me for those sorts of cases. While the cost of using the small claims court can put people off using it, the time and disruption, especially if they are running a business, can be more of an impediment. The ability to handle much of the case without having to attend in person would make the whole system much better, and if it reduced the costs it would make the small claims court more accessible to many people to seek redress from companies. There's also lots of potential to design the online system in such a way as to provide lots of help and advice to non-legal people to they can make their case batter, which should also make the whole process more effective and fair.

Comment It's an interesting idea, but this line is bunk (Score 4, Insightful) 298

This is an intersting idea, and it would be fun to see it developed further, but this line really stuck out.

"Shall we continue to get killed because it is easier to produce aircraft with a design from 1950s?"

Bullshit. I'm not saying some improvement in air crash survivability isn't a good thing, but the idea that people are regularly dying because their aeroplane can't disassemble in midair and parachute them to the ground it frankly offensive to all the engineers who have worked over the years to make large scale commercial flying unbelievably safe.

Total number of air craft fatalities worldwide in commercial flight has been significantly less than 1000 per year for the last couple of decades. Something like 3.6 billion passenger journeys will be completed in 2016 (IATA estimate).

Safety is the single worst reason to throw away a tried and tested basic design that is fantastically safe and replace it with a much more complicated and new system.

Comment Re:Bad reporting. (Score 3, Informative) 323

The $13.25 is a calculation by Uber, not by the journalist. The journalists re-ran some of the calculations and got slightly lower figure.

"Internal Uber calculations, provided to BuzzFeed News by Uber, based on data spanning more than a million rides and covering thousands of drivers in three major U.S. markets — Denver, Detroit, and Houston — suggest that drivers in each of the three markets overall earned less than an average of $13.25 an hour after expenses."

Assuming Uber are not lying about the $13.25, that would still mean that if you worked 40 hours a week, every week of the year, you'd make $27,560 a year.

Whether this is a good or bad depends on how much ordinary taxi drivers make for a similar amount of work.

Uber has frequently talked about how much a driver's gross income will be as a way of encouraging people to join up, which is a nice bit of marketing, and standard practice for companies like Uber.

Comment The question was settled in the 1970s. Just don't (Score 2) 103

Why must you record my phone calls?
Are you planning a bootleg LP?
Said you've been threatened by gangsters
Now it's you that's threatening me
Can't fight corruption with con tricks
They use the law to commit crime
And I dread, dread to think what the future will bring
When we're living in gangster time

Comment Re:Why has perl6 flopped? (Score 1) 281

Strange, I could have sworn that I replied to this with a very detailed and lengthy response... urg.

Anyway, upshot is this: Perl 6 hasn't yet had a chance to flop. It was released in beta in December of last year and continues to make steady progress. Users are checking it out slowly, but I don't expect a landslide migration. P6 will have to prove itself as a language.

Comment Re:Why has perl6 flopped? (Score 1) 281

I won't say, "I don't think it has," because it demonstrably has not.

The language has been released in open beta. It still has many properties that I think chase away those who approach it outside of language research communities. As a Perl 5 nostalgia fix, the learning curve is just too daunting, so as the beta progresses, I expect it to continue to build its own base of enthusiasts, the same way Perl did when it was first released.

So the language has not "flopped" yet because it hasn't had a chance to succeed yet.

It took Perl many years to go from a small toy that a trivial number of Usenet enthusiasts had heard of to a standard part of the Unix and Unix-like toolset. I don't think Perl 6 will gain traction any faster, especially given the learning curve. That's not flopping.

However, it has some substantial advantages over other languages. High on that list is the trivial nature of slinging highly functional grammars as first-class objects. That's something that you just can't do as easily in any other language that I know of. Perl 5 parsers and those of many other high level languages have some pretty severe performance penalties; yacc and its kin aren't dynamic enough; the various parser generators for Java are fast and mostly complete, but really painful to use.

Basically, you need a language that closely integrates grammars with the language itself in order gain the benefits of Perl 6. Here's and example parser I posted to reddit the other day:

A few other notable things that I think will draw people in:

The generalization of operators over iterable sequences and the hyper-operator version of reduction are features that you're going to hear a lot more about, I suspect. Perhaps in Perl 6, perhaps in other languages that adopt these ideas. I'm especially stunned by the utility of hyper-method-invocation (foo>>.method) which dispatches a given method over any iterable sequence of objects (whether they are the same type or not).

Full macros have not yet landed, not least because we've never had a full understanding of what macros would be. We know that they need to operate on the ASTs that represent code, and all of the self-hosting properties necessary to support that are there, but the exact syntax and semantics that are most Perl-friendly haven't fully gelled, yet. Once they do, I think that every language to have true macros in the past (mostly Lisp variants) has demonstrated the power of this tool.

A few other languages auto-generate accessors for classes, but I find the way Perl 6 does it to be a substantial improvement on the field, and it really is a joy to use. I think others will feel the same.

Speaking of objects, role composition will take some time for people to get used to, but as in other languages that have had similar features, I think this will be critical to Perl 6's adoption.

There are dozens of smaller features that are just quality-of-life benefits ranging from lexical variable/named parameter passing to the way any block can be turned into an anonymous closure and even curried. Some of these will be important to some, but not to others. It will be interesting to see it play out.

Comment Re:what do you think about the perl guy? (Score 1) 281

Well it's a combination of the automatic array flattening, and the fact that perl subroutines take an arbitrary number of arguments and just shift them off the list, allowing you to "overwrite" expected arguments with the flattened list -- ideally you might hope for an error with "too many arguments to function foo". The end result is that as long as you can force things to be arrays (or lists) when people don't necessarily expect that (like, say, the cgi module handling mutiple parameter definitions) you can get up to some serious funny business.

Slashdot Top Deals

Everybody likes a kidder, but nobody lends him money. -- Arthur Miller